Page 1
Energy Recovery in Water Distribution Systems by a Pump
Running as Turbine (PAT) – Grid Isolated Case
Bernardo Tomás Ribeiro de Brito Capelo
Thesis to obtain the Master of Science Degree in
Electrical and Computer Engineering
Supervisors: Prof. Paulo José da Costa Branco
Prof. Helena Margarida Ramos
Examination Committee
Chairperson: Prof. Rui Manuel Gameiro de Castro
Supervisors: Prof. Paulo José da Costa Branco
Members of Committee: Prof. Petra Amparo López Jiménez
May 2017
Page 3
iii
Acknowledgements
The accomplishment of this work would not have been possible without the help of a few people. First,
I would like to thank my supervisors, Prof. Paulo José da Costa Branco and Prof. Helena Margarida
Ramos, for their guidance and encouragement. I would like to thank Eng. Modesto Pérez Sánchez, for
his help and guidance through this project. In addition, I would like to thank to my family and girlfriend,
for constant support and motivation. Finally, I want to thank my group of colleagues and friends, for the
companionship and motivation during the difficult times.
Page 5
v
Abstract
The optimization of Water Distribution Systems (WDS) was the objective of many studies over the past
years. The reduction of the excessive water pressure in some nodes of a Water Distribution Network
(WDN), to reduce the water leakages, has been some of the main motivations of the studies of WDS.
Adding the high costs related to water pumping, the study changed from pressure reduction to combine
energy recovery and pressure reduction. Large efforts were made in this topic for the last years, and a
viable solution seems to be the use of Pump as Turbine (PAT), providing both the pressure control and
energy savings.
In this context, this work aims the analysis of the application of Pump as Turbine isolated from the grid.
A hydraulic pump is usually assembled with a squirrel cage induction generator, due to its low cost and
simplicity. Therefore, working as a PAT, the induction machine needs an external source of reactive
power. In this work, the proposed solution is the use of a bank of capacitors across the stator phases.
This induction machine application is named self-excited induction generator (SEIG). Both simulation
and experimental work is performed, in order to analyse the SEIG behaviour. The obtained results show
that this application can be an interesting solution for combining water pressure reduction with energy
generation, although for the specific system SEIG + PAT tested in laboratory, the global efficiencies
obtained were lower than expected, and so, an optimization of the system needs to be performed.
Keywords
Self-Excited Induction Generator (SEIG), Pump as Turbine (PAT), Water Distribution Networks (WDNs)
Page 7
vii
Resumo
A otimização de sistemas de distribuição de águas tem sido objetivo de muitos estudos nos últimos
anos. A redução do excesso de pressão em alguns nós de uma rede de distribuição de águas,
permitindo a redução de perdas de água, é uma das principais motivações destes estudos. Adicionando
os elevados custos relacionados com o bombeamento de água, novos estudos têm-se focado na
combinação da redução de pressão e recuperação de energia. Grandes esforços foram feitos neste
tema nos últimos anos, e uma possível solução viável é a utilização da bomba a funcionar como turbina
(PAT), permitindo tanto o controlo de pressão como recuperação de energia.
Neste contexto, este trabalho tem como foco a análise da aplicação da PAT isolada da rede. Uma
bomba hidráulica é normalmente usada com uma máquina de indução com rotor em gaiola, devido à
sua simplicidade e baixo custo. Funcionando como gerador, a máquina de indução requer de uma fonte
externa de reativa. Neste trabalho, a solução proposta consiste em colocar um banco de
condensadores ligado às fases do gerador. Esta aplicação, gerador isolado, tem o nome de gerador de
indução autoexcitado (SEIG). De forma a analisar o comportamento de um SEIG aplicado a um sistema
PAT, foi realizado trabalho de simulação e experimental. Os resultados obtidos demonstram que esta
aplicação é uma solução interessante na combinação da redução dos excessos de pressão e geração
de energia. Ainda assim, para o caso do sistema testado experimentalmente em laboratório, o
rendimento global obtido apresenta valores inferiores ao esperado sendo, por isso, necessário a
realização de uma otimização do sistema.
Palavras-chave
Bomba como Turbina, Gerador de Indução Autoexcitado, Redes de Distribuição de Água
Page 9
ix
Table of Contents
Acknowledgements ................................................................................................................................. iii
Abstract … ................................................................................................................................................v
Resumo .. ............................................................................................................................................... vii
Table of Contents .................................................................................................................................... ix
List of Figures .......................................................................................................................................... xi
List of Tables ......................................................................................................................................... xiii
List of Acronyms ..................................................................................................................................... xv
List of Symbols ..................................................................................................................................... xvii
1. Introduction ............................................................................................................................ 1
1.1. Objectives ................................................................................................................................ 2
1.2. Thesis Structure ....................................................................................................................... 3
2. State of Art: Review on Pump as Turbine (PAT) ................................................................. 5
2.1. PAT Regulation Designs .......................................................................................................... 6
2.1.1. Hydraulic Regulation ............................................................................................. 8
2.1.2. Electrical Regulation .............................................................................................. 8
2.1.3. Other Design Combinations .................................................................................. 8
2.2. Variable Operating Strategy (VOS) .......................................................................................... 9
2.2.1. VOS for Hydraulic Regulation .............................................................................. 10
2.2.2. VOS for Electrical Regulation .............................................................................. 13
2.2.3. Comparison of Results ........................................................................................ 16
2.3. System Effectiveness ............................................................................................................. 16
2.4. Economic Analysis ................................................................................................................. 18
2.4.1. VOS Economic Feasibility ................................................................................... 18
2.4.2. Optimal Management of Leakage Reduction and Energy Recovery .................. 20
2.5. Conclusions ............................................................................................................................ 21
3. Self – Excited Induction Generator .................................................................................... 23
3.1. Proposed System ................................................................................................................... 25
3.2. Simulation – Analytical and Computational Models ............................................................... 26
3.2.1. Analytical Model................................................................................................... 26
3.2.2. Computational Model ........................................................................................... 30
Page 10
x
3.3. Experimental Setup ................................................................................................................ 31
3.3.1. Induction Machine – characteristic parameters determination ............................ 33
3.3.2. DC Motor characteristic equations and its electric efficiency .............................. 36
3.4. Model Application: Analytical and Computational Simulation ................................................ 38
3.4.1. Analytical analysis – capacitance determination ................................................. 38
3.4.2. Dynamic analysis - computational model ............................................................ 42
3.4.3. Analysis of Simulations Results ........................................................................... 46
3.5. Experimental Work ................................................................................................................. 47
3.5.1. Tests and Results ................................................................................................ 47
3.6. Analysis and validation of simulation methods ...................................................................... 52
3.6.1. Comparison of required capacitances ................................................................. 52
3.6.2. Comparison of generator performances .............................................................. 55
3.7. Conclusions ............................................................................................................................ 57
4. SEIG application to a hydraulic PAT system .................................................................... 59
4.1. Experimental Setup ................................................................................................................ 59
4.2. Steady state operation: experimental tests and results ......................................................... 60
4.2.1. Analysis of the required capacitances and system behaviour ............................ 61
4.2.2. Global efficiency curves of the system ................................................................ 64
4.2.3. Hydraulic Q-H curve of PAT ................................................................................ 66
4.3. PAT – hydraulic efficiency curves .......................................................................................... 67
4.4. Analysis of transient state operation ...................................................................................... 68
4.5. Conclusions ............................................................................................................................ 75
5. Conclusions and Future Work ............................................................................................ 77
5.1. Conclusions ............................................................................................................................ 77
5.2. Future Work ........................................................................................................................... 78
Bibliography ......................................................................................................................................... 79
Apendix ................................................................................................................................................ A1
A. Equipment used in experimental applications ....................................................................... A1
B. Datasheet of the Induction Machine used in experimental and simulation analysis ............. A3
C. Tables with Experimental results of SEIG application ........................................................... A4
D. Tables with Experimental results of PAT application ............................................................. A8
Page 11
xi
List of Figures
Figure 1.1 – Illustrative scheme of the project in which this work is inserted. …………………………… 2
Figure 1.2 – Illustrative scheme of the case study of this work. …………………………………………... 3
Figure 2.1 – Characteristic curves for pump and turbine operating modes [7] …………………….……. 5
Figure 2.2 – Illustrative scheme of hydraulic and electrical regulation modes, adapted from [10] ...….. 7
Figure 2.3 – Scheme of PAT operating conditions for ER (right) and HR (left) [10] ..…………………... 7
Figure 2.4 – Daily Pattern from a PRV station [9] .………………………………………………………….. 9
Figure 2.5 – Working conditions of the PAT [9] ..…………………………………………………..………. 10
Figure 2.6 – Efficiency variation as function of BP, for different machines with HR mode [10] .…….. 12
Figure 2.7 – Regulation region and operating points for ER mode (BP = 22 m and D = 210 mm) [10]….13
Figure 2.8 – Efficiency variability of different machines, with different impeller diameter [10] .......…... 14
Figure 2.9 – Modified solution for the efficiency variability, with ER mode [10] ...…………………….... 15
Figure 3.1 – Illustrative image of both types of induction machines [19] ...……………………………… 22
Figure 3.2 – Equivalent circuit of an induction machine, adapted from [20] ...………………………….. 22
Figure 3.3 – Simplified equivalent of an induction machine, adapted from [20] .……………………..… 23
Figure 3.4 – Illustrative scheme of a IG system, grid connected [21] .…………………………………... 23
Figure 3.5 – Illustrative scheme of the self-excited induction generator system [21] ..……………….... 24
Figure 3.6 – Equivalent circuit of a Self-Excited induction generator, adapted from [27] .……..……… 25
Figure 3.7 – Equivalent circuit of SEIG with changes of (3.1) and (3.2) .………………………………... 26
Figure 3.8 – Simplified equivalent circuit of SEIG .………………………………………………………… 27
Figure 3.9 – Scheme of experimental installation used for the tests .……………………………………. 31
Figure 3.10 – Assemble Workbench for experimental tests .……………………………………………... 32
Figure 3.11 – No-load equivalent circuit of induction machine [30] .……………………………………... 33
Figure 3.12 – Equivalent circuit of an induction machine for blocked rotor case, adapted from [30] … 35
Figure 3.13 – DC Motor’s equivalent circuit, separate excitation case [31] .……………………………. 36
Figure 3.14 – Curve of Rotational Losses of the DC Motor .…………………………………………….... 38
Figure 3.15 – Description of the variation of bank’s capacitances with rotational speed, for specific
resistive loads .………………………………………………………………………………………………… 39
Figure 3.16 – Variation of the Capacitance as function of rotational speed, for different values of load
inductance – for a constant R = 230Ω .………………...………………………………………………..…... 41
Figure 3.17 – Illustrative scheme of the tests sequence of the simulation .……………………………... 42
Figure 3.18 – Simulation results of capacitance variation with rotational speed, for each load .……… 43
Figure 3.19 – Simulation results of SEIG efficiency curves .……………………………………………… 44
Figure 3.20 – Curve of capacitances that maximize efficiency, as function of the resistive load .….… 45
Figure 3.21 – Illustration of the sequence of the tests performed .………………………………………. 47
Page 12
xii
Figure 3.22 – Experimental results of capacitance required as function of rotational speed of generator,
for constant values of load .………………………………………………………………...…………………. 48
Figure 3.23 – Graphic of the SEIG Efficiency variation as function of rotational speed, for constant
load values ...……………………………………………………………………………………….………….. 49
Figure 3.24 – SEIG efficiency curve – average of most curves expressed in figure 3.23 …………….. 49
Figure 3.25 – Graphic of minimum capacitance in function of load .………..…………………………… 51
Figure 3.26 – Illustration of the comparison between the simulated and experimental results, for a
specific load of R = 230 Ω .…………………………………………………………...……………………….. 53
Figure 3.27 – Illustration of the minimum loads able to be handled by the generator, both experimental
and simulation cases .……………………………………………………………………………………….… 54
Figure 3.28 – Comparison between experimental and simulation efficiency variations, for
R = 230Ω …..…………………………………………………………………………………………………… 55
Figure 3.29 – Capacitance variation as function of load value – simulation and experimental comparison
– that maximizes the efficiency …………………………………………....…………………………………. 56
Figure 4.1 – Photos of the Installation assembled in Hydraulic Laboratory .…………………………..... 59
Figure 4.2 – Illustrative Scheme of the generating system .………………………………………………. 60
Figure 4.3 – Illustrative scheme of the hydraulic system assembled, adapted from [33] .…………….. 60
Figure 4.4 – Illustration of the sequence of the tests performed .………………………………………… 61
Figure 4.5 – Illustrative variation of the capacitance required as function of the rotational speed, for a steady flow regime …………………..……………………………..………………………………...……….. 62
Figure 4.6 – Global efficiency variations as function of the bank capacitances, for a constant flow
regime .…………………………………………………………………………………………………………. 62
Figure 4.7 – Curve of capacitances associated with the maximum global efficiency, as function of the
load values .……………………………………………………………………………………………………. 64
Figure 4.8 – Illustration of the sequence of tests performed with PAT .…………………………………. 64
Figure 4.9 – Illustration of the global efficiency curves as function of flow .…………………………….. 65
Figure 4.10 – Hydraulic Head/Flow Curves of the PAT .…………………………………………………... 66
Figure 4.11 – Illustration of the resultant PAT efficiency curves, as function of flow .………………….. 68
Figure 4.12 – Illustration of the hydraulic transient behaviour, pressure variation – close and open valve
cases – for N = 1560rpm .……………………………………………………………………………………... 69
Figure 4.13 – Illustration of the electrical transient phenomenon for N = 1560 rpm (R = 230Ω and
C = 15 µF) – close valve case .………………………………………………………………………………. 70
Figure 4.14 – Illustration of the electrical transient phenomenon for N = 1560 rpm (R = 230Ω and
C = 15 µF) – open valve case ….…………………………………………………………………………….. 71
Figure 4.15 – Illustration of the electrical transient phenomenon for N = 920 rpm (R = 230Ω and
C = 35 µF) – open valve case ………………………………………………………………………………... 73
Figure 4.16 – Simulink simulation results, of the generator excitation phenomenon, for the same load
and prime mover power input, with different capacitance values .………………………………………... 74
Page 13
xiii
List of Tables
Table 2.1 – Possible combinations of PAT regulation modes, adapted from [12]. …………………...…. 8
Table 2.2 – VOS results for Hydraulic Regulation [10]. .…………………………………………………... 12
Table 2.3 – VOS results for ER mode [10]. ….………………………………………….………………….. 15
Table 2.4 – Comparison between HR and ER – variation of BP influence on efficiency [10]. ……….. 16
Table 2.5 – VOS results for HR mode [11]. ….……………………………………………………………... 17
Table 2.6 – VOS results for ER mode [11]. …….…………………………………………………………... 18
Table 2.7 – Assessment of economic viability for both HR and ER modes, adapted from [10]. ….….. 19
Table 3.1 – Nameplate Information of both machines. ……………………………………………………. 31
Table 3.2 – Tested ranges of the resistive load and capacitances. …………………………………………... 33
Table 3.3 – Induction motor test results for no-load case. ………………………………………………… 34
Table 3.4 – Induction motor blocked rotor test results. ……………………………………………………. 35
Table 3.5 – Characteristic parameters of the induction machine in study. ……………………………… 36
Table 3.6 – Capacitance determination data results of figure 3.16. ……………………………………… 40
Table 3.7 – Limit reactance value associated to each different resistive value. ………………………... 41
Table 3.8 – Best efficiency points of the SEIG simulated results. ………………………………………... 45
Table 3.9 – Maximum efficiency points, for each tested load. ……………………………………………. 50
Table 3.10 – Comparison between obtained capacitance results for a specific load, R = 230 Ω. ….… 52
Table 3.11 – Comparison of results for the lower possible loads. ……………………………………….. 54
Table 4.1 – Synthetization of the Results .…………………………………………………………………... 63
Table 4.2 – Combinations of load and capacitance used to achieve the N values for the tests below
…………………………………………………………………………………………………………………… 65
Table 4.3 – Combinations of load and capacitance used to achieve the N for the tests below, with
correspondent 𝜂𝑒𝑙 .…………………………………………………………………………………………….. 67
Page 15
xv
List of Acronyms
AC Alternating Current
BEP Best Efficiency Point
CFD Computational Fluid Dynamics
DC Direct Current
ER Electrical Regulation
HR Hydraulic Regulation
IG Induction Generator
PAT Pump running as Turbine / Pump as Turbine
PRV Pressure Reducing Valve
SEIG Self-Excited Induction Generator
VOS Variable Operating Strategy
WDN Water Distribution Network
WDS Water Distribution System
WTS Water Transmission System
Page 17
xvii
List of Symbols
A Amp
𝑎 Stator frequency in values per unit
𝑏 Rotational speed in values per unit
𝐶 Capacitance value of the excitation system of the SEIG
𝑓 Induction generator actual stator frequency [Hz]
𝑓𝑠 Grid frequency [Hz]
F Farad
𝑔 Gravitational acceleration [m/s2]
𝐻𝑖 Available head drop [mwc]
Hz Hertz
𝐼0 Phase current for no load condition
𝐼𝑎 DC motor armature current
𝐼𝑐𝑐 Phase current for short-circuit / blocked rotor case
𝐼𝑓 DC motor field current
𝐼𝑚 Magnetizing current
𝐼𝑟′ Rotor current
𝐼𝑠 Stator phase current
𝐿𝐿 Load inductance
𝐿𝑚 Magnetizing inductance
𝐿𝑟 Rotor inductance
𝐿𝑠 Stator inductance
𝑁 Rotational speed of the induction machine [rpm]
𝑁1 Number of stator windings
𝑁2 Number of rotor windings
𝑁𝑠 Synchronous frequency [rpm]
𝑝 Number of pole pairs of the induction machine
Page 18
xviii
𝑃0 No load active power
𝑃𝑐𝑐 Short-circuit active power
𝑃ℎ𝑦𝑑 Hydraulic power available
𝑃𝐼𝑁 Total power at the entrance of the DC motor
𝑃𝑙𝑜𝑠𝑠 DC motor total losses
𝑃𝑚𝑒𝑐 Mechanical power
𝑃𝑂𝑈𝑇 DC motor mechanical output power
𝑃𝑟𝑜𝑡_𝑙𝑜𝑠𝑠 DC motor rotational losses
𝑃𝑠 Active power of the induction machine
𝑝𝑓 ≡ cos𝜑
Power factor
𝑄0 No load reactive power
𝑄𝑐𝑐 Short-circuit / blocked rotor reactive power
𝑄𝑖 Available flow discharge [l/s]
𝑄𝑠 Induction machine stator reactive power
𝑅𝐶 Induction machine magnetizing resistance, represents the core losses
𝑅𝑒 Resistance value of two stator phase of the induction machine
𝑅𝐿 Load resistance
𝑅𝑟′ Induction machine rotor resistance
𝑅𝑠 Induction machine stator resistance
𝑠 Slip
𝑆0 No load apparent power of the induction machine
𝑆𝑐𝑐 Short-circuit apparent power of the induction machine
𝑈𝑎 DC motor armature voltage
𝑈𝑓 DC motor field voltage
V Volt
𝑉0 Induction machine no load phase voltage
𝑉𝑐𝑐 Induction machine short-circuit phase voltage
𝑉𝑠 ≡ 𝑉𝑡 Induction machine phase voltage
Page 19
xix
𝑋𝑐 Capacitor Reactance
𝑋𝐿 Load inductive reactance
𝑋𝑟′ Induction machine rotor reactance
𝑋𝑠 Induction machine stator leakage reactance
𝑌𝐶 Admittance of the bank of capacitors
𝑌𝑖𝑛 Induction machine equivalent admittance
𝑌𝐿 Load admittance
𝑌𝑟 Equivalent admittance of rotor and magnetizing branch
𝑌𝑠 Stator equivalent admittance
𝑌𝑡 Total equivalent admittance of the SEIG
𝜔 Angular frequency of the machine [rad/s]
𝜔𝑟 Rotational speed of the rotor [rad/s]
𝜔𝑠 Synchronous frequency [rad/s]
𝜂𝑒𝑙_𝑠𝑖𝑚 SEIG electrical efficiency – Simulink simulation
𝜂𝑒𝑙 SEIG electrical efficiency
𝜂𝑃𝐴𝑇 PAT efficiency
𝜂𝑔𝑙𝑜𝑏𝑎𝑙 Global efficiency of the system
𝜌 Water Density [kg/m3]
Page 21
1
1. Introduction
The awareness of the depletion of natural resources and concern with the global environment has forced
a boost on the development of environmental friendly methods of energy generation. In this context,
investments were made in many renewable energy sources with small environmental impact, including
in hydroelectric power generation. Within the hydroelectric power generation, being one of the most
mature renewable energy sources, there has been an interesting increase in the small-scale and micro-
hydro power sources studies.
In some contexts, such as the case of rural and remote areas, the installation of micro-hydro power
source able to generate and store energy is a good alternative, however in certain cases it revealed
some problems related with costs, due to the need to miniaturize the turbine. With this constraint in
mind, a proposed solution was the use of a pump running in turbine mode (PAT), which represents a
high reduction in the implementation costs. Although it may not be as efficient alternative as a hydro-
turbine, it is considered to be an efficient method of generating energy, having some advantages such
as a simple construction and durability. The use of PAT is also considered a good and economic
approach, in locations where the water supply is relatively constant, such as water transmission systems
(WTS). Pump Manufacturers, in particular KSB, have been active investors in PAT for the past years.
The idea of using Pump running as Turbine is not exactly recent. During the 20st century, researchers
such as Stepanoff [1], Knapp [2], that in 1941 published the complete pump characteristics for a few
pump designs based on experimental investigations, and Kittredge [3], that in 1961 published a study
on centrifugal pumps used as hydraulic turbines (having an advantage, which is the initial cost
reduction), studied the work characteristics of a PAT. However, at that time there was still a lack of
technology to allow a proper use of PAT. Only in recent years, with the advances of technology, the
appearance of converters, sensors and so on, it allowed a completely new kind of control and use of
PAT.
The aim of this work is related to an idea for a new application for PAT that has recently emerged and
consists of using it to generate energy in water distribution systems (WDS). The existence of specific
locations in a water distribution network (WDN) where due to high water pressures, there is an increase
in water leakages and can cause serious damages in the pipes, lead to studies with the objective of
reducing the leakages. Currently, a solution for that consists of installing pressure-reducing valves
(PRVs) [4]–[6], reducing the water pressure and leakages, and so preventing investments in
rehabilitation of the pipelines. Therefore, the use of PRVs has been increasing over the last years. By
changing the PRV for hydropower devices, such as PATs, could be an interesting investment, and could
lead to large energy savings. The number of studies in this area has increased over the last few years,
with new ideas for using and controlling PATs in order to achieve the best performances and recover
higher amounts of energy.
Herein, a new approach for PAT application was proposed, more specifically, the application of PAT to
WDSs, with system isolated from the electrical grid.
Page 22
2
1.1. Objectives
This work is part of a project that aims to install a PAT in a PRV station, located in a water distribution
network. Figure 1.1 shows an illustrative scheme of the global project. The project is divided into three
different partial projects. The work performed in this thesis represents the partial project 1.
Figure 1.1 – Illustrative scheme of the project in which this work is inserted.
Usually, a hydraulic pump is assembled with an induction machine. An induction machine, to work,
needs a constant source of reactive power. As a generator connected to the electrical grid, the source
of reactive power is the grid itself. However, when the PAT system is isolated, the induction generator
needs another source of reactive power to be able to work.
A proposed solution consists of installing a bank of capacitors across the phases of the machine, as
shows figure 1.2, in order to provide the reactive power required to the induction generator.
Page 23
3
The objectives of this thesis can be divided into the following topics:
1. Complete study of the induction generator isolated from the grid;
2. Propose a solution system (figure 1.2);
3. Develop analytical and computational simulation models of the proposed system, and then an
experimental work in laboratory, with induction generator isolated from the grid, to analyse and
verify the work of the system;
4. Experimental work with the proposed solution applied to a hydraulic PAT system. The
experiment consists of assembling a hydraulic circuit and analyse the behaviour of the isolated
PAT + Induction Generator.
Figure 1.2 – Illustrative scheme of the case study of this work.
1.2. Thesis Structure
This thesis is divided into five chapters, described as follows:
- In Chapter 1 is provided an introduction and overview of the work;
- In Chapter 2 is done a brief review of the state of the art related to the use of PAT in WDS,
focusing on the latest works;
- In Chapter 3 is performed a complete study of the Isolated Induction Generator, named the Self-
Excited Induction Generator (SEIG), with both simulation and experimental work;
- In Chapter 4, a pilot hydraulic installation is assembled, in order to perform an experimental
study of a PAT + Induction Generator system, isolated from the grid, for both steady and
transient state operations;
- In Chapter 5, some general conclusions of the performed work are given, and also some
comments on the future work in this theme are made.
Page 25
5
2. State of Art: Review on Pump as Turbine (PAT)
In water supply systems, two types of machine can be used for energy generation, turbines of pumps
working as turbines (PATs). The traditional turbines provide better performances, as they are designed
for that specific application, however they present an expensive investment. Pumps are a more
economic solution, as they can be bought from the market, cheaper than turbines.
Figure 2.1 shows the characteristic curves of an example of centrifugal pump, working in pump operation
and turbine operation.
Figure 2.1 - Characteristic curves for pump and turbine operating modes [7].
The highlighted head and flow points represent the pump best efficiency points, both for pump and
turbine operations, BEP. Besides the head/flow curves, the efficiency (η) and power (P) curves, as
function of flow, are plotted in the figure, for a constant rotational speed.
In water transmission systems, pressure-reducing valves are installed in order to reduce the large
variations and dissipate residual head, therefore the power available is relatively constant, allowing the
use of traditional hydropower devices. In those cases, the use of PAT can be an efficient form of power
generation. In WDSs, the case is completely different, the necessity to fulfil the user demand is the most
important objective. In these systems, there are big variations of demand. Therefore, when designing a
water distribution network, flexibility to provide water for various demands is the main objective. With
this, problems started to emerge. The excessive pressure in some parts of the network can cause pipe
damages, increasing the water leakages. Besides this, another problem is related to the large amount
of energy spent in WDS. Energy consumption is about 80% of the total expenses in water distribution
[8]. The transportation of water to higher levels, done by using pumps, usually represents the
predominant component of the total operating cost of the system. Studies in this topic were done and,
in order to avoid expensive investments in pipes replacement in urban distribution networks, a
suggested solution was by placing PRVs. The valves provided a partial reduction of water losses, by
Page 26
6
reducing the pressure. This solution has proven to be effective in the reduction of water losses and
pressure control and so, the use of PRVs in water distribution systems has increased significantly.
Over the last years, studies were performed in order to replace the PRVs with small hydropower devices
to generate energy, in order to reduce the energy costs related with the WDSs. Adding this to the existent
problem of water leakages emerged the idea of using PAT. As it was already stated, PAT is considered
a cheap and sustainable solution. However nowadays there is still a lack of information related to PATs
behaviour, which can be a setback.
An important issue when analysing the economical convenience of this application of PAT is the power
available in the specific node of the WDN chosen to install the PAT. In WDSs, unlike the WTSs, the flow
rate and head drop present high variations depending on the user demands, and so the available
hydraulic power changes.
Ramos, Carravetta, Fecarotta and Giudice [9]–[12] described some design strategies for application of
PAT in WDS. Two main regulation designs were proposed, named Hydraulic Regulation [9] and
Electrical Regulation [10], both considered viable solutions. Recent studies also analysed some
combinations of both [12].
Furthermore in this chapter, a brief description of two design procedures proposed is performed, named
Variable Operating Strategy (VOS) [9], [10] and System Effectiveness [11]. The VOS design procedure
is a method to predict the PAT behaviour and find the optimal solution that maximizes the efficiency for
specific working conditions. However, this design does not have into account that the application of a
PAT in WDS means that the system will operate far from the best efficiency points, due to the high
variability in the working conditions (flow and pressure), and so it will cause a decrease in the reliability
of the system. Therefore, to take into account these important factors, System Effectiveness design was
proposed.
2.1. PAT Regulation Designs
An important aspect of this topic is the study of different ways to operate the PAT, in order to achieve
the best efficiencies and generate higher quantities of energy. The efficiency curve of a PAT, as function
of the flow discharge, presents a point of maximum efficiency (QBT, HB
T), named the best efficiency point
(BEP). In a WDN, the working conditions usually achieve values very far from the BEP of the PAT, so
an important objective was to find and study regulation designs that could adapt according with the
operating region, by changing PAT characteristic curves or the working conditions, in order to achieve
better performances. Here, a brief analysis of the different proposed regulation modes is performed,
focusing on the main designs.
An illustrative scheme of a hydropower plant installation is shown in figure 2.2, below. In the scheme
are represented both main regulation designs, electrical and hydraulic. In figure 2.3 are represented the
PAT operating conditions for both regulations.
Page 27
7
Figure 2.2 – Illustrative scheme of hydraulic and electrical regulation modes, adapted from [10].
Figure 2.3 – Scheme of PAT operating conditions for ER (right) and HR (left) [10].
Page 28
8
2.1.1. Hydraulic Regulation
As shows figure 2.2, this regulation mode consists of a PAT connected to the grid, with a series/parallel
set of hydraulic valves to help control and adjust the working conditions to a suitable area. The work of
the valves can be described into two different procedures:
1. In case the available head (𝐻𝑖) is higher than the head-drop deliverable by the machine (𝐻𝑖𝑇),
described by the points to the left side of the PAT characteristic curve (figure 2.3 – left), the
valve A will dissipate the excess pressure.
2. In case the flow (𝑄𝑖) is higher, given by the points to the right side of the PAT characteristic
curve, the PAT will produce a head-drop higher than the available head, then the valve B will
open in order to reduce the flow 𝑄𝑖.
2.1.2. Electrical Regulation
This design, instead of valves, consists of installing a regenerative converter between the PAT and the
grid. As shows the characteristic curves of figure 2.3 (right), the idea of this regulation mode, using the
converter, is to change the frequency of the PAT, and so, change its rotational speed to match the load
conditions that are determined by the instant flow discharge and head drop values. Basically, by
changing the rotational speed, the PAT characteristic curve changes, in order to match the required
values of the hydraulic parameters (𝑄𝑖 , 𝐻𝑖).
2.1.3. Other Design Combinations
Besides the two main modes of regulation already described, other possible combinations of both can
be assembled. From the scheme of figure 2.2, table 2.1 presents the other combinations using both the
inverter and one or two valves.
Table 2.1 – Possible combinations of PAT regulation modes, adapted from [12].
Device Case I (HR) Case II (ER) Case III Case IV Case V
Inverter No Yes Yes Yes Yes
Valve A Yes No No Yes Yes
Valve B Yes No Yes No Yes
Analysing all the regulation modes above proposed, cases III – V represent different combinations of
ER and HR.
Page 29
9
2.2. Variable Operating Strategy (VOS)
This model was proposed with the purpose of analysing the working conditions, the operating region of
a specific case, and help estimate the best PAT solution for that case. The VOS analysis was proposed
for both electrical and hydraulic regulations, comparing the obtained results for both types of PAT
application.
For the study, a daily demand pattern measured from a PRV station located in Pompeii, Italy, was
considered, as shows figure 2.4, containing the variations of flow rate (𝑄) and upstream pressure head
(𝑃𝐻). The value of head drop (𝐻) depends on the minimum pressure required to ensure a correct
pressure distribution through all pipes, named backpressure (𝐵𝑃), and it is given by equation (2.1). In
figure 2.4 is also shown the variation of available hydraulic power (𝑃).
𝐻 = 𝑃𝐻 − 𝐵𝑃 (2.1)
Figure 2.4 - Daily Pattern from a PRV station [9].
For this range of power, the installation of a turbine can be a problem due to its high costs because of
miniaturization of the regulation devices of the turbine. Therefore, the reduction of both manufacturer
and maintenance costs can be achieved by using a PAT instead of traditional turbine. This strategy,
VOS, allows to obtain the optimal PAT solution for each case (𝑄 − 𝐻 and 𝐵𝑃 required), choosing the
best PAT characteristics, impeller diameter and rotational speed.
When designing a hydropower plant in a WDN, there are two main problems: the lack of complete series
of characteristic curves for industrial PAT, and the absence of a strategy for turbine selection and overall
efficiency computation.
Therefore, to determine both characteristic and efficiency curves of a PAT, three different possible ways
were proposed: experimentally [13], by computational fluid dynamics (CFD) [14] and by one-dimensional
Page 30
10
methods [1]. The most reliable is the experimental analysis, but it requires a huge number of
experimental tests. Therefore, simulation by CFD could be considered a valid alternative. After
determining the characteristic curve of a specific PAT, the results could be extended to any PAT of the
same type (e.g. centrifugal, semi axial,…), similar to the reference PAT (different diameter and rotational
speed), using the Suter parameters [15]. Two different PATs, of the same type, are considered similar
if they have the same specific speed, either turbine (𝑁𝑆𝑇) or pump (𝑁𝑆
𝑃).
𝑁𝑆𝑇 = 𝑁𝐵
𝑇 𝑃𝐵𝑇12
𝐻𝐵𝑇43
𝑁𝑆𝑃 = 𝑁𝐵
𝑃 𝑄𝐵𝑃12
𝐻𝐵𝑃34
(2.2)
The parameters of equation (2.2) are at the best efficiency point (BEP), and are the rotational speed
(𝑁𝐵), flow rate (𝑄𝐵), head drop (𝐻𝐵) and power (𝑃𝐵), for both turbine and pump operating modes.
2.2.1. VOS for Hydraulic Regulation
The working conditions shown in figure 2.4 are expressed below, as function of discharge and head
drop, in figure 2.5.
Figure 2.5 – Working conditions of the PAT [9].
Also, in the figure above is shown the characteristic curve of a PAT example. The area filled by the
working points is named the operating region, as shown.
The proposed model, VOS, was characterized by the following steps:
1. measured pattern of discharge and pressure conditions and the available head is determined
based on the required value of BP;
2. consider a type of PAT (centrifugal, axial…);
3. consider a set of different PAT characteristic curves for the operating region;
Page 31
11
4. computation of the overall plant efficiency, using the equation below,
𝜂𝑝 =∑ 𝐻𝑖
𝑇𝑄𝑖𝑇𝜂𝑖
𝑇Δ𝑡𝑖𝑛𝑖=1
∑ 𝐻𝑖𝑄𝑖Δ𝑡𝑖𝑛𝑖=1
(2.3)
𝜂𝑝 - Overall plant efficiency; 𝑛 - number of available operating points; 𝐻𝑖𝑇 and 𝑄𝑖
𝑇 - head drop
and discharge delivered by the PAT; 𝜂𝑖𝑇 - PAT efficiency; 𝐻𝑖 and 𝑄𝑖 – available head drop and
discharge; Δ𝑡𝑖 – time interval discretization of the operating points;
5. the PAT that maximizes the overall efficiency is considered the optimal design solution;
6. optimal machine selected from the market, and then the actual efficiency is verified.
However, as already referred before, to perform steps 3 and 4, a wide set of characteristic curves was
needed, with their BEPs characterized. This represented a problem due to the lack of data available to
technicians. Therefore, a proposed solution was using the affinity law [16], equation (2.4), where having
a reference estimated characteristic curve for a PAT (𝐼), allowed to estimate the curves for similar
machines (that verify condition of equation 2.2).
𝑁𝐵𝐼
𝑁𝐵𝐼𝐼 =
𝐷𝐼𝐼
𝐷𝐼(𝐻𝐵𝐼
𝐻𝐵𝐼𝐼)
1/2
= (𝑄𝐵𝐼𝐼
𝑄𝐵𝐼 )
1/2
(𝐻𝐵𝐼
𝐻𝐵𝐼𝐼)
3/4
= (𝑃𝐵𝐼𝐼
𝑃𝐵𝐼 )
1/2
(𝐻𝐵𝐼
𝐻𝐵𝐼𝐼)
5/4
(2.4)
Knowing the diameter, 𝐷𝐼𝐼 and rotational speed 𝑁𝐵𝐼𝐼 of the similar PAT, it was possible to estimate its
BEPs (𝑄𝐵𝐼𝐼, 𝐻𝐵
𝐼𝐼) and produced power (𝑃𝐵𝐼𝐼). This estimation can be used for any machine similar to the
reference machine.
The developed model was applied in two different cases [9], [10]. Here is described an application to a
centrifugal pump, single stage, with a specific speed 𝑁𝑆𝑇 = 44 [rpm, kW], in order to compare with the
Electrical Regulation application [10]. The data used for the application was from the measured pattern
of the PRV station in Pompeii, shown in figure 2.3.
As already said, the available head depended on the backpressure (𝐵𝑃) value. The 𝐵𝑃 is a design
parameter that depends on the downstream network characteristics, and different 𝐵𝑃 values will
produce different scenarios. The considered 𝐵𝑃 values were between 0 and 70 m, in order to include a
wide set of scenarios.
The results are expressed in table 2.2, applied for different impeller diameters 𝐷 [mm] and rotational
speeds 𝑁 [rpm]. Below, in figure 2.6, is shown the variation of efficiency as function of 𝐵𝑃 value of
different machines, for different rotational speeds, as well as for different diameters for the same 𝑁 =
3000 rpm.
For 𝑁 = 3000 rpm, the bold curve, 𝜂𝑝𝐻𝑅|𝑚𝑎𝑥, represented the maximum efficiency curves. For different
diameter machines, the intersection with maximum efficiency curve, i.e., 𝜂𝑝𝐻𝑅|𝑚𝑎𝑥 (𝐵𝑃) = 𝜂𝑝
𝐻𝑅|𝐷 (𝐵𝑃)
represented the points where the correspondent diameter, 𝐷, was the VOS solution.
Page 32
12
Table 2.2 – VOS results for Hydraulic Regulation [10].
BP [m] D [mm] N [rpm] ηpHR |max
10 186 3000 0.59
15 190 3000 0.59
20 191 3000 0.59
25 187 3000 0.56
30 181 3000 0.53
35 176 3000 0.48
40 208 3000 0.47
45 219 1500 0.51
50 232 1500 0.54
55 237 1500 0.55
60 221 1000 0.50
Figure 2.6 – Efficiency variation as function of BP, for different machines with HR mode [10].
Page 33
13
2.2.2. VOS for Electrical Regulation
For ER mode, as already said, the application of the PAT consists of inserting a regenerative converter
between the machine and the electrical grid. The converter allows changing the frequency of the
machine, and so, the rotational speed will change too. The rotational speed can be determined by the
expression of equation (2.5), as function of the frequency 𝑓 [Hz] and number of pole pairs 𝑝.
𝑁 = 60𝑓
𝑝 (2.5)
For safety reasons, the frequency variation must be limited to a specific range of values. The upper limit
must not be higher than 60 Hz, which the maximum designed frequency. The lower limit should not be
lower than 50 % of the grid frequency, so 25 Hz.
Below, in figure 2.7, are expressed the working conditions for ER mode, for a specific machine with an
optimal diameter 𝐷 = 210 mm and a 𝐵𝑃 = 22 m. In the figure are expressed the characteristic curves
for different frequencies, representing the regulation region, as well as the operating points from the
data shown in figure 2.4. As it is possible to observe, the BEP line crossed all the characteristic curves
in their maximum efficiency points.
Figure 2.7 – Regulation region and operating points for ER mode (𝐵𝑃 = 22 m and 𝐷 = 210 mm) [10].
As shown in the figure above, the highest values of the net head correspond to the highest speeds, and
for higher flow-rates it is required a frequency reduction, in order to reduce the rotational speed.
For ER mode, the VOS was similar to the model developed for HR mode, but with some modifications.
In this model, there was only one design parameter, which was the impeller diameter. Then, for this
regulation mode, the VOS model was described by the following steps:
1. measured pattern of discharge and pressure conditions and the available head is determined
based on the required value of BP;
Page 34
14
2. consider a type of PAT (centrifugal, axial…);
3. consider a set of different PAT characteristic curves for the operating region;
4. change, for each PAT selected, the rotational speed within the range limit;
5. computation of the overall plant efficiency, using the equation below,
𝜂𝑝𝐸𝑅 =
∑ 𝐻𝑖𝑄𝑖𝜂𝑖𝑇(𝑓)Δ𝑡𝑖
𝑛𝑖=1
∑ 𝐻𝑖𝑄𝑖Δ𝑡𝑖𝑛𝑖=1
(2.6)
𝑛 - number of available operating points; 𝜂𝑖𝑇 - PAT efficiency, dependent on the frequency, 𝑓;
𝐻𝑖 and 𝑄𝑖 – available head drop and discharge; Δ𝑡𝑖 – time interval;
6. the PAT that maximizes the overall efficiency is considered the optimal design solution;
7. near-optimal machine selected from the market, and then the actual efficiency is verified.
Just like already said for HR mode, due to the lack of characteristic and efficiency curves for a set of
PATs, a proposed solution was using the affinity law [16], equation (2.4), where having a reference
estimated characteristic curve for a PAT, allowed to estimate the curves for similar machines.
The application of this design procedure was done to the same centrifugal pump of HR mode (single
stage, 𝑁𝑆𝑇 = 44 [rpm, kW]). In order to compare with the HR results, this application was performed for
the same range of BP values. For each BP value, an optimal diameter was found.
Figure 2.8 shows the ER results, for different machines, as function the BP value. For example, for a
𝐵𝑃 = 22 m, as shown in figure 2.7, the diameter chosen should be 𝐷 = 210 mm, which provided the
higher efficiency of 57 %.
In addition, from figure 2.8, it is possible to observe that for BP values higher than 41 m, the electrical
regulation became impossible for the simulated machines, and so, the ER could not ensure the work for
all the BP values.
Figure 2.8 – Efficiency variability of different machines, with different impeller diameter [10].
Page 35
15
To increase the elasticity of the ER system, a 𝛥 = 10 % variation of the BP was taken into account, i.e.,
for a specific BP value, for example 22 m, the chosen machine should be able to work between 19.8 m
and 26.4 m. In this specific case, the chosen machine would have an impeller diameter 𝐷 = 215 mm,
instead of the 𝐷 = 210 mm previously chosen. In addition, the new highest 𝐵𝑃 ensured was of 37 m,
and the new best efficiency was of 54.4 %, presenting an average decrease of efficiency of 16.3 %.
Below in figure 2.9 is plotted the same curve of figure 2.8, but considering the 𝛥 = 10 % variation for
every scenario.
Figure 2.9 – Modified solution for the efficiency variability, with ER mode [10].
The VOS results of the ER mode, for the different scenarios considered, are presented in table 2.3,
containing for each 𝐵𝑃 value, the best impeller diameter with the correspondent efficiency.
Table 2.3 – VOS results for ER mode [10].
BP [m] D [mm] ηpER |max
10 238 0.31
15 230 0.40
20 220 0.50
23 216 0.54
25 219 0.54
30 232 0.49
35 251 0.40
37 270 0.28
40 - -
Page 36
16
2.2.3. Comparison of Results
Looking at both hydraulic and electrical regulation results, in tables 1 and 2, is it possible to observe that
the HR mode presented higher efficiencies for all the 𝐵𝑃 values, although 𝐵𝑃 between 23 m and 37 m
they showed similar efficiencies.
In ER mode, by changing the frequency, the characteristic curve could adjust to the operating point, as
shows figure 2.6. However, the operating point being in the characteristic curve was not enough, and in
some points, the operating point was too far from the BEP, then presenting a low hydraulic efficiency.
When talking about a WDN, the values measured should not be considered true. The parameter values
depend on the demand at each moment. Therefore, in table 2.4 are expressed the efficiency impact of
a 𝐵𝑃 variation, for the ER and HR results previously expressed. As it is possible to see, the HR usually
presented higher efficiency and higher flexibility than ER.
Table 2.4 – Comparison between HR and ER – variation of 𝐵𝑃 influence on efficiency [10].
% variation of ηpHR % variation of ηp
ER
BP [m] ηpHR - 10% BP + 10% BP ηp
ER - 10% BP + 10% BP
10 0.586 -0.485 0.387 0.305 -2.349 2.244
20 0.585 1.045 -1.492 0.502 -2.418 2.246
30 0.531 3.356 -5.285 0.488 -4.497 4.126
35 0.484 6.405 -12.791 0.403 -8.473 7.450
2.3. System Effectiveness
The Variable Operating Strategy (VOS), proposed by Carravetta et al [9], [10] and described above,
analysed the overall efficiency and flexibility of different PATs for different scenarios, choosing the
optimal solution for each one. However, this application is for a WDN, where there is a high variation of
the working conditions, forcing the PAT to work far from its BEP, decreasing its reliability.
Therefore, based on the VOS, another design strategy was proposed by Carravetta et al [11],
considering three different factors:
- Capability of the PAT performance;
- Flexibility of the PAT under small variation of the working conditions;
- Reliability of the PAT based on the lifecycle of its components.
The effectiveness of PAT is described by equation (2.7). The factors that compose the equation are the
Page 37
17
system’s capability (𝜂𝑝), flexibility (𝜙p) and reliability (𝜇𝑃).
𝐸 = 𝜂𝑃𝜙𝑃𝜇𝑃 (2.7)
The effectiveness factors represent, respectively, if the performance of the PAT is according with
expectations, how is the performance for a 𝐵𝑃 variation around the design value and if it operates for a
given time without failing.
The three effectiveness indicators were analysed by Carravetta et al [11]. The effectiveness strategy
was applied for the same centrifugal pump used for the VOS model (single stage, 𝑁𝑆𝑇 = 44 [rpm, kW]),
both for HR and ER mode. Previously, the VOS model focused on the overall efficiency of the PAT, with
the objective of finding the optimal machine for specific work conditions (VOSη). Here, an application of
the system effectiveness was proposed (VOSE).
In tables 2.5 and 2.6 are shown the obtained results for both HR and ER modes.
Table 2.5 – VOS results for HR mode [11].
VOS BP [m] 10 20 30 35
VOSη
D [mm] 186 191 181 176
ηp 0.590 0.595 0.532 0.483
φp 0.997 0.984 0.952 0.893
µp 0.878 0.666 0.868 0.603
E 0.517 0.390 0.440 0.260
VOSE
D [mm] 179 178 171 163
ηp 0.567 0.549 0.499 0.439
φp 0.999 0.990 0.956 0.923
µp 0.933 0.976 0.985 0.971
E 0.529 0.531 0.471 0.394
Looking at table 2.6, it is possible to observe that, for VOSη, the ER mode was not flexible at all, problem
that was already discussed in the previous section 2.2.
Comparing the VOSη and VOSE, it is possible to see that the decrease of capability when applying the
effectiveness was very small. However, the results of the new procedure showed an 𝐸 in HR mode
better than ER mode for all considered 𝐵𝑃 values. The reason of this difference rely on the fact that both
capability and reliability presented lower values when compared to HR mode. On the other side, the
Page 38
18
flexibility of both modes presented close values.
Table 2.6 – VOS results for ER mode [11].
VOS BP [m] 10 20 30 35
VOSη
D [mm] 237 216 222 234
ηp 0.313 0.527 0.549 0.513
φp 0.000 0.000 0.000 0.000
µp 0.704 0.914 0.917 0.899
E 0.000 0.000 0.000 0.000
VOSE
D [mm] 239 221 231 250
ηp 0.298 0.495 0.495 0.410
φp 0.977 0.976 0.957 0.924
µp 0.686 0.894 0.891 0.809
E 0.199 0.432 0.422 0.306
2.4. Economic Analysis
The study of the installation of PAT in WDN instead of PRVs, allowed the development of application
designs that maximize the overall efficiencies and produced energy for different scenarios.
Besides this, an effort was made to develop economic assessments for each regulation designs
proposed. Carravetta et al [10] performed an economic analysis of VOS for both HR and ER modes.
In addition, Tricarico et al [17], developed a novel methodology for WDS pressure management, to
accomplish both leakage reduction and minimization of the difference between annual pumping costs
and PAT income.
2.4.1. VOS Economic Feasibility
After the strategies and simulations above analysed, Carravetta et al [10] also performed an economic
assessment of both regulation modes (for VOS results of tables 2.2 and 2.3), for the different scenarios
presented in table 2.4.
Page 39
19
The results are shown in table 2.7. The parameters included in the table are described below.
- PB – power of the machine at the BEP;
- Pmax – maximum output power of the generator;
- CPAT – PAT cost, considered to be 230 €/kW;
- Cgen – generator cost was evaluated on the Pmax on each scenario, considered to be 115 €/kW;
- Cvalve – cost of the hydraulic valves used in HR mode, considered to be 2500€ each;
- Cinve – cost of the inverter used in ER mode, dependent on the generator power, with a cost of
200 €/kW;
- ADEP – average daily energy production;
- PPeme – represents the payback period.
The value considered for the cost of produced energy was of 0.20 €/kWh. The costs of civil works and
maintenance of the equipment were not taken into account because they were more or less similar for
both cases.
Table 2.7 – Assessment of economic viability for both HR and ER modes, adapted from [10].
BP [m] 10 20 30 35
Mode HR ER HR ER HR ER HR ER
D [mm] 186 238 191 220 181 232 176 251
ηp 0.59 0.31 0.59 0.50 0.53 0.49 0.48 0.40
PB [kW] 8.49 29.12 9.69 19.65 7.41 25.63 6.44 37.99
Pmax [kW] 12.79 12.29 10.86 11.92 7.97 9.78 6.71 8.51
CPAT [€] 1953 6698 2230 4520 1704 5895 1481 8738
Cgen [€] 1471 1414 1248 1372 916 1125 772 978
Cinv [€] - 2459 - 2385 - 1957 - 1701
Cvalve [€] 5000 - 5000 - 5000 - 5000 -
Ctot [€] 8423 10570 8478 8277 7620 8977 7253 11418
ADEP [kW/h/day] 226.30 139.92 146.13 123.84 159.67 147.62 84.59 70.49
Daily Income [€/day] 53.26 27.98 29.23 24.77 31.93 29.52 16.92 14.10
PPeme [days] 158 378 290 334 238 304 428 810
Page 40
20
The obtained results showed that the payback period was shorter for HR mode than for ER mode, for
all the considered scenarios. In the case of the HR mode, the payback period for all the BP considered
was around one year. For ER mode, the case was the same except for the 𝐵𝑃 = 35 m, where the
payback period was higher than two years.
However, as already said, this payback period did not consider the maintenance and civil works.
Therefore, the payback period of the whole hydropower system could be much longer.
2.4.2. Optimal Management of Leakage Reduction and Energy Recovery
As already said, the costs of pumping represent a predominant component of the total operating costs
of the system. With that in mind, Tricarico et al [17] developed an optimization approach, by using the
PAT [9]–[12] application to generate energy and reduce the pumping energy costs.
This study was performed, to the case of the trunk mains model of Sorrento Peninsula, Italy. It supplies
eight urban areas, with around 90,000 user daily, and 70,000 additional user, during the summer. The
proposed simulation method consisted of replacing PRV in pumping stations with PAT. The number of
PATs to place were 10 or 15.
In the results analysis, they concluded that replacing PRVs with PATs presented a good solution and,
for that particular case, if the actual PRVs presented in the system were replaced by PATs, the income
generated could cover almost all of the operation costs related with pumping.
Page 41
21
2.5. Conclusions
Over the past years, a large effort has been made to find alternatives to PRV, that allow not only the
pressure control and leakage reduction, but also to generate energy. In addition, in WDS, the costs of
pumping represent a predominant component of the total operating costs of the system. Joining both
problems, a proposed solution consisted of the replacement of the PRVs with PATs, allowing large
energy savings.
Carravetta et al [9]–[12], proposed a few regulation modes, with two main designs, Electrical and
Hydraulic regulations. In addition to the regulation modes, two strategies were proposed, that have the
purpose of determining the optimal machine for each specific scenario. According to the obtained
results, both were considered viable solutions, with interesting estimated payback periods. However,
the constant dependence that the working conditions have on the users demand still represents an
issue, due to its difficulty of prediction.
The lack of available characteristic and performance curves represent another constraint. Simulation
methods, like computational fluid dynamics, and the use of affinity law, are considered options to bypass
the problem. However, they can lead to error, which can be an issue if the chosen PAT does not match
the expected characteristic operating conditions, endangering the viability of the project. To contradict
these issues, as already said, some pump manufacturers have been active investors in PAT over the
past years.
Besides this, Tricarico et al [17] developed a method to analyse the influence of the replacement of
PRVs with PATs in the overall costs of operations related to pumping. Again, the results showed that
the used of PAT in WDN can be an interesting application, and can provide large quantity of energy
savings.
Concluding, according to all researches considered here, the replacement of PRVs with PATs it shows
to be a very promising solution for both leakage reduction and energy savings.
Further, in this work, a different approach of PAT application design was proposed, with the system
isolated from the grid.
Page 43
23
3. Self – Excited Induction Generator
An Induction Machine is an alternating current machine. It operates by supplying the stator directly with
alternating current, which will supply the rotor by induction from the stator. It has the same physical
stator as a synchronous machine with a different rotor construction, no needing of DC field current. Due
to its work mode, induction in the rotor by the stator, its rotational speed is always lower than the
synchronous speed (as a motor), and this variation is named slip. There are two types of three-phase
induction machines, wound and squirrel-cage rotor. A wound rotor is composed by a three-phase
winding, and can be accessed by a set of carbon brushes. A squirrel-cage rotor is composed by
conducting bars embedded in slots in the rotor, short-circuited at each end by conducting end rings [18].
Due to its simplicity, the squirrel-cage design presents high advantages, and so it is the most commonly
used. For motor type operation, induction machines are by far the most common type used in industrial
and commercial applications.
Figure 3.1 - Illustrative image of both types of induction machines [19].
The equivalent circuit of an induction machine is expressed in figure 3.2. The circuit represents one
phase of the machine.
Figure 3.2 – Equivalent circuit of an induction machine, adapted from [20].
Page 44
24
The equivalent circuit can be divided into three different parts, stator, magnetization and rotor. The
resistance 𝑅𝑠 represents the resistance of the stator winding and 𝑋𝑠 represents the stator leakage
reactance. The resistance 𝑅𝐶 represents the core losses and 𝑋𝑚 represents the magnetization
reactance. For the air gap, 𝑁1 and 𝑁2 represent an ideal transformer between stator and rotor windings.
Finally, for the rotor side, 𝑋𝑟 and 𝑅𝑟
𝑠 represent the rotor’s reactance and resistance, respectively.
Below, in figure 3.3, is expressed a simplified model, in which the air gap (ideal transformer) area is
neglected, and so the rotor values are now referred to the primary (stator).
Figure 3.3 – Simplified equivalent of an induction machine, adapted from [20].
Working in generator regime, the induction machine offers advantages in terms of cost and simplicity
for both hydro and wind applications. However, an induction generator needs to have a constant reactive
power source to work. When connected to the grid, as shows figure 3.4, the reactive power is supplied
by the grid. If the purpose is to isolate the system (e.g. for remote areas), an external reactive power
source must be implemented.
Figure 3.4 – Illustrative scheme of a IG system, grid connected [21].
Herein, an analysis of the Induction Generator isolated from the grid is performed. This type of IG
application is named the Self-Excited Induction Generator (SEIG) [22].
Page 45
25
This analysis is divided in two main parts, Simulation and Experimental. In the simulation, two different
models are developed for the SEIG analysis. Then, an experimental work is executed, in order to obtain
both capacitance and performance curves of a specific induction machine. In addition, a comparison
between the simulation and experimental results is performed.
3.1. Proposed System
In this work, the Pump running in turbine mode (PAT) is going to be analysed isolated from the grid. Like
most hydraulic pumps, being assemble with a squirrel-cage induction machine, it cannot operate
isolated from the grid by itself, and so needs an external excitation source.
As expressed below, in figure 3.5, the proposed system is composed by connecting a capacitor bank
across the stator terminals to supply the required reactive power to the induction machine.
When the machine starts rotating, driven by a mechanical prime mover, a small voltage is induced in
the stator with a frequency proportional to the rotational speed of the rotor, due to residual magnetism
present in the rotor from previous utilizations. With this small voltage available, the capacitors produce
current, providing reactive power, 𝑄, increasing the voltage, and stablishing a magnetizing flux in the
machine. So, the induction machine is excited.
Figure 3.5 – Illustrative scheme of the self-excited induction generator system [21].
After the machine is excited, depending on the available power provided by the prime mover, the
generator will produce active power, 𝑃, that will be delivered to the load, and will consume reactive
power, 𝑄, as shown in the scheme of figure 3.5. However, the reactive power produced by the capacitors
needs to be sufficient for both the load and the machine excitation. The main drawback of this SEIG
system is its poor voltage and frequency regulations under variable loads. A change in the load
impedance directly influences the excitation capacitance of the system, because the reactive power
produced by the capacitors is shared by both the machine and load. It also influences the generator
frequency (and consequently, rotational speed) and stator voltage.
Page 46
26
The prime mover can be any mechanical rotating device. In this work, two different prime movers were
used. For the SEIG analysis, the prime mover chosen was a DC Motor. For the PAT analysis, the prime
mover was the Pump itself.
In figure 3.5, the presented bank of capacitors is connected in a delta design, but it is not the only
connection design. In this project, the wye connection was chosen for the experimental and simulation
procedures. As for the load, there are many options depending on the purpose of the work. Here, the
general type of load considered was RL (resistive-inductive) even though, for experimental procedures,
the considered load was purely resistive.
3.2. Simulation – Analytical and Computational Models
In this section, two different models for the analysis of a self-excited induction generator are developed.
The two models play different roles in the analysis of the system, both of them important. The first
consists of an analytical model that allows determining the capacitance required for different cases. The
second model is a simulation model developed in the MATLAB tool, Simulink.
3.2.1. Analytical Model
The minimum capacitance needed to keep the generator excited for the no load case can be easily
determined by analysing the magnetizing curve of the induction machine that can be obtained
experimentally, for no load [23]. However, when a load is applied, the capacitance value changes,
depending not only on the parameters of the machine, but also on the load applied to the system and
the rotational speed of the generator. Therefore, before starting experiment with the machine or even
assemble the system, it would come in handy to have a model that depending on the machine and load
in question, allows making a first estimation of the range of capacitances needed for that specific
application.
So, with that in mind, in this section was developed a general analytical model that allows determining
the range of capacitances required, for any case. Over the last years, many studies were done in the
SEIG [24]–[26], both simulation and experimentally. The most seen analytical model is the Loop Analysis
model, which works with the equivalent circuit impedances [23]–[29], however it is an iterative model,
turning the analysis more complex. Herein, a different model was developed, focusing on the
admittances of the circuit, based on [27], [28].
The equivalent circuit of a SEIG, based on the equivalent circuit of figure 3.3, is expressed in figure 3.6.
In each phase of the machine, a load and capacitor were added. Considering a load type RL, in this
section the objective is to obtain an expression for the minimum capacitance depending on the rotational
speed, parameters of the machine, the generator’s frequency and the load RL applied to the system.
Page 47
27
Figure 3.6 – Equivalent circuit of a self-excited induction generator, adapted from [27].
For the SEIG, the stator frequency 𝑓 [Hz] can be different from the grid frequency, 𝑓𝑠 = 50 Hz, therefore
in this analysis is convenient to change the frequency to values per unit, in order to facilitate its
determination. For the rotational speed 𝑁𝑟 [rpm], the same changes were done, converting it in values
per unit as function of the nominal frequency from the stator connected to the grid, 𝑁𝑠 [rpm]. These
changes are expressed in equation (3.1), and they were also applied to the slip equation (3.2).
𝑎 =
𝜔
𝜔𝑠=
2𝜋𝑓
𝑝2𝜋𝑓𝑠𝑝
=𝑓
𝑓𝑠
𝑏 =𝜔𝑟
𝜔𝑠=
𝑁𝑟
𝑁𝑠=
𝑁𝑟
60𝑓𝑠𝑝
(3.1)
𝑠 =𝜔−𝜔𝑟
𝜔=
𝜔
𝜔𝑠−𝜔𝑟𝜔𝑠
𝜔
𝜔𝑠
=𝑎−𝑏
𝑎 (3.2)
So, 𝑎 [pu] and 𝑏 [pu] represent the stator frequency and rotational speed, respectively, in values per
unit. Then, from equations (3.1) and (3.2), the first unknown variable of this model is defined, the
frequency 𝑎 [pu], which only with its determination for each specific case, allows computing the
capacitance values required. Applying the changes of equations (3.1) and (3.2) in the equivalent circuit
of figure 3.6, resulted in figure 3.7, below.
Figure 3.7 – Equivalent circuit of SEIG with changes of (3.1) and (3.2).
Page 48
28
Then, dividing all parameters by frequency 𝑎, the figure 3.8 is obtained, being this equivalent circuit the
base for the further simplifications and computations.
Figure 3.8 – Simplified equivalent circuit of SEIG.
Applying the nodal analysis to the terminal voltage, 𝑉𝑡, equation (3.3) is obtained. The total
admittance, 𝑌𝑡, is given by equation (3.4),
𝑉𝑡
𝑎∙ 𝑌𝑡 = 0 (3.3)
𝑌𝑡 = 𝑌𝑖𝑛 + 𝑌𝐿 + 𝑌𝐶 (3.4)
where 𝑌𝐶 represents the capacitors admittance, given by equation (3.5) and 𝑌𝐿 represents the load
admittance, described by equation (3.6). The admittance 𝑌𝑖𝑛 represents the induction machine’s
equivalent admittance, as can be detailed expressed in equations (3.7 – 3.9).
𝑌𝐶 = −𝑎2
𝑗𝑋𝐶 (3.5)
𝑌𝐿 =1
𝑅𝐿𝑎+𝑗𝑋𝐿
(3.6)
𝑌𝑖𝑛 =𝑌𝑠𝑌𝑟
𝑌𝑠+𝑌𝑟 (3.7)
𝑌𝑠 =1
𝑅𝑠𝑎+𝑗𝑋𝑠
(3.8)
𝑌𝑟 =1
𝑅𝑟′
𝑎−𝑏+𝑗𝑋𝑟
′+
1
𝑗𝑋𝑚 (3.9)
As shown in the admittance equations above, another unknown variable is the capacitors reactance,
𝑋𝐶 [Ω]. Knowing that the terminal voltage is always different from zero, the condition of equation (3.10)
is obtained.
𝑌𝑡 = 0 ⇔ 𝑅𝑒𝑌𝑡 = 0
𝐼𝑚𝑌𝑡 = 0 (3.10)
Page 49
29
There are two unknown variables, so to solve the condition shown above, in equation (3.10), the
objective is to simplify both real and imaginary expressions, in order to have the real equation only as
function of frequency 𝑎. Then, it cannot depend on 𝑋𝐶. Solving, in equation (3.11) is expressed the
resultant equation for the real part.
𝐶4𝑎4 + 𝐶3𝑎
3 + 𝐶2𝑎2 + 𝐶1𝑎
1 + 𝐶0 = 0 (3.11)
The coefficients 𝐶𝑛, n = 0,1,2,3 and 4, are described in equations below,
𝐶4 = 𝑅𝐿(𝐿32) + 𝑅𝑟𝑋𝐿
2𝑋𝑚2 + 𝑅𝑠𝑋𝐿
2(𝐿12) (3.12)
𝐶3 = − [2𝑅𝐿𝑏(𝐿32) + 𝑅𝑟𝑋𝐿
2𝑏𝑋𝑚2 + 2𝑅𝑠𝑋𝑙
2𝑏(𝐿12)] (3.13)
𝐶2 = 𝑅𝐿2𝑅𝑟
′𝑋𝑚2 + 𝑅𝐿
2𝑅𝑠(𝐿12) + 𝑅𝐿𝑅𝑟
′ 2(𝐿22) + 2𝑅𝐿𝑅𝑟
′𝑅𝑠𝑋𝑚2 + 𝑅𝐿𝑅𝑠
2(𝐿12) +
𝑅𝐿𝑏2(𝐿3
2) + 𝑅𝑠𝑋𝐿2𝑅𝑟
2′ + 𝑅𝑠𝑋𝐿2𝑏2(𝐿1
2) (3.14)
𝐶1 = −[𝑅𝐿2𝑅𝑟
′𝑋𝑚2 𝑏 + 2𝑅𝐿
2𝑅𝑠𝑏(𝐿12) + 2𝑅𝑟
′𝑅𝐿𝑅𝑠𝑏(𝑋𝑚2 ) + 2𝑅𝐿𝑅𝑠
2𝑏(𝐿12) (3.15)
𝐶0 = 𝑅𝐿2𝑅𝑟
′ 2𝑅𝑠 + 𝑅𝐿2𝑅𝑠𝑏
2(𝐿12) + 𝑅𝐿𝑅𝑟
′ 2𝑅𝑠2 + 𝑅𝐿𝑅𝑠
2𝑏2(𝐿12) (3.16)
and the coefficients 𝐿𝑛 , n = 1… 3, are described in equations below.
𝐿1 = 𝑋𝑚 + 𝑋𝑟 (3.17)
𝐿2 = 𝑋𝑚 + 𝑋𝑠 (3.18)
𝐿3 = 𝑋𝑚𝑋𝑟′ + 𝑋𝑠(𝑋𝑚 + 𝑋𝑟
′) (3.19)
After obtaining the roots of equation (3.11), the real and positive root corresponds to the frequency. If
there are not any real and positive roots, it means that the generator cannot be excited. By replacing
the frequency in the imaginary part that, after simplification, can be described by equation (3.20), it is
possible to obtain the capacitors reactance.
𝑋𝐶 = −𝐴4𝑎
4+𝐴3𝑎3+𝐴2𝑎
2+𝐴1𝑎1
𝐵2𝑎2+𝐵1𝑎
1+𝐵0 (3.20)
The coefficients from equation (3.20) are given below.
𝐴4 = − 𝑋𝐿𝐿3 (3.21)
𝐴3 = 𝑋𝐿𝑏𝐿3 (3.22)
𝐴2 = 𝑅𝐿𝑅𝑟′𝐿2 + 𝑅𝐿𝑅𝑠𝐿1 + 𝑅𝑠𝑅𝑟
′𝑋𝐿 (3.23)
𝐴1 = −𝑅𝐿𝑅𝑠𝑏𝐿1 (3.24)
Page 50
30
𝐵2 = 𝑋𝐿𝐿1 + 𝐿3 (3.25)
𝐵1 = −(𝑋𝐿𝑏𝐿1 + 𝑏𝐿3) (3.26)
𝐵0 = −𝑅𝑟′(𝑅𝐿 + 𝑅𝑠) (3.27)
Finally, knowing both the value of frequency, 𝑎, and capacitors reactance, 𝑋𝐶, it is possible to determine
the minimum capacitance required to keep the generator excited for a specific load and intended
rotational speed, shown in equation (3.28).
𝐶𝑚𝑖𝑛(𝑁𝑟 , 𝑅𝐿 , 𝐿𝐿 , 𝑅𝑠, 𝐿𝑠, 𝐿𝑚, 𝑅𝑟 , 𝐿𝑟) =1
2𝜋∙𝑎∙50∙𝑋𝐶 (3.28)
3.2.2. Computational Model
The analytical model developed in 3.2.1 provides the tools to compute and analyse the range of
capacitances required for a specific induction machine. Here, with the purpose of performing a dynamic
analysis of the SEIG system, in terms of efficiencies, voltage, current and power of the induction
generator, a computational model is developed. By simulating the intended system, the results obtained,
being close or not to reality, represent a first idea and view of what to expect from the application of a
specific SEIG.
To develop the simulation model, the software chosen was a MATLAB tool, Simulink that provides a big
library, with a wide variability of useful tools in Electrical Systems - Power Systems area. The simplified
scheme of the simulations model is the circuit of figures 3.5 and 1.2.
In the simulation model, unlike the proposed system figure 3.5, there is a three-phase power source. As
already stated, in a real induction machine working as a SEIG, when achieves a considerate rotational
speed, a small voltage appears in the stators terminals due to residual magnetization from previous
works, which allows the capacitors to produce reactive power, exciting the machine. However, when
simulating a SEIG work, that small voltage does not exist, it behaves like ideal work conditions.
Therefore, in order to be able to work isolated from the grid, in the beginning of each simulation test, the
machine needs to be connected to the grid, an only after a very short period it is switched to the load
and capacitors bank, isolating the system.
Page 51
31
3.3. Experimental Setup
In order to analyse the behaviour of a SEIG, an experimental system was assembled in Electrical
Machines Laboratory. The objective of these experimental tests was to analyse the performance of a
SEIG, efficiency curve of the generator and required capacitors bank, as well as the bank influence on
the system.
The scheme of the experimental system is represented in figure 3.9. In the assembled system, the prime
mover chosen was a DC motor, connected to the SEIG currently in study.
Figure 3.9 – Scheme of experimental installation used for the tests.
An important step is to check the parameters of both machines, to understand the limit values for both
of them. Therefore, in table 3.1 are expressed the machines’ nameplate information. In addition, the
datasheet of the Induction Machine can be visualized in appendix B.
Table 3.1 – Nameplate Information of both machines.
3 phase Induction Machine Siemens DC Machine Siemens
Grid Frequency 50 Hz Armature Voltage 230 V
Nominal Power 550 W Armature Current 4.4 A
Nominal Phase Current 1.6/2.8 A (Υ/Δ) Nominal Speed 1500 rpm
Nominal Phase Voltage 230/400 V (Υ/Δ) Nominal Power 1 kW
Power Factor 0.74 Field Voltage 220 V
Nominal Speed 910 rpm Field Current 0.55 A
Page 52
32
The assembled workbench is shown in the figure 3.10, below. In the legend of the figure are described
all the materials used in the experimental tests. The detailed information about some equipment can be
seen in Appendix A.
Figure 3.10 – Assemble Workbench for experimental tests: a) 1 - resistive load, 2 – digital multimeter
and 3 – DC power source; b) 4 – Induction generator and 5 – DC motor; c) 6 – auto-transformer; d) 7
– digital multimeter FLUKE, 8 – digital tachometer, 9 – bank of capacitors, 10 – digital multimeter, 11 –
digital amperimeter and 12 – Full-bridge diode rectifier.
The load chosen for the experiments was purely resistive. In table 3.2 are expressed both load and
capacitance ranges used.
12
Page 53
33
Table 3.2 – Tested ranges of the resistive load and capacitances.
In this experimental work, two different machines were used, as shown in table 3.1 and in figure 3.11 –
b) 4 and 5. Before starting the experimental tests, an important step was the characterization of both
machines.
3.3.1. Induction Machine – characteristic parameters determination
As shown in the equivalent circuit, in figure 3.3, an induction machine is characterized by a set of
parameters that allow simulating its operation.
Therefore, in order to simulate the proposed system, an estimation of the parameters needs to be done.
One way to make the estimation is by experimental tests with the machine, by testing the machine in
two different conditions, no load and blocked rotor test [18].
Before starting the tests, with recourse to a digital multimeter, connecting it between two phases of the
machine, it was possible to estimate the stator resistance value.
𝑅𝑒 = 2𝑅𝑠 = 34.9 Ω ; 𝑅𝑠 = 17.45 Ω
For the tests executed in this section, a small set of materials was required: an autotransformer, a digital
multimeter FLUKE and digital tachometer.
No load Tests
The equivalent circuit for the no load case is expressed in figure 3.11. With these tests, an information
about the exciting current and no load losses was given.
Figure 3.11 – No load equivalent circuit of induction machine [30].
Minimum Maximum
Load – RL (Ω) 40 600
Bank – C (µF) 5 175
Page 54
34
As shows figure 3.11, the no load equivalent circuit considers only the stator and magnetizing
parameters. Performing the tests with the induction motor applying different voltages, measuring the
phase current (𝐼0), the active power (𝑃0), the apparent power (𝑆0), power factor and rotational speed (𝑁),
the obtained results are below, in table 3.3.
Table 3.3 – Induction motor test results for no load case.
Using the equations below, with the set of results #1 of table 3.3, a set of calculations can be done, in
order to determine the magnetizing resistance (𝑅𝐶) and reactance (𝑋𝑚) parameters.
𝑄0 = 𝑆0 sin 𝜑 (3.29)
𝐺𝑚 =
𝑃0
𝑉02 =
1
𝑅𝐶
𝐵𝑚 =𝑄0
𝑉02 =
1
𝑋𝑚
(3.30)
Test Phase V0 [V] I0 [A] P0 [W] S0 [VA] cos(ϕ) N [rpm]
#1 L1
L2
L3
Total
228
225
230
-
1.46
1.44
1.47
-
72
68
65
199
333
324
338
995
0.220
0.185
0.193
0.199
996
#2 L1
L2
L3
Total
200
198
202
-
1.13
1.12
1.14
-
49
40
43
133
228
222
230
692
0.216
0.181
0.189
0.194
991
#3 L1
L2
L3
Total
161.3
158.8
163.1
-
0.83
0.82
0.85
-
31
22
27
74
134
130
139
403
0.226
0.178
0.198
0.197
984
#4 L1
L2
L3
Total
119.3
118.5
120.6
-
0.58
0.58
0.58
-
15
14
16
45
69
69
70
208
0.231
0.200
0.217
0.218
976
#5 L1
L2
L3
Total
80.7
80.2
82.1
-
0.38
0.39
0.39
-
8
7
9
25
30
31
32
93
0.273
0.239
0.283
0.263
950
Page 55
35
Blocked Rotor Tests
Inhibiting the rotor from rotating, the objective was to increase the phase voltage until the current
achieves the nominal value, 𝐼𝑐𝑐 = 1.6 A. When that point was achieved, the measures taken were the
same as for no load. The induction machine’s equivalent circuit for this case can be simplified as shows
figure 3.12, where the magnetization branch of the circuit is neglected and also the slip is considered
𝑠 = 1, due to the speed being 𝑁 = 0 rpm.
Figure 3.12 – Equivalent circuit of an induction machine for blocked rotor case, adapted from [30].
Table 3.4 contains the results of the experimental tests with the blocked rotor.
Table 3.4 – Induction motor blocked rotor test results.
Phase Vcc [V] Icc [A] Pcc [W] Scc [VA] cos(ϕ)
L1
L2
L3
72
71.4
73.7
1.52
1.54
1.58
82
79
86
109
109
116
0.753
0.727
0.745
Total - - 247 334 0.743
Using the equations (3.31 – 3.32) below, with the results of table 3.4, a set of calculations can be done,
in order to determine the stator and rotor reactances (𝑋𝑠 and 𝑋𝑟′), as well as the rotor resistance (𝑅𝑟
′).
𝑄𝑐𝑐 = 𝑆𝑐𝑐 sin𝜑 (3.31)
𝑅 = 𝑅𝑠 + 𝑅𝑟
′ =𝑃𝑐𝑐
𝐼𝑐𝑐2
𝑋 = 𝑋𝑟′ + 𝑋𝑠 =
𝑄𝑐𝑐
𝐼𝑐𝑐2 𝑎𝑛𝑑 𝑋𝑟
′ ≅ 𝑋𝑠 (3.32)
Page 56
36
Finally, after performing the experimental tests, in tables 3.3 and 3.4, and using those results in
equations above (3.29 to 3.32), the parameters of the induction machine were estimated.
Table 3.5 – Characteristic parameters of the induction machine in study.
For simplification reasons, the resistance 𝑅𝑐 was neglected further in this work, both for analytical and
computational simulations of the induction machine. Therefore, the magnetization branch was simply
given by the magnetizing reactance, 𝑋𝑚.
3.3.2. DC Motor characteristic equations and its electric efficiency
For the DC machine, its equivalent circuit for steady state condition is simpler to estimate. As shows
figure 3.13, both armature and field circuits, in steady state operation, can be described by resistances.
The values of these resistances can be determined, using a digital multimeter – function ohmmeter – in
the terminals of both circuits. Below, are the obtained values for both resistances.
Armature: 𝑅𝑎 = 6 Ω; Field: 𝑅𝑓 = 318,7 Ω
Figure 3.13 – DC Motor’s equivalent circuit, separate excitation case [31].
In this work, the purpose of the DC machine was to work as a prime mover, in order to help analyse the
behaviour of the induction generator in study. Therefore, an analysis of the DC machine was important.
First, get the nameplate information that was already expressed in table 3.1. Then, the important was
to analyse the DC motor [31], its efficiency, and so, its losses. This allowed, further in the analysis of the
Parameter 𝑹𝒔(Ω) 𝑳𝒔 =𝑿𝒔
𝟐𝝅𝒇 [mH] 𝑹𝒄(Ω) 𝑳𝒎 =
𝑿𝒎
𝟐𝝅𝒇 [H] 𝑹𝒓
′(Ω) 𝑳𝒓 =𝑿𝒓
′
𝟐𝝅𝒇 [mH]
Value 17.45 49.40 783.68 0.51 16.96 49.40
Page 57
37
experimental results, to separate DC motor efficiency and induction generator efficiency.
Considering the case of separate excitation, shown in figure 3.13, the DC Motor steady state conditions
can be expressed by the set of equations (3.33) and (3.34).
𝐼𝑓 = 𝑈𝑓
𝑅𝑓
𝑈𝑎 = 𝑅𝑎 ∙ 𝐼𝑎 + 𝑘∅∅𝑢(𝐼𝑓) ∙ 𝜔𝑟
𝑘∅∅𝑢(𝐼𝑓) ∙ 𝐼𝑎 = 𝑇𝑐
(3.33)
𝑇 = −(𝑘∅∅𝑢)
2
𝑅𝑎∙ 𝜔𝑟 +
𝑈𝑎
𝑅𝑎∙ 𝑘∅∅𝑢 (3.34)
The power flow in the DC Motor, in a more simplified form, can be expressed by equations (3.35), (3.36)
and (3.37).
𝑃𝐼𝑁 = 𝑈𝑎 ∙ 𝐼𝑎 + 𝑈𝑓 ∙ 𝐼𝑓 (3.35)
𝑃𝑙𝑜𝑠𝑠 = 𝑅𝑎 ∙ 𝐼𝑎2 + 𝑅𝑓 ∙ 𝐼𝑓
2 + 𝑃𝑟𝑜𝑡_𝑙𝑜𝑠𝑠 (3.36)
𝑃𝑂𝑈𝑇 = 𝑃𝐼𝑁 − 𝑃𝑙𝑜𝑠𝑠 (3.37)
In equation (3.36), the losses by Joule effect in 𝑅𝑎 and 𝑅𝑓 can easily be determined because they only
depend on the values of armature and field currents, measured during the tests. On the other way, the
rotational losses 𝑃𝑟𝑜𝑡_𝑙𝑜𝑠𝑠 are not that easy to determine. For this experiment, an approximate estimation
was made by testing the DC Motor for no load conditions. In no load conditions, all of the power
measured in armature and field are losses. Therefore, by separating the Joule losses from the rotational
losses (mechanical and magnetic), an estimation of the last ones can be achieved.
Therefore, for no load condition the resultant equation (3.38) comes from joining equations (3.35) and
(3.36).
𝑃𝑟𝑜𝑡_𝑙𝑜𝑠𝑠 = 𝑃𝐼𝑁 − (𝑅𝑎 ∙ 𝐼𝑎2 + 𝑅𝑓 ∙ 𝐼𝑓
2) (3.38)
Figure 3.14 shows the experimental results obtained. For further tests, the rotational losses curve was
approximated to a quadratic function, also in figure 3.14. This estimation, made for no load conditions,
is going to be considered constant, for different values of load.
Page 58
38
Figure 3.14 – Curve of Rotational Losses of the DC Motor.
3.4. Model Application: Analytical and Computational Simulation
In this section, an application of the simulation models is performed. The induction machine
characteristic parameters, already estimated, are expressed in table 3.5.
3.4.1. Analytical analysis – capacitance determination
The analytical model, developed in 3.2.1, can be used for any induction machine and for any resistive-
inductive (RL) load. Using the simplified equations (3.11), (3.20) and (3.28), an estimation of the required
capacitances can be achieved. From equation (3.11), an estimation of the generator frequency is
obtained, then to be applied in equation (3.20), to obtain the reactance of the bank. Finally, with equation
(3.28), using the estimated values of capacitor reactance and generator frequency, the capacitance
value is determined. In this case, the model was applied to the induction machine in study, characterized
in tables 3.1 and 3.5.
To facilitate the analysis of the model, the calculations were performed with recourse to MATLAB. The
tests were carried out by stablishing a load value and then analyse the variation of the capacitance
needed to keep the generator excited as function of the rotational speed. Two types of load were used,
resistive (R) and resistive-inductive (RL) loads.
Prot_loss = 6E-06N2 + 0,0099N + 8,5082R² = 0,9996
0
5
10
15
20
25
30
35
40
0 200 400 600 800 1000 1200 1400 1600 1800
Pro
t_lo
ss
(W)
N (rpm)
Page 59
39
Load – R
Stablishing the load resistance, in figure 3.15 is described the variation of capacitance with rotational
speed of the machine. The speed values chosen were between 500 rpm and 1500 rpm. The range of
loads was between 40 Ω and 600 Ω.
Figure 3.15 – Description of the variation of bank’s capacitances with rotational speed, for specific
resistive loads.
Analysing figure 3.15, for a specific load, the capacitance value applied influenced highly the rotational
speed of the system. The capacitance required to keep the generator excited increased as the speed
decreased, so lower the speed higher the capacitance needed. The same occurred for the load variation,
where lower the resistance, higher the capacitance required, for the same rotational speed. For high
values of load, as shown above, for 300 Ω ≤ 𝑅𝐿 ≤ 600 Ω, the capacitance variation curve was very
similar, and so, the variation of capacitance with speed presented more or less the same results for the
different loads applied.
Another important aspect is that when the load resistance was lower, as shown in 𝑅𝐿 = 100 Ω and
𝑅𝐿 = 85 Ω, the range values of rotational speed in which the generator could operate decreased more
and more. For 𝑅𝐿 = 100 Ω, the range of rotational speeds available was 𝑁 = [600 1300] rpm, and for
𝑅𝐿 = 85 Ω was 𝑁 = [700 950] rpm. For 𝑅𝐿 < 85 Ω, according to the analytical results, it should not be
possible to excite the generator.
In table 3.6 is expressed a more detailed description of the results obtained in the graphic of figure 3.15.
For each load are expressed the capacitance required and the resultant stator frequency, for a specific
rotational speed.
500 600 700 800 900 1000 1100 1200 1300 1400 15000
50
100
150
200
250
300
350
400
N (rpm)
C (
µF
)
R = 600
R = 300
R = 150
R = 100
R = 85
Page 60
40
Table 3.6 – Capacitance determination data results of figure 3.15.
RL=600 Ω RL=300 Ω RL=150 Ω RL=100 Ω RL=85 Ω
N (rpm) C (µF) f (Hz) C (µF) f (Hz) C (µF) f (Hz) C (µF) f (Hz) C (µF) f (Hz)
1500 6.7 72.1 8.8 69.7 16.7 64.7 - - -
1400 8.3 67.3 10.6 65.1 19.4 60.5 - - -
1300 10.4 62.4 13.1 60.4 23.2 56.2 55.1 50.0 -
1200 13.2 57.5 16.6 55.6 28.3 51.9 58.6 47.0 -
1100 17.2 52.6 21.4 50.9 35.5 47.5 67.8 43.5 -
1000 23.0 47.7 28.5 46.2 45.9 43.1 82.9 39.7 -
950 27.0 45.3 33.3 43.8 53.0 40.9 93.6 37.7 161.2 34.7
900 31.9 42.8 39.2 41.4 61.9 38.7 107.3 35.7 172.1 33.2
800 46.2 37.9 56.5 36.6 87.8 34.2 148.3 31.6 227.7 29.6
700 70.9 32.9 86.5 31.8 133.2 26.4 224.1 27.3 379.1 25.1
600 117.8 27.9 144.2 26.9 223.7 25.0 398.5 22.7 -
500 222.3 22.8 275.9 21.9 453.2 20.0 - - -
Load - RL
Here, a brief analysis of load RL is performed and compared to the previous numerical results for loads
R. Stablishing both resistance and inductance values, an analysis of the variation of the required
capacitance with rotational speed is performed. In figure 3.16, for a constant resistive load 𝑅𝐿 = 230 Ω,
are expressed the variation curves of capacitance as function of rotational speed, for different values of
load inductance (𝐿𝐿 [H]).
As shows figure 3.16, the increase of the load inductance, represented a both an increase of the
capacitance required and a huge decrease on the range of rotational speeds in the operating region.
For this specific resistive load, the maximum possible inductive reactance was 𝐿𝐿 = 0.14 H. The
frequency value obtained for this limit was 21 Hz.
By changing the resistive load, other simulations were performed, and the same variation behaviour
was obtained. As smaller the resistance value, shorter was the working range of inductances. Therefore,
the added inductive load had a considerable influence on the capacitances required and the range of
rotational speeds in the operating region of the generator.
This behaviour was already expected, as when this type of load is applied, the capacitance required
would need to be sufficient for both the machine and load. Therefore, when compared to the resistive
Page 61
41
load, it presents more limitations.
Figure 3.16 – Variation of the Capacitance as function of rotational speed, for different values of
inductive load – for a constant 𝑅𝐿 = 230 Ω.
For the different values of resistive load, there was a maximum limit reactance associated. Below, in
table 3.7, are expressed the simulated resistance values and its maximum possible inductance. Also, in
the table are registered the frequency values and rotational speeds for each case.
Table 3.7 – Limit reactance value associated to each different resistive value.
RL (Ω) 85 100 120 150 170 200 230 300 450 600
LL (H) 0.009 0.045 0.080 0.092 0.105 0.124 0.140 0.172 0.223 0.271
f (Hz) 27.3 25 25 23.1 23.2 21.5 21.4 19.9 20 19.5
N (rpm) 750 700 700 650 650 600 600 550 550 550
As already said, the limit reactance decreased with the decrease of the resistance value. For the
minimum resistance value possible (as shown in the numerical analysis for load - R), 85 Ω, the maximum
possible inductance was 𝐿𝐿 = 9 mH.
Therefore, from the perspective of capacitance and rotational speed ranges of the working region of the
induction generator, as it is observed above, in figure 3.16, when applying a load of RL type, the
inductance value presented a high influence on the working ranges. Therefore, if this type of load is
applied to a SEIG system, it should be done a thorough analysis of the system, in order to fully
500 600 700 800 900 1000 1100 1200 1300 1400 15000
100
200
300
400
500
600
700
N(rpm)
C(µ
F)
L = 3 mH
L = 65 mH
L = 80 mH
L = 95 mH
L = 110 mH
L = 125 mH
L = 140 mH
Page 62
42
understand its limits as function of the load.
Further, in this work, the RL load was not used, being both computational and experimental analysis
performed for a purely resistive load.
3.4.2. Dynamic analysis - computational model
The Simulink model developed in section 3.2.2 was applied here for the induction machine characterized
in tables 3.1 and 3.5. The results of the analytical model application were used as a base for the
simulation ranges of capacitances.
This simulation had two main objectives. First, the analysis of the required capacitances, as already
performed in the analytical model. Second, the analysis of the induction generator’s efficiency variations
with load, capacitance and rotational speeds proposed.
In the analytical model, the simulation was performed for both resistive and resistive-inductive loads.
Here however, the main purpose was to execute analysis of capacitance ranges and efficiency, in order
to compare with analytical results and have an estimation of what to expect from the experimental
analysis of the SEIG system. Therefore, the load chosen was purely resistive, with the same load range
chosen for both analytical simulation and experimental work, between 40 Ω and 600 Ω.
The flowchart of figure 3.17 shows the sequence in which the tests were performed, for each load value.
Figure 3.17 – Illustrative scheme of the tests sequence of the simulation.
Page 63
43
As shows the flowchart, first, the load value was stablished, and then the prime mover power, 𝑃𝑝𝑚, was
applied. The simulations were performed by determining the minimum capacitance required, and then
increasing its value and running it again, registering all the measured values. The measured values
corresponded to the stator voltages (𝑉𝑠) and currents (𝐼𝑠), the active power (𝑃𝑠), apparent power (𝑆𝑠),
power factor, rotational speed (𝑁) and mechanical power (𝑃𝑚𝑒𝑐).
Required Capacitance Determination
First, just like the results of the analytical model, figure 3.18 shows the variation of the capacitor values
with the intended rotational speed, for each specific load.
According to the results, for rotational speeds 𝑁 = [500 1500] rpm, the range of capacitances required
to make sure that the machine can reach any position in that speed interval, for any load value, it should
be between [7.5 180] µF.
Figure 3.18 – Simulation results of capacitance variation with rotational speed, for constant load.
Efficiency Analysis
As already said, when connected to the grid, the induction generator efficiency is as function of the slip.
In turn, the slip is as function of the synchronism speed, which is related to the grid frequency (𝑓𝑠 =
50 Hz). As an isolated generator, the case is completely different. The stator frequency is not the grid
frequency anymore, and is constantly changing, as function of the rotational speed of the rotor.
Therefore, the efficiency curves will also change significantly.
From the same simulation results of figure 3.18, an analysis of the efficiency curves of the induction
generator can be performed. In each simulated point, both mechanical and electrical power output
400 600 800 1000 1200 1400 1600 18000
20
40
60
80
100
120
140
160
180
N (rpm)
C (
µF
)
R = 600
R = 200
R = 150
R = 120
R = 100
R = 85
Page 64
44
values were registered, so its efficiency can be calculated, simply by using the equation below (3.39).
𝜂𝑒𝑙_𝑠𝑖𝑚 =𝑃𝑠
𝑃𝑚𝑒𝑐 (3.39)
Figure 3.19 shows the results of the efficiency as function of rotational speed. For different loads, the
efficiency curve changes. For most loads, the maximum efficiency variation was small, at most 8% of
efficiency. However, as the load decreased, the operating ranges of rotational speed decreased as well.
This occurrence was more visible for 𝑅𝐿 ≤ 120 Ω, where there was also a big decrease in the
efficiencies. According to the simulation results, with the exception of the particular cases of 𝑅𝐿 ≤ 120 Ω,
the best efficiencies of the generator were registered for the range of rotational speed, 𝑁 =
[1000 1350] rpm, where the efficiencies were usually above 60 %.
Figure 3.19 – Simulation results of SEIG efficiency curves.
Analysing both figures above, for each load, the efficiency curve presented a maximum efficiency point,
and that point as a capacitance and rotational speed associated with it. In the table 3.8, below are
expressed those points for each simulated resistive load.
400 600 800 1000 1200 1400 1600 180010
20
30
40
50
60
70
N (rpm)
Eff
icie
ncy (
%)
R = 600
R = 200
R = 150
R = 120
R = 100
R = 85
Page 65
45
Table 3.8 – Best efficiency points of the SEIG simulated results.
Sim. Test RL (Ω) C (µF) N (rpm) η (%)
#1 600 17,5 1119 68.9
#2 365 17,5 1188 69.9
#3 300 17,5 1151 71.1
#4 265 20 1185 69.2
#5 230 25 1094 69.0
#6 200 25 1145 67.9
#7 170 30 1105 65.3
#8 150 30 1184 63.8
#9 100 60 1083 44.6
#10 90 85 993 33.8
#11 85 115 898 25.6
Figure 3.20 illustrates the content of table 3.8. The required capacitance presented very small variations
for 𝑅𝐿 = [200 600] Ω. However, for 𝑅𝐿 < 200 Ω the capacitances showed a variation almost exponential.
For loads lower than 85 Ω, the excitation of the generator became impossible.
Figure 3.20 – Curve of capacitances that maximize efficiency, as function of the resistive load.
0
15
30
45
60
75
90
105
120
135
0 50 100 150 200 250 300 350 400 450 500 550 600 650
C (
µF
)
RL(Ω)
𝐶 = 495𝑒−0.02𝑅𝐿
, 𝑓𝑜𝑟 𝑅𝐿 ≤ 150 Ω
𝐶 = −6 ∙ 10−7𝑅𝐿3 + 0.0008𝑅𝐿
2 − 0.37𝑅𝐿 + 71.97 , 𝑓𝑜𝑟 𝑅𝐿>150 Ω
Page 66
46
3.4.3. Analysis of Simulations Results
The purpose of the simulation work performed in this section was to estimate the behaviour of a specific
induction machine working as a SEIG. Both models developed in section 3.2 were applied here.
Although the results of the analytical model were also simulated in Simulink, the usefulness of the
analytical model is in the automatic computation of the required capacitance ranges, something that
with the Simulink model it needed to be done by trial and error. Therefore, by using analytical simulation
it allowed to estimate the capacitances, then used in the Simulink simulation.
According with the obtained results, as already said, the capacitance required influences highly the
speed of the generator, and depends on the load applied. For higher values of load resistance, the
capacitance required presented small variations, however, as the resistance decreased the needed
capacitance increased more and more. In addition, there is an expected limit for the resistive load of
𝑅𝐿 = 85 Ω, in which for lower values the generator cannot be kept excited. Both simulations obtained
this load limit. However, when analysing the capacitance ranges, for each resistive load, they presented
considerable differences. In the analytical case, as the rotational speed decreased, at some point, the
capacitance started to increase rapidly, achieving much greater values than for the Simulink case.
In the analytical results of table 3.6, the frequency values for each case were included, to enhance and
verify the constant variation of its value, proportional to the rotational speed, and also dependent on the
capacitance and load values applied to the SEIG. This high dependence proved the difficulty in
frequency control, already stated in the beginning of this chapter.
As for the resistive-inductive load, only a brief analytical analysis was performed, in order to see the
influences of the load inductance in the system. As it was expected, it presented a high influence the
capacitance needed. When inductive load was added to the system, the bank started to provide reactive
power to the load as well, and so, it needed to be sufficient to satisfy both load and machine. Therefore,
in case this load is applied to the system, a careful analysis must be performed, in order to fully
characterize the bank and the system’s limitations.
The simulation efficiency results (figure 3.20), as already stated, for most of the load values used, the
best efficiencies of the generator were registered for the range of rotational speed, 𝑁 =
[1000 1350] rpm, with values usually above 60 %. Different loads applied presented different maximum
efficiency values. For most loads, the efficiency curve presented small variations, however, for 𝑅𝐿 ≤
120 Ω, the performance got worse. In addition, for lower values of load, the ranges of rotational speed
in the working region decreased significantly.
Concluding, this two developed simulation methods presented some differences of results, in some
cases considerable differences, but the general behaviour of the SEIG was equal on both simulations,
and so, even before the experimental application, the simulation performed here shows that the work of
a SEIG presents considerable limitations, and is highly influenced by the bank capacitance and load
values.
Page 67
47
3.5. Experimental Work
The experimental setup assembled, as shown before in figures 3.9 and 3.10, consisted in a system
composed by a DC motor connected to a SEIG. Herein, an experimental analysis of the required range
for the bank of capacitors was performed, as well as an estimation of the efficiency curve of the generator
for different loads, as function of the rotational speed.
3.5.1. Tests and Results
The flowchart of figure 3.21 shows the followed sequence of the tests performed. For each load value,
first, using a DC power source, a constant current 𝐼𝑓 and voltage 𝑈𝑓 is applied in the DC motor field
circuit. Then, with recourse to a three-phase auto-transformer and a diode rectifier, apply current to the
armature of the DC motor to a value around its nominal value, register the armature voltage 𝑈𝑎 and
measure the resultant rotational speed 𝜔𝑚. Determine the minimum value of capacitance able to keep
the generator excited. Register all the measured values of induction generator (stator voltage - 𝑉𝑠, stator
current - 𝐼𝑠, active power - 𝑃𝑠, reactive power - 𝑄𝑠 and power factor - cos 𝜑). Then, increasing the
capacitance, repeat and register.
In the beginning of the experimental work, an important occurrence observed was the inability of the
system to be excited with a load applied. Every time the generator started working, during the excitation
phenomenon, the generator had to be at no load condition, an only after the generator was excited, the
switch changed position, applying load to the system. This was the reason of the presence of a switch
in the scheme of the experimental installation, in figure 3.9.
Figure 3.21 – Illustration of the sequence of the tests performed.
Page 68
48
The tests were done for the range of loads and capacitances expressed in table 3.2. The estimation of
the efficiency is described by equations (3.40) and (3.41). In equation (3.41), 𝑃𝑠 is the active power of
the generator, and 𝑃𝑚𝑒𝑐 is the mechanical power, and it is given by the total power in the dc motor minus
all the losses associated with it, as it was described in equations (3.35-3.38), in section 3.3.2.
𝜂𝑒𝑙 = 𝑃𝑠
𝑃𝑚𝑒𝑐 (3.40)
𝑃𝑠 = 3 ∙ 𝑉𝑠 ∙ 𝐼𝑠 ∙ cos𝜑
𝑃𝑚𝑒𝑐 = 𝑃𝐼𝑁 − 𝑃𝑙𝑜𝑠𝑠
(3.41)
Analysis of Required Capacitance values
Below, in figure 3.22 is represented the capacitance variation with rotational speed. The experimental
results for the capacitances behaviour with rotational speed were in accordance with the simulation
results. The increase in capacitors resulted in a decrease of rotational speed. The first difference that
stand out was the minimum possible load in which the generator could work with. Here, the minimum
load obtained was around 𝑅𝐿 = 120 Ω.
Figure 3.22 – Experimental results of capacitance required as function of rotational speed of
generator, for constant values of load.
Generator Efficiency Curves
From the same set of tests results of figure 3.22, the efficiency curves of the generator were estimated,
using equations 3.40 and 3.41, described above. Figure 3.23 illustrates the obtained results.
0
20
40
60
80
100
120
140
160
180
200
400 650 900 1150 1400 1650
C (
µF
)
N (rpm)
R = 600 Ω
R = 365 Ω
R = 300 Ω
R = 265 Ω
R = 230 Ω
R = 200 Ω
R = 170 Ω
R = 150 Ω
R = 120 Ω
Page 69
49
Figure 3.23 – Graphic of the SEIG efficiency variation as function of rotational speed, for constant
load values.
As the results show, with different loads, for the same rotational speed, the efficiencies were not always
the same. For most of the loads applied (200 Ω ≤ 𝑅𝐿 ≤ 365 Ω), the efficiencies obtained were very
similar, as it is possible to verify, having very small variations (for the exact same rotational speed),
usually less than 5 % of efficiency.
However, for low or high loads (𝑅𝐿 = 600 Ω and 𝑅𝐿 < 200 Ω), the variation of the efficiency presented
considerable differences. For both of them, the efficiencies were lower than for the other loads tested.
Figure 3.24 – SEIG efficiency curve – average of most of the curves expressed in figure 3.23.
10
20
30
40
50
60
70
300 500 700 900 1100 1300 1500 1700
Eff
icie
ncy (
%)
N (rpm)
R = 600 Ω
R = 365 Ω
R = 300 Ω
R = 265 Ω
R = 230 Ω
R = 200 Ω
R = 170 Ω
R = 150 Ω
R = 120 Ω
ηel = -5E-05N2 + 0,1445N - 36,185
0
10
20
30
40
50
60
70
0 200 400 600 800 1000 1200 1400 1600 1800
Eff
icie
ncy (
%)
N(rpm)
Page 70
50
Besides these differences, depending on the load, the interest of this application was to find the
capacitance that allowed achieving the best efficiency of the system. So, with that in mind, one efficiency
curve of generator was estimated, shown in figure 3.24. This curve was estimated by considering an
average efficiency of the curves of figure 3.23. Although, due to presenting high discrepancies, for this
estimation the curves for 𝑅𝐿 = 600 Ω, 𝑅𝐿 = 170 Ω, 𝑅𝐿 = 150 Ω and 𝑅𝐿 = 120 Ω were not taken into
account. This curve was useful for efficiency calculations, further in this work.
The data of all the result analysis performed here can be accessed in appendix C, with tables for each
different applied load value.
Minimum and Ideal Capacitances
Analysing the efficiency curves of the generator, for each load there was a point of maximum efficiency.
In this case, for all the tested loads, the point of maximum efficiency was the minimum capacitance
needed to keep the generator excited. In table 3.9 are described those specific points for each load.
Figure 3.26 illustrates table 3.9.
Table 3.9 – Maximum efficiency points, for each tested load.
For higher values of resistive load, the capacitance needed did not change much as shows figure 3.25.
However, as the load decreased (in this case, when 𝑅𝐿 ≤ 200 Ω), its value started to increase
exponentially, and with it, the maximum speed the generator can achieve decreased the same way. As
can be observed in table 3.9, this caused a limitation of efficiency for those cases, because the maximum
efficiencies of the generator were achieved for speeds of 𝑁 = [1200 1500] rpm, where 𝜂𝑒𝑙 ≥ 60 %, and
so, if the maximum speed decreased, at some point, it was not able to reach the best efficiency zone.
Test Load –RL (Ω) C (µF) Global Efficiency (%) N (rpm)
#1 600 12.5 62.8 1515
#2 365 15.5 63.8 1380
#3 300 15.5 63.0 1438
#4 265 15.5 64.5 1472
#5 230 15.5 61.8 1553
#6 200 19.5 61.2 1440
#7 170 31.4 53.1 1190
#8 150 40.0 52.9 1134
#9 120 62.0 45.0 965
Page 71
51
Figure 3.25 – Graphic of minimum capacitance as function of load.
Therefore, the SEIG application presents some limitations that need to carefully analysed, because
depending on the load applied and rotational speed intended, this induction machine may or may not
be a good fit for that specific application.
0
10
20
30
40
50
60
70
0 100 200 300 400 500 600 700
C (
µF
)
RL(Ω)
𝐶𝑚𝑖𝑛 = 281.59𝑒−0.013𝑅𝐿 , 𝑓𝑜𝑟 𝑅𝐿 ≤ 230
𝑎𝑠𝑑𝑠𝑠𝑠𝐶𝑚𝑖𝑛 = −0.0086𝑅𝐿 + 17.916 , 𝑓𝑜𝑟 𝑅𝐿 > 230
Page 72
52
3.6. Analysis and validation of simulation methods
In this chapter, simulation and experimental work was performed to analyse a Self-Excited Induction
Generator. For simulation, two models were used, an analytical and a Simulink model, representing two
different ways to simulate the SEIG system. Then, an experimental setup was assembled to analyse
the real performance of the SEIG.
Herein, a comparison between experimental and simulation results was performed, analysing both the
results of the required capacitances and of the performance curves of the SEIG.
3.6.1. Comparison of required capacitances
For a chosen load value, 𝑅𝐿 = 230 Ω, an analysis of the required capacitor values obtained, both by
simulation and experimentally, was performed. The results are expressed in table 3.10, below.
In addition, in table 3.10 are expressed the variation in percentage between the simulation and
experimental results.
Table 3.10 – Comparison between obtained capacitance results for a specific load, 𝑅𝐿 = 230 Ω.
Capacitance Values Required – C (µF) Variation (%) : (Csim – Cexp)*100/Cexp
N (rpm) Experimental Simulink Analytical Simulink Analytical
1550 15.60 13.50 9.60 -13.46 -38.46
1315 19.65 18.00 15.10 -8.40 -23.16
1045 31.41 27.25 28.70 -13.24 -8.63
1000 34.90 29.60 32.70 -15.19 -6.30
920 39.30 34.85 41.90 -11.32 6.62
799 59.50 46.30 64.50 -22.18 8.40
635 88.00 76.00 135.30 -13.64 53.75
560 121.20 102.00 208.50 -15.84 72.03
Figure 3.26 illustrates, for the same load value 𝑅𝐿 = 230 Ω, the variation of the required capacitance
with the rotational speed, for both simulation and experimental results.
Page 73
53
Figure 3.26 – Illustration of the comparison between the simulated and experimental results, for a
specific load of 𝑅𝐿 = 230 Ω.
Analysing both figure 3.26 and table 3.10, some conclusions are important to enhance. First, comparing
the Simulink with experimental results, the capacitance values simulated were very similar, although
they were always lower than the experimental, with an error more or less constant, around 13%. Second,
looking at the experimental and analytical results, for 𝑁 ≥ 800 rpm, they presented more or less similar
values, with exception for 𝑁 = 1550 rpm, in which the error variation is about 38%. However, for speeds
𝑁 < 800 rpm, the behaviour got a lot different. Lowering the speed caused an increase of the
capacitance, much faster than in the other two cases, achieving for 𝑁 = 560 rpm, an error of 72 %, as
presented in table 3.10. These differences are probably related to some factors that were not taken into
account in the analytical equations, and that for lower rotational speeds have high influences on the
capacitance values.
For most of the analysed loads, the behaviour was the same as for 𝑅𝐿 = 230 Ω. Although, there was
another difference between simulations and experimental results that is important to enhance. As the
load resistance value decreased, for both computational and analytical simulations, the limit resistance
value in which the generator could be excited is 𝑅𝐿 = 85 Ω. However, the experimental results showed
otherwise, having the limit resistance around the 𝑅𝐿 = 120 Ω. Below, in table 3.11 and figure 3.27, are
expressed both the experimental and simulation results, for the limit load cases.
As show the results of table 3.11, there were different limit load values for experimental and simulation
cases. The areas filled with white, represent the experimental and simulation operating points, while the
areas filled with grey, represent the non-operating points for experimental and simulation results.
400 600 800 1000 1200 1400 1600 18000
50
100
150
200
250
300
N (rpm)
C (
µF
)
Experimental
Analytical
Simulation
Page 74
54
Table 3.11 – Comparison of results for the lower possible loads.
RL = 150 Ω RL = 120 Ω RL = 85 Ω
N
(rpm)
Cexp
(µF)
Csimulink
(µF)
Canalytical
(µF)
Cexp
(µF)
Csimulink
(µF)
Canalytical
(µF)
Cexp
(µF)
Csimulink
(µF)
Canalytical
(µF)
1500 - 21.5 16.7 - - 25.7 - - -
1490 - - 31.7 - - -
1250 - 27.5 25.5 - 36.9 35.5 - - -
1135 40 32 32.6 - 41.7 44.1 - - -
1035 49.9 37.5 41.8 - 47.5 55.2 - - -
965 50.8 42.3 50.7 62 53 66.2 - - 163.3
930 - 112.5
850 53.5 73.2 66 93.9 - 122.5 194.2
750 78.2 68 107.1 104.8 84 136.3 - 154 279.1
700 79 133.2 96.5 169.5 - 190 379.1
650 104.8 92.5 169.9 175.1 114.5 216.8 - - -
555 175.1 135 297.2 - 170 391 - - -
Figure 3.27 – Illustration of the minimum loads able to be handled by the generator, both experimental
and simulation cases.
500 600 700 800 900 1000 1100 1200 1300 1400 15000
50
100
150
200
250
300
350
400
N (rpm)
C (
µF
)
Exp - R = 150
Num - R = 150
Sim - R = 150
Exp - R = 120
Num - R = 120
Sim - R = 120
Num - R = 85
Sim - R = 85
Page 75
55
3.6.2. Comparison of generator performances
In the previous sections, 3.4.2 and 3.5.1, an analysis of the generator efficiencies was performed by
simulation and experimentally. Herein, for a chosen load, 𝑅𝐿 = 230 Ω, a comparison between the
experimental and simulated results was made.
Figure 3.28 shows the variation of the efficiency with rotational speed. The obtained results presented
considerable differences. The simulated efficiency presented a maximum around the 1200 rpm, as the
experimental efficiency presented a maximum around the 1450 rpm. However, thinking more in general,
both curves presented more or less the same area where the efficiencies were above 50 %.
These results were more or less the same for other resistance values, with exception of the lower
resistances mentioned before.
Figure 3.28 – Comparison between experimental and simulation efficiency variations, for 𝑅𝐿 = 230 Ω.
As shows the figure above, besides the rotational speeds associated with the best efficiencies, another
difference between the curves was the maximum efficiency achieved. In simulation, the generator
achieved efficiencies of around 70 %, while in experimental results the maximum efficiency was around
65 %. However, for the load value of figure 3.28, the maximum efficiency obtained experimentally was
around 61 %.
In both simulation and experimental analysis, a curve of ideal capacitance as function of the applied
load was presented, in order to observe the load influence on the bank. Below, in figure 3.29, are plotted
both variation curves. As already concluded and verified above, there are some differences in the
obtained results.
400 600 800 1000 1200 1400 16000
10
20
30
40
50
60
70
N (rpm)
Eff
icie
ncy (
%)
effe
effs
Page 76
56
Figure 3.29 – Capacitance variation as function of load value – simulation and experimental
comparison – that maximizes the efficiency.
0
20
40
60
80
100
120
140
0 100 200 300 400 500 600 700
C (
µF
)
RL (Ω)
Experimental
Simulation
Page 77
57
3.7. Conclusions
The work performed in this chapter had the purpose of analysing the operation and performance of a
SEIG, further to be applied in a PAT system. To estimate and understand the range of the needed bank
of capacitors, two models were developed. First, a theoretical analytical model, allowing the computation
of the range of capacitances, for a given load. Then, for the estimated values, a simulation model was
developed, to analyse the behaviour of the SEIG operation. These two models allowed a complete
estimation of its performance. In order to verify and understand the real behaviour of a SEIG, an
experimental workbench was assembled, for a specific induction machine.
When working connected to the grid, the induction machine presents an efficiency variation as function
of the slip, which is referred to the grid frequency. However, in this case, as showed the results of
sections 3.4 and 3.5, for an isolated system, the efficiency presented different results. The frequency of
the system was not fixed, being constantly changing, as function of rotational speed, capacitance and
load values. Then, in this case, the efficiency variation was analysed as function of rotational speed.
As shown in section 3.6, comparison of experimental and simulation results, the estimated performances
differed considerably from the real ones, in some cases. In capacitance determination, the Simulink
model presented similar results for most applied loads. The analytical model is simpler and easier to
use, when compared to the Simulink model, but showed more variations, then being less reliable.
Therefore, the simulation in Simulink is the more reliable for estimating the range of capacitances.
However, analysing the simulated and experimental performances, there were considerable differences.
Therefore, the use of the developed theoretical and computational simulation models can be a useful
tool for the estimation of the range of capacitances required, and to have a first idea of the expected
performance of the generator. However, an experimental analysis of the SEIG is always advised, to
verify the real values of capacitances required and performance of the generator.
The SEIG, as shown in the experimental work results, presents considerable limitations. Therefore,
before choosing the generator, it is important to perform a full characterisation of the intended
application, the intended load for the system, in order to avoid problems related to the SEIG limitations
observed here.
For the particular application to a hydraulic PAT system, there is an interest in regulating the rotational
speed depending on the hydraulic parameters, head and flow, available at each moment, to increase
the system performances. With this proposed SEIG system, that interest can be accomplished by the
bank of capacitors.
Page 79
59
4. SEIG application to a hydraulic PAT system
In chapter 3, it was done an exhaustive study about the self-excited induction generator, where analytical
and computational models were developed in order to help to predict its behaviour. In addition,
experimental tests were performed to compare and verify the numerical results. All of the SEIG study
was with the purpose of its application to a PAT system.
Here, a pilot installation was assembled in the Hydraulic Laboratory, in order to perform experimental
work to analyse the behaviour of the system, constituted by a SEIG connected to a hydraulic PAT. First,
a set of experimental tests was executed in order to analyse the capacitance - rotational speed - load
relationship, for a steady flow regime. Then, an analysis of the system’s behaviour was performed, for
different working conditions. Finally, an analysis of the transient state imposed to system was executed.
4.1. Experimental Setup
The assembled workbench is presented in Figure 4.1. In the legend are described all the equipment
and materials used in the experimental work. The detailed information of some equipment can be
accessed in appendix A. The used pump was a KSB Etanorm 32-125, and can be accessed at [32].
Figure 4.1 – Photos of the installation assembled in Hydraulic Laboratory: a) 1 – PAT, 2 – induction
generator, 3 – compressed air reservoir, 4 – pressure sensors; b) 5 – digital multimeter FLUKE, 6 –
three-phase wye resistive load; c) 7 – bank of capacitors, 8 – switch, 9 – digital multimeter.
Page 80
60
In figures 4.2 and 4.3 are represented two schemes of the assembled experimental system. In figure
4.2 is expressed the electrical part, with the induction generator, capacitors, switches and load. Figure
4.3 shows the general hydraulic circuit assembled in hydraulic laboratory for the experimental tests. The
system was composed by a compressed air reservoir, a v notch weir, a pump, a set of valves allowing
to change the flow, a flow meter, pressure sensors and the PAT.
Figure 4.2 – Illustrative Scheme of the generating system.
Figure 4.3 – Illustrative scheme of the hydraulic system assembled, adapted from [33].
4.2. Steady state operation: experimental tests and results
Herein, an analysis of the PAT system for a steady state operation was performed. The experimental
procedures were divided into two different types. First, the analysis of the capacitances required to keep
the generator excited by changing the load and the intended rotational speed. Second, with the results
of the first set of tests, specific operation points were chosen (load, capacitance and rotational speed),
stablishing a rotational speed and then the system was tested for different hydraulic conditions, in order
to estimate the global efficiency curves of the system.
Page 81
61
4.2.1. Analysis of the required capacitances and system behaviour
The experimental setup, shown in figures 4.1, 4.2 and 4.3, consisted in a system composed by a closed
hydraulic circuit and a pump running as turbine (PAT) with an induction generator. The objectives of
these first experimental tests, as already stated, were to analyse the system efficiency dependence on
the capacitance value used for the induction generator’s excitation system, and to analyse the relation
between capacitance value, rotational speed and load.
These tests were made using the sequence shown in the flowchart of figure 4.4. For a chosen load, first,
a constant flow regime is applied in the system’s pipeline (usually around 𝑄𝑖 = 4.6 l/s). Then, determine
the minimum value of capacitance able to keep the induction generator excited for the applied load.
Register all the measured values (Hydraulic –𝑄𝑖,𝐻𝑖; Mechanical - 𝑁; Electrical - 𝑉𝑠,𝐼𝑠,𝑆𝑠,𝑃𝑠 and 𝑝𝑓).
Increasing the capacitance value, repeat the sequence. The capacitance was increased until the
rotational speed reached a low value, around 600 rpm.
Figure 4.4 – Illustration of the sequence of the tests performed.
The range of capacitances used is expressed in table 3.2, and the range of resistive loads 𝑅𝐿 ≤ 265 Ω.
The global efficiency of the system was estimated using the measured values, and is expressed by the
equations below (4.1 and 4.2), where 𝑃𝑠 represents the active power of the generator, and 𝑃ℎ𝑦𝑑
represents the available hydraulic power.
𝜂𝑔𝑙𝑜𝑏𝑎𝑙 = 𝑃𝑠
𝑃ℎ𝑦𝑑 (4.1)
𝑃𝑠 = 3 ∙ 𝑉𝑠 ∙ 𝐼𝑠 ∙ cos 𝜑
𝑃ℎ𝑦𝑑 = 𝜌 ∙ 𝑔 ∙ 𝑄𝑖 ∙ 𝐻𝑖
(4.2)
Page 82
62
Below, in figure 4.5, are expressed the obtained results for the capacitance variation with rotational
speed, for each applied load.
Figure 4.5 – Illustrative variation of the capacitance required as function of the rotational speed, for a
steady flow regime.
In this case, different from the SEIG results, the lowest possible resistive load observe was 𝑅𝐿 = 113 Ω.
In figure 4.6, are plotted the same results of figure 4.5, but for the efficiency variation with the
capacitance value.
Figure 4.6 – Global efficiency variations as function of the bank capacitances, for a constant flow
regime.
0
20
40
60
80
100
120
140
160
500 700 900 1100 1300 1500
C (
µF
)
N (rpm)
R = 265 Ω
R = 230 Ω
R = 200 Ω
R = 170 Ω
R = 150 Ω
R = 125 Ω
R = 113 Ω
5
10
15
20
25
30
0 15 30 45 60 75 90 105 120 135 150
Eff
icie
ncy (
%)
C (µF)
R = 265 Ω
R = 230 Ω
R = 200 Ω
R = 170 Ω
R = 150 Ω
R = 125 Ω
R = 113 Ω
Page 83
63
As already stated in the previous experimental work with the SEIG, it is possible to observe that for the
same resistive load, the increase of the capacitance caused a decrease in the rotational speed. As the
load decreased, at some point, the range of rotational speeds in the operating region of the system also
decreased. For instance, for loads 𝑅𝐿 = 125 Ω and 𝑅𝐿 = 113 Ω, the maximum speeds were around 𝑁 =
850 rpm, as observed in figure 4.5.
As for the efficiency, the results showed that the higher efficiency values were related to the lower
capacitances (higher rotational speeds). However, these values described the generator efficiencies
with the big influence of the PAT efficiency related to the specific flow point. Therefore, an important
procedure is to test the system for different hydraulic work conditions, to estimate the global performance
of the system, and later, the PAT efficiency curves.
In table 4.1 and figure 4.7, are expressed and represented the synthetized results of the tests, showing
the best efficiency operating points for each load and respective capacitance. Once again, proving the
results of previous tests, the results regarding the capacitances needed to keep the generator excited
depending on the load showed the same as the results shown in figure 3.26. So, the behaviour of the
system as the load resistance value decreased started to become exponential, making it impossible at
some point, for the available capacitances, and probably for any capacitance, to keep the generator
excited.
Table 4.1 – Synthetization of the Results.
However, there is one difference from the results of the previous work with the SEIG. For the SEIG case,
the minimum capacitances needed were also the best efficiency points for each load. In this case as
observed in figure 4.6, for the higher loads, the minimum and ideal capacitances were different values.
This occurred due to being different systems. In this case, Generator + PAT, the maximum efficiency
points differed, as it can be observed in table 4.1, where they were obtained for rotational speeds around
𝑁 = 1200 rpm, different from results of table 3.9 (only generator).
Test Load –RL (Ω) C (µF) Global Efficiency (%) N (rpm)
#1 265 19.5 24.6 1233
#2 230 19.5 25.1 1276
#3 200 24.8 24.6 1213
#4 170 31.4 23.1 1140
#5 150 40.0 21.1 1136
#6 125 68.4 17.4 854
#7 113 78.2 15.9 857
Page 84
64
Figure 4.7 – Curve of capacitances associated with the maximum global efficiency, as function of the
load values.
4.2.2. Global efficiency curves of the system
After the analysis of the required capacitors and its influence on the system, the next step was to retrieve
the global efficiency curves of the system. Being a hydraulic system, the efficiency values depend highly
on the flow and head values. As already stated and verified before, other efficiency dependences of the
system are the rotational speed of the generator and its load value.
Therefore, as illustrates the flowchart of figure 4.8, for this experimental procedure, the first step was to
choose the load and capacitance values (so, setting the rotational speed of the system). Then, apply a
constant flow (𝑄𝑖 [𝑙/𝑠]) to the circuit. Register all the measured parameter values (Hydraulic –𝑄𝑖,𝐻𝑖;
Mechanical - 𝑁𝑟; Electrical - 𝑉𝑠,𝐼𝑠,𝑆𝑠,𝑃𝑠 and 𝑝𝑓). Then, change the working conditions, and repeat the
same tests, until reaches the limit of the operating region.
Figure 4.8 – Illustration of the sequence of tests performed for performance analysis.
0
10
20
30
40
50
60
70
80
90
50 100 150 200 250 300
C (
µF
)
Load - RL (Ω)
𝐶 = −3 ∙ 10−5𝑅𝐿3 + 0.0198𝑅𝐿
2 − 4.7788𝑅𝐿 + 406.69
Page 85
65
The tests were done for different values of load and capacitance, in order to be able to stablish different
values of rotational speeds, and with this analysing the hydraulic behaviour of the system. In table 4.2
are expressed the different pairs of load/capacitance chosen for this experimental procedures. In figure
4.9 are expressed the obtained efficiency curves.
Table 4.2 – Load and capacitance combinations used to stablish the speed values for the tests below.
As shown in figure 4.9, the best efficiency points of each curve were usually registered for flow values
between 4,5 l/s and 5,0 l/s. The highest efficiencies were achieved for the curve 𝑁 = 1200 rpm, with
values around 27 %. For rotational speeds higher or lower, the efficiency values decreased gradually.
Figure 4.9 – Illustration of the global efficiency curves as function of flow.
N (rpm) RL (Ω) C (µF)
1525 230 15.5
1370 230 18.5
1200 265 19.5
1130 230 24.8
1050 230 31.4
931 230 34.9
813 230 49.9
700 150 78.2
5
10
15
20
25
30
2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,0
Eff
icie
ncy (
%)
Qi (l/s)
N = 1525
N = 1370
N = 1200
N = 1130
N = 1050
N = 931
N = 813
N = 700
Page 86
66
In addition, for flows lower than 3.5 l/s, the system efficiency was never higher than 20 %, and if the flow
went below 2.5 l/s, the system stopped generating energy, a situation to avoid. Therefore, when applying
this system to a WDN, where the flow and pressure have high variations, the overall efficiency may
decrease even more.
Another aspect to enhance is that, as lower as the speed was, more flat was the efficiency curve, and
for higher speeds the shape of the curve was the opposite, increasing the slope. So, lower the speed,
higher was the range of flows of the operating region, like for the curve of 𝑁 = 700 rpm.
All the detailed information about the data results of figure 4.9 is available in appendix D, with tables
correspondent to each different curve observed.
4.2.3. Hydraulic Q-H curve of PAT
From the same data of the previous results of 4.2.2, in figure 4.10 are expressed the hydraulic curves
of head and flow for each rotational speed and for the runaway working condition.
Figure 4.10 – Hydraulic head/flow curves of the PAT.
An important observation of the results of figure 4.10 was the existence of an inferior limit of flow in
which for lower values, the generator stopped being excited and so it stopped generating energy. This
value depended on the rotational speed, which was a reflexion of the capacitance value applied to the
generator. As shows figure 4.10, higher the rotational speed, lower the range of flow in the operating
region. The lowest value observed was about 2.5 l/s, only for 𝑁 = 700 rpm case.
0
2
4
6
8
10
12
14
16
1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5
Head (
mw
c)
Qi (l/s)
N = 1525 N = 1370
N = 1200 N = 1130
N = 1050 N = 930
N = 813 N = 700
Runaway
Page 87
67
4.3. PAT – hydraulic efficiency curves
In sections 3.5 and 4.2 the experiment was focused on two parts. In section 3.5 was performed the study
of the generator itself, and the results provided an estimated efficiency depending on the rotational
speed. Section 4.2 was centred on the overall system, allowing to retrieve the global efficiencies of the
whole system.
Therefore, using the results of the previous sections, here the focus was turned to the PAT efficiencies.
𝜂𝑔𝑙𝑜𝑏𝑎𝑙(𝑁) = 𝜂𝑒𝑙(𝑁) ∙ 𝜂𝑃𝐴𝑇(𝑁) (4.3)
The system efficiencies are related as it shows equation (4.3), with 𝜂𝑔𝑙𝑜𝑏𝑎𝑙 represented in figure 4.9,
and 𝜂𝑒𝑙 shown in figures 3.23 and 3.24. The variations of these two efficiencies shown in their respective
figures were expressed as function of different variables. The electrical efficiency was expressed as a
function of rotational speed, whereas the global efficiency was expressed as function of flow rate, for a
constant rotational speed. For the PAT efficiency, the variation expected is as function of flow rate, being
the factor that influences the most the turbine performance. For the same rotational speed, the electrical
efficiency is the same no matter what is the hydraulic power available.
Therefore, using the 𝜂𝑒𝑙 results of figures 3.23 and 3.24, expressed below in equation (4.4), and the
rotational speed values expressed in table 4.2, it was possible to achieve the 𝜂𝑒𝑙 shown in table 4.3.
𝜂𝑒𝑙(𝑁) = −5 ∙ 10−5 ∙ 𝑁2 + 0,1445 ∙ 𝑁 − 36,185 (4.4)
Table 4.3 – Combinations of load and capacitance used to achieve the speeds for the tests below,
with correspondent 𝜂𝑒𝑙.
N (rpm) RL (Ω) C (µF) ηel (%)
1525 230 15.5 61.03
1370 230 18.5 61.78
1200 265 19.5 59.81
1130 230 24.8 58.17
1050 230 31.4 55.69
931 230 34.9 50.81
813 230 49.9 44.58
Page 88
68
Then, using equation (4.3), the PAT efficiency curves were determined, as shows figure 4.11.
Figure 4.11 – Illustration of the resultant PAT efficiency curves, as function of flow.
The PAT efficiency, just like in global efficiency results, showed its maximum values for flows between
4,5 l/s and 5,0 l/s. However, the maximum efficiencies of the turbine were obtained for rotational speeds
of 𝑁 = 810 rpm and 𝑁 = 930 rpm. For 𝑁 = [1050 1200] rpm the efficiency values were close to the best
efficiencies for the flow ranges mentioned above. Although, for speeds 𝑁 > 1200 rpm, the efficiency
decreased considerably. Therefore, the maximum efficiency areas of the PAT and generator are not the
same, and so, the system is not optimized.
4.4. Analysis of transient state operation
The steady state analysis of the SEIG applied to a PAT system, study material until now, is of great
importance. However, it only considers constant working conditions, without any variations, that are
situations that occur a lot in a real application, in a WDN. Therefore, an experimental procedure was
performed, considering big variations of the working conditions. Different flow variations were analysed,
closing totally or partially the system valve. Herein, an analysis of the transient behaviour was
performed, for a total close of the valve, for the electrical side of the system. In addition, an illustrative
example of the hydraulic transient phenomenon was given.
15
20
25
30
35
40
45
50
2,5 3,0 3,5 4,0 4,5 5,0 5,5
Eff
icie
ncy (
%)
Qi (l/s)
N = 1525
N = 1374
N = 1200
N = 1128
N = 1050
N = 930
N = 813
Page 89
69
This experimental procedure was divided in two parts:
- Open valve: the excitation of the SEIG, beginning with the machine standstill, opening the circuit
valves, for a specific hydraulic parameters, and then analysing the excitation phenomenon, first
for no load, and then switching to the load applied.
- Close valve: the system was working in steady state condition for a specific hydraulic working
conditions. The circuit valve was closed, stopping the flow of water through the PAT, and then
the behaviour of the SEIG was registered.
The study was performed for the different rotational speeds presented in table 4.2. Both hydraulic and
electrical transient phenomenon were registered. The electrical transient was registered with recourse
to an oscilloscope. Herein, an analysis of the transient was performed for a specific case, for rotational
speed 𝑁 = 1560 rpm.
Case – rotational speed 𝑵 = 𝟏𝟓𝟔𝟎 𝐫𝐩𝐦
Figure 4.12 illustrates the pressure variations, correspondent to the hydraulic transient. Two variations
are visible in the figure, which describe both closing and opening the valves.
The experiment started with the valve open (Figure 4.1a, blue valve in the left bottom of the picture), the
system was excited and the water was flowing through the PAT, with a constant flow and constant
values of upstream and downstream pressure (so, constant head drop). Then, the valve was closed
manually (highlighted block A), and as it is possible to observe, there was a big variation of pressure,
that can cause high amount of water leakages, as occurred many times during these experimental tests.
The maximum peak of pressure achieved was about 7.44 bar (77.33 m w.c.). In addition, as it was
suppose, the pressure drop (upstream and downstream) disappeared.
Figure 4.12 – Illustration of the hydraulic transient behaviour, pressure variation – close and open
valve cases – for 𝑁 = 1560 rpm.
0
10
20
30
40
50
60
70
80
90
0 10 20 30 40 50 60 70 80
Pre
ssure
(m
wc)
Time (s)
Pressure UpstreamMeasured (mwc)Pressure DownstreamMeasured (mwc)
A
B
Page 90
70
After stabilise the pressure, at 𝑡𝑖𝑚𝑒 = 60 s, the valve was opened (highlighted block B), stablishing
again the pressure drop. Here, the transient was more soft, did not present big variations, when
compared to the situation of closing the valve.
For the electrical transient, both voltage and current variations were observed and registered, when
opening or closing the valve. Figures 4.13 and 4.14, below, show their evolution in both situations. In
the figures, the voltage signal read by the oscilloscope corresponded to the phase voltage subjected to
an amplitude reduction, using a transformer with a 400 V – 20 V transformation, so 1/20 ratio.
Figure 4.13 – Illustration of the electrical transient phenomenon for 𝑁 = 1560 rpm (𝑅𝐿 = 230 Ω and
𝐶 = 15 µF) – close valve case: a) Full registered form; b) zoomed plot of highlighted block C.
When the valve was closed (Figure 4.12 – A), the water stopped flowing through the PAT, and so the
generator stopped generating energy, as shows highlighted block C of figure 4.13.
As the valve as opened (figure 4.14), two different phenomenon occurred. As the generator started
rotating, before connecting to the resistive load, the excitation needed to be done for no load, only
-5
-3
-1
1
3
5
0 1 2 3 4 5 6 7 8 9 10
Voltage (
V)
| C
urr
ent
(A)
Time (s)
Voltage
Current
-3
-2
-1
0
1
2
3
4,8 5 5,2 5,4
Voltage (
V)
| C
urr
ent
(A)
Time (s)
Voltage
Current
C
a)
b)
Page 91
71
connected to the capacitors. Only when the machine was excited, the system switched to connect the
resistances in parallel with the machine and capacitors.
Therefore, when the valve was opened and the machine started rotating and was magnetized by the
bank (highlighted block D), both voltage and current increased significantly. These initial peaks of
voltage and current are important, because if they achieve values higher than accepted and remain for
a considerable period of time, can cause damages to the machine.
𝐼𝑠𝑝𝑒𝑎𝑘
= 5 A ; 𝑉𝑠𝑝𝑒𝑎𝑘
= 205 V
In this case, the peak voltage values were within the limit, but the peaks of current were very high,
compared to the nominal current of the generator, 𝐼𝑠𝑛𝑜𝑚𝑖𝑛𝑎𝑙 = 1.6 A. In this registered variation, the
excitation procedure only lasted three seconds, so there was no problem. However, if in a real
application this situation of high peaks of current occurs frequently, and for longer periods, there is the
danger of overheating. To prevent these situations of current peaks, and also of voltage peaks, an option
is to excite the machine with a different capacitance value, to reduce their peak amplitude, and only
when the system is switched to the load, the capacitance is stablished in the intended value.
Figure 4.14 – Illustration of the electrical transient phenomenon for 𝑁 = 1560 rpm (𝑅𝐿 = 230 Ω and
𝐶 = 15 µF) – open valve case: a) Full registered form; b) zoomed plot of highlighted block D; c)
zoomed plot of highlighted block E.
-12
-8
-4
0
4
8
12
0 1 2 3 4 5 6 7 8 9 10
Voltage (
V)
| C
urr
ent
(A)
Time (s)
Voltage
CurrentD
E
-15
-10
-5
0
5
10
15
3,1 3,2 3,3 3,4 3,5 3,6
Voltage (
V)
| C
urr
ent
(A)
Time (s)
VoltageCurrent
-8
-6
-4
-2
0
2
4
6
8
6 6,5 7 7,5 8
Voltage (
V)
| C
urr
ent
(A)
Time (s)
VoltageCurrent
a)
b) c)
Page 92
72
Finally, when the machine was excited, the load was added to the system by a switch (highlighted block
E). The voltage and current slowly stablished into their final and steady values. However, this case was
one of the cases in which the stablishing takes more than usual. As can be observed, when switching,
the current and voltage values first decreased a lot, the effected of the load presence, and constantly
increased and decreased, evolving gradually to the steady final value.
Proposed solution for the high peaks of phase current during excitation
When applying an isolated PAT system in a WDN, there is a constant change in the flow and head drop,
and so in the hydraulic available power. The idea studied in this work was to stablish the capacitance of
the bank, depending on the working conditions at each instant, in order to adjust to the rotational speed
that allowed achieving the best efficiencies.
As already shown in the steady state study of the system, there is a minimum limit flow, and if the water
flow goes below that value, the system cannot generate energy. Therefore, every time the flow reaches
that kind of values, the generator will stop generating, and when the flow increases, has to restart the
excitation procedure again.
Related to that, the presence of high peaks of current associated with the excitation of the machine can
become a problem, if at some point, due to the high variability of the flow, the system as to constantly
restart itself, being every time submitted to high currents. If this situation extends for long periods, hours,
the overheating is a problem, and is important to avoid.
A proposed solution, instead of stablishing the ideal capacitor for the PAT work, for specific working
conditions, was changing the capacitor value in excitation moment, in order to reduce the peaks of
current, and only after the generator was excited and connected to the load, the capacitance was
restored to its intended value.
As an example, below is shown the electrical transient experimental results during the opening of the
system valve, for the same conditions as before, for a rotational speed 𝑁 = 920 rpm.
In this case, the initial peak current was lower, 𝐼𝑠𝑝𝑒𝑎𝑘
= 3 A, and even before connecting to the load, the
current peak values decreased instantly to 𝐼𝑠𝑝𝑒𝑎𝑘
= 1.5 A. Also, the transition when the system was
connected to the load presented a more soft evolution (highlighted block G), compared to the previous
case (highlighted block E), stablishing faster to its final value. The capacitance and load values
associated to both cases are expressed in table 4.2. Therefore, for the same load 𝑅𝐿 = 230 Ω, changing
the capacitor value from 15 µF to 35 µF, allowed a significant reduction of current peaks during excitation
of the generator.
Page 93
73
Figure 4.15 – Illustration of the electrical transient phenomenon for 𝑁 = 920 rpm (𝑅𝐿 = 230 Ω and
𝐶 = 35 µF) – open valve case: a) Full registered form; b) zoomed plot of highlighted block F; c)
zoomed plot of highlighted block G.
The same analysis was done using the Simulink model developed in chapter 3. For the same input
mechanical power, increasing the capacitance allowed to control the peak values of the phase current
during the excitation of the generator.
Below, in figure 4.16, are expressed the simulation results for the excitation of the machine for different
capacitor values. In both cases, a) and b), the two different areas were highlighted, showing two switch
variations, at the third and seventh seconds of the simulation. As already said in the introduction of the
Simulink model, in chapter 3, in the beginning of each simulation, the machine was connected to the
grid, and only after three seconds, the system switched and became isolated. So, for 𝑡𝑖𝑚𝑒 = [3 7] s, the
excitation phenomenon was registered, and then, at 𝑡𝑖𝑚𝑒 = 7 s, the load was applied to the system.
As showed the results of the figure 4.16, an increase in the capacitance, besides decreasing the
rotational speed of the generator, also reduces the current values in the excitation period, before
switching the system to the load. For the phase voltage, the same evolution was observed.
-4
-3
-2
-1
0
1
2
3
4
4 5 6 7 8 9 10
Curr
ent
(A)
Time (s)
𝐼𝑠𝑝𝑒𝑎𝑘 = 3 𝐴
G
F
-4
-3
-2
-1
0
1
2
3
4
4,2 4,4 4,6 4,8 5 5,2
Curr
ent
(A)
Time (s)-2
-1,5
-1
-0,5
0
0,5
1
1,5
2
6 6,2 6,4 6,6 6,8 7
Curr
ent
(A)
Time (s)
a)
c) b)
Page 94
74
Figure 4.16 – Simulink simulation results, of the generator excitation phenomenon, for the same load
and prime mover power input, with different capacitance values.
Therefore, this proposed solution can be a viable way to reduce the peaks of current and voltage during
excitation. In future work, a more detailed study of this should be done, in order to analyse if this solution
it is worth for different cases.
3 4 5 6 7 8 9 10-4
-3
-2
-1
0
1
2
3
4
Time (s)
Cu
rre
nt
(A)
3 4 5 6 7 8 9 10-4
-3
-2
-1
0
1
2
3
4
Time (s)
Cu
rre
nt
(A)
a)
𝐶𝑒𝑥𝑐 = 15 µF
𝑁 = 1460 rpm
𝑅𝐿 = 230 Ω
𝐼𝑠𝑝𝑒𝑎𝑘
= 3.2 A
b)
𝐶𝑒𝑥𝑐 = 40 µF
𝑁 = 960 rpm
𝑅𝐿 = 230 Ω
𝐼𝑠𝑝𝑒𝑎𝑘
= 2.1 A
Page 95
75
4.5. Conclusions
In this chapter, an experimental work was performed to analyse the Self-Excited Induction Generator
applied to a PAT system in a hydraulic closed circuit. The main focus was a steady state analysis of the
system, but a brief analysis of the transient state was also performed.
The steady state experimental work, shown in figures 4.9 and 4.10, was important in order to find the
operating conditions that allows achieving the best efficiency, depending on the available working
conditions, flow and head. By using a bank of capacitors as external excitation system of the generator,
when changing its capacitance, the rotational speed changes as well and so, it is possible to adjust the
system to better working conditions, according to the available flow at each moment. The highest value
of efficiency obtained was around 𝜂𝑔𝑙𝑜𝑏𝑎𝑙 = 27 %, a low value, compared to the maximum SEIG electrical
efficiency obtained in the previous section, of about 𝜂𝑒𝑙 ≈ 65 %. This occurred due to the presence of
the PAT, which had a maximum efficiency for rotational speeds around 𝑁 = 1000 rpm, of 𝜂𝑃𝐴𝑇 = 45 %,
different from the generator’s ideal working speeds, and so affecting the performance of the global
system.
The work performed allowed to retrieve and verify the capacitance variation with rotational speed, for
different loads applied to the system. In addition, the characteristic curves of the PAT were determined,
providing the lower limit values of flow, important data information for further application in real case.
For the transient behaviour, the electrical transient in the excitation of the generator presented in some
cases high peaks of current and voltages. In WDS, there is a high variation of the working conditions,
achieving many times flow values lower than the limit of the generating system (Figure 4.10). Therefore,
in the worst scenario, the generator will need to be excited constantly, for long periods, and so it can
become an issue, because of the high peaks obtained during excitation. The solution proposed here,
consisting of increasing the capacitance value during the excitation phenomenon, reducing the
amplitude of the peak values, represents an interesting alternative. After the system is excited and the
load is included, the capacitance can be changed to the value that better fits the working conditions,
and so, it does not affect the performance of the system.
Page 97
77
5. Conclusions and Future Work
5.1. Conclusions
The study of the application of PAT in WDS has increased over the last years. Different methods for
PAT control and regulation have been proposed and analysed. According to recent works, Electrical
and Hydraulic regulation modes are the two designs more studied, both of them applications with the
system connected to the electrical grid.
In this work, a new approach was proposed, by isolating the system from the grid. As main objective,
an analysis of the performance of a SEIG applied to a hydraulic PAT system was performed. Inserted
in project that aims the PAT application in a specific WDN, this work intended to analyse a specific PAT
+ IG system, further to be applied in the real case. The work was divided into two parts: SEIG analysis
and PAT analysis.
Concerning the SEIG study, the chosen form of excitation was by connecting a bank of capacitor across
the generator stator phases, to provide the needed reactive power. The SEIG performance presents
very different results from the grid connected IG application. There is a big difficulty on controlling the
frequency, which is constantly changing, as function of the rotational speed. Therefore, the analysis of
the generator efficiency was performed as function of rotational speed, capacitance and load applied. A
high dependence on the capacitance value was observed. When working with this type of system, in
order to obtain the best performances, a thorough study of the intended load as well as the system
dependence with the bank’s capacitance, and the ranges required for the bank, is of great importance.
In addition, the rotational speed is highly dependent of the capacitor value, and its variation needs to be
well analysed and described.
The application of the SEIG in a hydraulic PAT system, allowed estimating the both global and PAT
efficiency curves, as well as the characteristic curves of the PAT as function of flow. From the global
efficiency obtained, a conclusion to retrieve is that the system is not optimized. The best performance
areas of the SEIG and the PAT are obtained for very different values of speed, so an optimization of the
system is required. In addition, in experimental work, a set of limitations of the system were verified, of
both load and flow conditions. Under variable load conditions, the load value highly influences the
needed capacitance and rotational speed. Therefore, the intended application needs to be well defined
an analysed, in order to correctly choose the generating system, and avoid the occurrence of this
problems.
From a general perspective, the application of this isolated IG + PAT system can be an interesting
solution for combining the pressure reduction in WDS with energy generation. Although it does not
present very promising efficiency ranges, the presence of the bank of capacitors allows, by changing its
value, to change the rotational speed, and adapt it to the working conditions at each moment, in order
to achieve better performances.
Page 98
78
5.2. Future Work
From the work performed here, some problems and limitations were observed. To improve the system
performance and prepare it for the real case application, in the future some important work should be
done. The development of a controlling system that, depending on the working conditions, applies
different capacitance value, in order to achieve higher efficiencies, is an important part of this system
application in a real case. In addition, an optimization of the system, finding a more suitable pair of
SEIG+PAT for the system, allows achieving higher efficiencies, and so, generation of more energy.
So, the future work related to this project can be summarized by the following topics:
Assemble a new hydraulic pilot circuit, to perform more experimental tests with the isolated
PAT;
Develop a controlling electronic system to maximize the global efficiency, according to the
working conditions;
Optimization of the system for a real case study, according to its working specifications, by
selecting a more suitable hydraulic pump and induction generator;
Page 99
79
Bibliography
[1] A. Stepanoff, “Centrifugal and Axial Flow Pumps,” Ed. J. Wiley., 1957.
[2] R. Knapp, “Centrifugal Pump Performance as Affected by Design Features,” Trans. ASME, vol.
63, no. 2, pp. 251–260, 1941.
[3] C. P. Kittredge, “Centrifugal Pumps Used as Hydraulic Turbines,” J. Eng. Power, vol. 83, no. 1,
pp. 74–78, 1961.
[4] L. Araújo, H. Ramos, and S. Coelho, “Pressure control for leakage minimisation in water
distribution,” Water Res. Manag., vol. 20, no. 1, pp. 133–149, 2006.
[5] T. Walsky, W. Bezts, E. Posluzny, B. Withman, and M. Weir, “Modeling leakage reduction
through pressure control,” J. Am. Water Work. Assoc, vol. 98, no. 4, pp. 148–155, 2006.
[6] S. Prescott and B. Ulanicki, “Improved control of pressure reducing valves in water distribution,”
J. Hydraul. Eng., vol. 134, no. 1, pp. 56–65, 2008.
[7] “Manual on Pumps used as Turbines.” [Online]. Available: http://www.nzdl.org. [Accessed March
2017].
[8] I. Sarbu, “A Study of Energy Optimisation of Urban Water Distribution Systems Using Potencial
Elements,” Water, 2016.
[9] A. Carravetta, G. Del Giudice, O. Fecarotta, and H. M. Ramos, “Energy Production in Water
Distribution Networks: A PAT Design Strategy,” Water Resour. Manag., vol. 26, no. 13, pp. 3947–
3959, 2012.
[10] A. Carravetta, G. del Giudice, O. Fecarotta, and H. Ramos, “PAT design strategy for energy
recovery in water distribution networks by electrical regulation,” Energies, vol. 6, no. 1, pp. 411–
424, 2013.
[11] A. Carravetta, G. Del Giudice, O. Fecarotta, and H. Ramos, “Pump as turbine (PAT) design in
water distribution network by system effectiveness,” Water, vol. 5, no. 3, pp. 1211–1225, 2013.
[12] A. Carravetta, O. Fecarotta, G. Del Giudic, and H. Ramos, “Energy recovery in water systems
by PATs: A comparisons among the different installation schemes,” Procedia Eng., vol. 70, pp.
275–284, 2014.
[13] S. Derakhshan and A. Nourbakhsh, “Experimental study of characteristic curves of centrifugal
pumps working as turbines in different specific speeds,” Exp. Therm. Fluid Sci., vol. 32, pp. 800–
807, 2008.
[14] A. Rodrigues, P. Singh, A. Williams, F. Nestmann, and E. Lai, “Hydraulic analysis of a pump as
turbine with CFD and experimental data,” in Proceedings of IMechE Seminar, Computational
Fluid Dynamics for Fluid Machinery, 2003.
Page 100
80
[15] P. Suter, “Representation of pump characteristics for calculation of water hammer,” Sulzer Tech,
no. 4, pp. 45–48, 1966.
[16] A. Carravetta, O. Fecarotta, R. Martino, and L. Antipodi, “PAT efficiency variation with design
parameters,” Procedia Eng., vol. 70, pp. 285–291, 2014.
[17] C. Tricarico et al., “Optimal Water Supply System Managment by Leakage Reduction and Energy
Recovery,” Procedia Eng., vol. 89, pp. 573–580, 2014.
[18] A. E. Fitzgerald, C. Kingsley, and S. D. Umans, Electric Machinery, 6th editio. McGraw-Hill, 2003.
[19] “Three-phase asynchronous motor.” [Online]. Available: ref.
http://tesinasalvatorecoppola.altervista.org/Inglese.html. [Accessed February 2017].
[20] S. McFadyen, “Induction Motor Equivalent Circuit,” 2014. [Online]. Available:
http://myelectrical.com/notes/entryid/251/induction-motor-equivalent-circuit. [Accessed
February 2017].
[21] K. Daware, “Induction Generator Working Theory.” [Online]. Available:
http://www.electricaleasy.com/2014/12/induction-generator-working.html. [Accessed February
2017].
[22] G. K. Singh, “Self-excited induction generator research — a survey,” vol. 69, no. August 2003,
pp. 107–114, 2004.
[23] R. A. Valadão, “Análise do comportamento de um Gerador de Indução Trifásico Autoexcitado
por Capacitores,” Universidade Federal do Rio de Janeiro, 2012.
[24] M. K. Arya, “Steady State Analysis of Self-Excited Induction Generator for Balanced and
Unbalanced Conditions,” Thapar University, Patiala, 2009.
[25] M. E. A. Farrag and G. A. Putrus, “Analysis of the Dynamic Performance of Self-Excited Induction
Generators Employed in Renewable Energy Generation,” Energies, vol. 7, pp. 278–294, 2014.
[26] G. Ofualagba, “Analysis of the Dynamic Characteristics of an Isolated Self-Excited Induction
Generator Driven by a Wind-Turbine,” Int. J. Sci. Eng. Res., vol. 2, no. 2, pp. 1–10, 2012.
[27] A. M. Eltamaly, “New formula to determine the minimum capacitance required for self-excited
induction generator,” in 33rd Power Electronics Specialists Conference, 2002.
[28] A. Kumari, A. G. Thosar, and S. S. Mopari, “Determination of Excitation Capacitance of a Three
Phase Self Excited Induction Generator,” IJAREEIE, vol. 4, no. 5, pp. 4069–4074, 2015.
[29] A. Abbou, M. Barara, A. Ouchatti, M. Akherraz, and H. Mahmoudi, “Capacitance Required
Analysis For Self- Excited Induction Generateur,” vol. 55, no. 3, pp. 382–389, 2013.
[30] J. S. Paiva, Slides da Unidade Curricular - Fundamentos de Energia Eléctrica, Instituto Superior
Técnico, Universidade de Lisboa.
[31] A. Dente, "Máquina de Corrente Contínua", Notas complementares da Unidade Curricular -
Máquinas Eléctricas. IST/DEEC - Energia.
Page 101
81
[32] “Standardised Water Pump: Etanorm.” [Online]. Available:
https://shop.ksb.com/ims_docs/72/727F1E521F557577E10000000AD50628.pdf. [Accessed
April 2017]
[33] M. C. M. Simão, “Hidrodinâmica e desempenho em turbinas de baixa potência : concepção,
modelação e ensaios,” Instituto Superior Técnico - Universidade de Lisboa, Dissertação para
obtenção do Grau Mestre em Engenharia Civil, Lisboa, 2009.
Page 103
A1
Apendix
A. Equipment used in experimental applications
Power Logger Fluke 1735
Tektronix TDS 2001/2012C Oscilloscope
Model: Fluke 1735 three-phase power logger
Memory: 4MB flash memory – 3.5MB for
measuring data
Sample Rate: 10.24 kHz
V-RMS wye resolution: 0.1V
Operating Error: ±0.5% of measured value +10
digit
A-RMS resolution: 0.01A
Operating Error: ±1% of measured value +10
digit
Brand: Tektronix
Model: TDS 2001/2012C
Analog Bandwidth: 100 MHz
Sample Rate: 2 GS/s
Record Length: 2.5k Point
Analog Channels: 2
Page 104
A2
Capacitor banks Esselte Studium
Three-phase bank of resistances Oficel
Brand: ESSELTE STUDIUM
Power: 3.2 kVAr
Frequency: 50 Hz
Maximum capacitance: 3·60 µF
Connections:
Δ - 3·220 V ; 1.2 - 8.4 A
Y - 3·380 V ; 0.7 – 4.9 A
/// - 220V ; 2.1 – 14.5 A
Brand: Oficel
Range: 40 – 600 Ω
Brand: ESSELTE STUDIUM Connections:
Power: 5.1 kVAr Δ - 3·220 V ; 1.9 - 13.4 A
Frequency: 50 Hz Y - 3·380 V ; 1.1 – 7.7 A
Maximum capacitance: 3·115 µF /// - 220V ; 3.3 – 23.1 A
Page 105
A3
B. Datasheet of the Induction Machine used in experimental
and simulation analysis
Page 106
A4
C. Tables with Experimental results of SEIG application
Load resistance – R = 600 Ω
DC MOTOR INDUCTION GENERATOR
Ua (V) Ia (A) Uf (V) If (A) Pin
(W)
Pmec
(W)
N
(rpm)
Us (V) Is (A) Ps (W) Ss
(VA)
fp C (µF) Eff
(%)
200.0 2.24 141.8 0.34 496.2 392.0 1515 232 1.05 246 721 0.346 12.4 62.8
183.8 4.32 139.0 0.33 839.9 659.7 1378 287 1.65 367 1433 0.258 15.6 55.6
173.1 4.20 139.0 0.33 772.9 601.3 1281 266 1.65 308 1311 0.235 18.7 51.2
166.8 4.26 141.6 0.33 757.6 584.2 1215 253 1.69 292 1292 0.226 19.7 50.0
154.7 4.30 141.6 0.33 711.5 538.9 1110 231 1.72 240 1191 0.201 24.8 44.5
139.9 4.20 141.6 0.33 634.4 469.6 985 202 1.69 189 1026 0.184 31.4 40.2
134.3 4.10 141.9 0.33 597.6 438.6 945 190 1.66 169 952 0.177 34.9 38.5
128.0 4.32 141.3 0.33 599.0 431.2 885 180 1.72 147 930 0.156 39.3 34.1
123.0 4.20 149.0 0.37 571.0 401.7 837 166 1.69 127 839 0.155 49.9 31.6
119.7 4.23 149.0 0.37 560.7 390.5 807 160 1.69 114 814 0.15 50.8 29.2
114.5 4.23 149.0 0.37 539.5 368.9 763 146 1.65 104 725 0.146 59.5 28.2
112.0 4.26 149.3 0.37 532.4 360.7 740 143 1.66 95 710 0.134 60.6 26.3
102.0 4.27 149.5 0.37 490.9 320.3 650 125 1.69 57 632 0.09 78.2 17.8
Load resistance – R = 365 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia (A) Uf (V) If (A) Pin (W) Pmec
(W)
N
(rpm)
Us
(V)
Is
(A)
Ps (W) Ss
(VA)
fp C (µF) Eff (%)
181.3 4.15 140.6 0.38 806.2 622.6 1380 228 1.40 397 948 0.42 15.6 63.8
168.5 4.13 140.4 0.38 749.0 570.1 1278 215 1.40 347 889 0.392 18.67 60.9
160.9 4.18 140.6 0.38 725.6 546.3 1204 205 1.41 324 858 0.377 19.65 59.3
148.3 4.18 140.4 0.38 672.7 496.3 1093 188 1.41 274 802 0.341 24.82 55.2
134.8 4.24 140.5 0.38 624.2 447.6 977 171 1.44 231 744 0.309 31.41 51.6
131.7 4.23 141.7 0.38 610.7 434.7 939 165 1.45 214 725 0.295 34.9 49.2
124.1 4.27 141.5 0.38 583.3 406.9 871 155 1.48 187 689 0.267 39.3 46.0
118.9 4.14 141.6 0.38 545.6 376.7 830 144 1.48 163 631 0.260 49.9 43.3
115.3 4.16 141.7 0.38 532.9 363.8 796 139 1.48 149 610 0.243 50.8 41.0
111 4.24 141.2 0.38 523.7 351.3 758 131 1.49 138 580 0.239 59.5 39.3
108.4 4.24 141.6 0.38 512.7 341.0 733 127 1.49 127 566 0.225 60.6 37.2
97.2 4.10 141.6 0.38 451.6 288.6 640 113 1.50 85 499 0.172 78.2 29.4
94.7 4.24 141.7 0.38 454.7 285.2 611 104 1.50 78 470 0.166 88.0 27.4
89.9 4.23 141.8 0.37 433.2 265.4 568 94 1.51 64 423 0.151 104.8 24.1
79.5 4.13 141.5 0.37 381.1 219.7 485 73 1.43 40 306 0.130 157.8 18.2
77.9 4.18 141.8 0.37 378.5 214.9 466 67 1.42 34 285 0.120 175.1 15.8
Page 107
A5
Load resistance – R = 300 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia
(A)
Uf (V) If (A) Pin
(W)
Pmec
(W)
N
(rpm)
Us
(V)
Is (A) Ps (W) Ss
(VA)
fp C
(µF)
Eff (%)
200.0 4.20 141.5 0.39 895.8 705.3 1438 221 1.44 462 962 0.483 15.6 63.1
188.1 4.29 149.0 0.39 865.4 673.6 1332 212 1.46 423 936 0.452 18.7 62.8
174.4 4.24 149.6 0.39 798.4 610.9 1245 198 1.45 377 862 0.436 19.7 61.7
158.7 4.31 149.3 0.38 741.0 555.7 1129 184 1.46 322 802 0.399 24.8 57.9
144.0 4.29 149.6 0.37 673.7 494.2 1003 166 1.47 263 725 0.363 31.4 53.2
138.5 4.25 149.5 0.37 644.5 468.0 961 159 1.49 240 690 0.348 34.9 51.3
130.1 4.29 149.5 0.37 613.6 437.2 890 148 1.47 208 652 0.319 39.3 47.6
126.2 4.40 149.5 0.37 610.6 429.4 856 144 1.50 197 646 0.307 49.9 45.9
121.9 4.37 149.5 0.37 588.0 409.1 820 138 1.51 179 618 0.289 50.8 43.7
117.0 4.35 149.5 0.37 564.3 387.2 780 127 1.50 160 571 0.282 59.5 41.3
114.0 4.35 149.5 0.37 551.2 374.7 754 124 1.49 150 556 0.267 60.6 40.0
103.0 4.27 149.3 0.37 495.1 324.3 662 109 1.50 102 489 0.210 78.2 31.4
Load resistance – R = 265 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia
(A)
Uf (V) If (A) Pin (W) Pmec (W) N
(rpm)
Us (V) Is
(A)
Ps
(W)
Ss
(VA)
fp C (µF) Eff (%)
199.5 4.22 140.5 0.41 899.8 702.7 1472 210 1.45 468 888 0.523 15.6 64.5
184.5 4.29 141.2 0.39 845.7 653.6 1371 196 1.42 421 835 0.504 18.7 64.4
173.8 4.12 141.3 0.39 771.2 589.8 1283 183 1.38 368 756 0.486 19.7 62.4
159.3 4.25 141.3 0.39 731.8 547.5 1155 172 1.4 323 722 0.448 24.8 59.0
144.0 4.27 140.9 0.39 669.3 487.4 1024 156 1.42 273 668 0.41 31.4 56.0
137.8 4.16 141.2 0.39 627.6 452.6 977 149 1.42 246 623 0.395 34.9 54.3
129.9 4.23 141.4 0.39 603.9 427.0 903 141 1.43 219 600 0.364 39.3 51.3
125.0 4.27 141.2 0.38 587.8 410.1 865 135 1.43 199 572 0.351 49.9 48.5
119.1 4.02 141.4 0.38 532.9 368.5 822 127 1.38 175 524 0.334 50.8 47.5
114.3 4.04 141.3 0.38 515.7 351.4 780 118 1.4 159 493 0.325 59.5 45.2
112.3 4.18 141.5 0.38 523.3 352.8 758 117 1.42 155 499 0.308 60.6 43.9
100.2 4.06 141.6 0.38 460.6 298.1 658 104 1.43 110 441 0.249 78.2 36.9
97.3 4.23 141.3 0.38 464.8 295.1 630 98 1.44 100 421 0.236 88.0 33.9
92.6 4.21 141.5 0.38 443.5 275.0 585 89 1.44 84 386 0.217 104.8 30.5
83.6 4.25 141.5 0.38 408.6 239.9 505 71 1.44 55 303 0.18 157.8 22.9
81.0 4.25 141.7 0.38 398.0 229.1 482 65 1.43 46 280 0.16 175.1 20.1
Page 108
A6
Load resistance – R = 230 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia (A) Uf (V) If (A) Pin (W) Pmec
(W)
N
(rpm)
Us (V) Is
(A)
Ps
(W)
Ss
(VA)
fp C (µF) Eff
(%)
210 4.20 141.1 0.40 938.7 743.0 1553 193 1.44 459 826 0.562 15.6 61.8
195 4.30 141.5 0.37 890.9 701.8 1415 186 1.45 432 810 0.535 18.7 61.6
186 4.29 141.5 0.37 849.4 663.7 1315 179 1.43 405 780 0.52 19.7 61.0
166.3 4.21 141.5 0.37 752.5 573.6 1194 157 1.41 325 661 0.495 24.8 56.7
148.6 4.32 141.5 0.37 693.6 512.8 1045 145 1.4 278 612 0.455 31.4 54.2
142.8 4.29 141.5 0.37 665.1 486.5 1004 140 1.39 257 584 0.442 34.9 52.8
132.4 4.19 141.5 0.37 607.1 435.4 920 130 1.36 219 534 0.41 39.3 50.3
128.1 4.21 141.5 0.37 591.7 419.8 880 126 1.37 203 519 0.394 49.9 48.4
123.5 4.22 141.5 0.37 573.5 401.9 842 121 1.37 187 494 0.376 50.8 46.5
118.6 4.25 141.5 0.37 556.4 384.2 799 114 1.38 174 476 0.366 59.5 45.3
114.8 4.22 141.5 0.37 536.8 366.6 770 110 1.37 158 450 0.349 60.6 43.1
101.1 4.06 141.7 0.37 462.9 302.6 665 97 1.35 113 392 0.288 78.2 37.3
98 4.14 141.8 0.37 458.2 294.5 634 92 1.36 102 377 0.273 88.0 34.6
93.4 4.14 141.8 0.37 439.1 276.2 592 84 1.39 87 350 0.249 104.8 31.5
89.8 4.18 141.6 0.37 427.8 263.4 559 78 1.39 75 326 0.232 121.2 28.5
84.2 4.18 141.3 0.37 404.2 240.6 513 67 1.39 58 276 0.206 157.8 24.1
81.6 4.15 141.7 0.37 391.1 229.3 491 62 1.37 49 254 0.192 175.1 21.4
Load resistance – R = 200 Ω
DC MOTOR INDUCTION GENERATOR
Ua (V) Ia
(A)
Uf (V) If
(A)
Pin (W) Pmec (W) N
(rpm)
Us
(V)
Is (A) Ps
(W)
Ss
(VA)
fp C
(µF)
Eff (%)
191.2 4.30 149.3 0.37 877.4 687.6 1440 170 1.49 421 762 0.550 19.7 61.2
174.2 4.38 149.3 0.37 818.2 629.0 1298 157 1.51 369 695 0.531 24.8 58.8
154 4.27 149.3 0.37 712.8 532.6 1121 140 1.42 295 599 0.493 31.4 55.4
148.7 4.35 149.3 0.37 702.1 518.9 1071 137 1.44 281 588 0.479 34.9 54.2
137.9 4.3 149.3 0.37 648.2 469.7 977 127 1.42 243 539 0.448 39.3 51.7
133.8 4.33 149.3 0.37 634.6 455.5 932 125 1.43 226 528 0.432 49.9 49.6
128.5 4.31 149.3 0.37 609.1 432.0 888 119 1.45 209 506 0.415 50.8 48.4
124.2 4.38 149.3 0.37 599.2 419.3 846 112 1.44 194 486 0.402 59.5 46.3
121 4.37 149.3 0.37 584.0 405.3 813 110 1.44 184 474 0.385 60.6 45.4
107.8 4.37 149.3 0.37 526.3 349.7 703 99 1.45 136 422 0.322 78.2 38.9
104.3 4.35 149.3 0.37 508.9 333.9 670 92 1.44 120 393 0.305 88.0 35.9
99.6 4.40 149.3 0.37 493.5 316.6 626 85 1.45 102 369 0.279 104.8 32.2
90 4.35 149.3 0.37 446.7 273.9 544 67 1.44 66 288 0.228 157.8 24.1
Page 109
A7
Load resistance – R = 170 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia
(A)
Uf (V) If (A) Pin (W) Pmec (W) N
(rpm)
Us (V) Is
(A)
Ps
(W)
Ss
(VA)
fp C (µF) Eff (%)
158.8 4.23 141.5 0.37 724.1 544.3 1190 129 1.44 289 554 0.52 31.4 53.1
151.8 4.19 141.5 0.37 688.4 512.2 1125 124 1.43 266 524 0.51 34.9 51.9
139.2 4.22 141.5 0.37 639.5 464.4 1015 117 1.39 233 482 0.482 39.3 50.2
134.2 4.24 141.5 0.37 621.4 446.2 965 112 1.39 218 467 0.469 49.9 48.9
128.3 4.25 141.6 0.37 597.7 423.1 913 108 1.37 200 444 0.45 50.8 47.3
123.3 4.28 141.6 0.37 580.1 404.7 878 101 1.36 188 404 0.446 59.5 46.4
120.4 4.29 141.6 0.37 568.9 393.8 842 100 1.37 177 410 0.428 60.6 45.0
105.8 4.27 141.5 0.37 504.1 332.4 718 90 1.36 133 362 0.367 78.2 40.0
102.4 4.19 141.8 0.37 481.2 314.7 682 85 1.38 122 352 0.346 88.0 38.8
97.1 4.22 141.8 0.37 461.7 294.9 634 78 1.38 103 325 0.319 104.8 34.9
87.7 4.25 141.8 0.37 424.6 257.8 549 63 1.38 69 263 0.263 157.8 26.8
85.7 4.23 141.8 0.37 414.4 248.9 530 59 1.39 59 245 0.243 175.1 23.7
Load resistance – R = 150 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia (A) Uf (V) If (A) Pin (W) Pmec (W) N
(rpm)
Us (V) Is
(A)
Ps
(W)
Ss
(VA)
fp C (µF) Eff (%)
157.0 4.28 144.7 0.38 726.9 543.6 1134 123 1.53 288 564 0.513 40.0 53.0
144.7 4.29 144.7 0.38 675.7 494.2 1033 113 1.50 249 508 0.493 49.9 50.4
137.7 4.35 144.7 0.38 654.0 470.6 971 109 1.49 231 483 0.479 50.8 49.1
131.7 4.33 144.7 0.38 625.2 444.0 920 103 1.46 210 456 0.463 59.5 47.3
127.1 4.40 144.7 0.37 613.2 431.0 871 101 1.47 199 444 0.448 60.6 46.2
111.0 4.37 144.7 0.37 539.0 361.0 741 90 1.44 152 389 0.390 78.2 42.1
106.5 4.29 144.7 0.37 510.9 337.6 705 84 1.42 132 357 0.370 88.0 39.1
101.1 4.35 144.7 0.37 493.8 318.4 649 78 1.43 114 335 0.340 104.8 35.8
91.6 4.30 144.7 0.37 446.7 277.2 567 62 1.42 76 264 0.288 157.8 27.4
90.0 4.36 144.7 0.37 445.2 272.8 555 59 1.42 66 250 0.264 175.1 24.2
Load resistance – R = 120 Ω
DC MOTOR INDUCTION GENERATOR
Ua
(V)
Ia (A) Uf (V) If (A) Pin (W) Pmec
(W)
N
(rpm)
Us (V) Is
(A)
Ps
(W)
Ss
(VA)
fp C (µF) Eff
(%)
137.9 4.22 141.3 0.37 634.8 459.7 965 94 1.55 207 434 0.475 62.0 45.0
122.9 4.26 141.3 0.37 576.4 402.0 834 85 1.53 170 389 0.434 78.2 42.3
122.0 4.31 141.3 0.37 578.7 401.9 822 80 1.55 157 376 0.417 88.0 39.1
113.8 4.31 141.3 0.37 543.3 367.7 756 74 1.52 132 338 0.389 104.8 35.9
111.0 4.28 141.3 0.37 528.4 354.6 735 68 1.53 113 310 0.364 121.2 31.9
107.5 4.21 141.3 0.37 505.4 336.0 705 59 1.55 85 272 0.315 157.8 25.3
99.4 3.91 141.7 0.37 441.6 287.8 655 53 1.46 66 228 0.300 175.1 22.9
Page 110
A8
D. Tables with Experimental results of PAT application
Rotational Speed – N = 1525 rpm
Rotational Speed – N = 1370 rpm
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
5.010 2.153 1.168 10.047 1525
112 0.57 230 15.5 493.791 22.682
4.761 2.127 1.163 9.833 1525
107 0.57 230 15.5 459.245 23.299
4.620 2.135 1.164 9.904 1525
107 0.57 230 15.5 448.880 23.837
4.510 2.108 1.158 9.690 1525 91.4 0.65 176 101 0.57 230 15.5 428.716 23.559
4.461 2.092 1.154 9.568 1525 88.00 0.63 165 95 0.57 230 15.5 418.701 22.689
4.421 2.082 1.151 9.496 1526 86.50 0.61 157 90 0.57 230 15.5 411.850 21.853
4.272 2.052 1.145 9.251 1530 79.60 0.57 135 77 0.57 230 15.5 387.711 19.860
4.176 2.033 1.141 9.098 1532 76.40 0.54 123 70 0.57 230 15.5 372.730 18.780
4.060 2.017 1.137 8.976 1536 72.00 0.81 110 63 0.57 230 15.5 357.502 17.622
3.878 1.996 1.13 8.833 1544 65.50 0.47 93 52 0.57 230 15.5 336.043 15.474
3.845 1.987 1.131 8.731 1546 63.00 0.45 85 48 0.57 230 15.5 329.336 14.575
3.752 1.97 1.124 8.629 1551 59.50 0.43 75 42 0.57 230 15.5 317.616 13.224
3.599 1.956 1.122 8.507 1561 55 0.39 63 35 0.57 230 15.5 300.343 11.653
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
4.820 2.057 1.168 9.068 1374 54.7 0.4 60 107 0.55 230 18.5 428.764 24.955
4.750 2.047 1.164 9.007 1374 56.0 0.40 65 104 0.55 230 18.5 419.685 24.780
4.620 2.029 1.159 8.874 1374 91.9 0.72 182 101 0.55 230 18.5 402.189 25.113
4.495 1.993 1.154 8.558 1374 88.7 0.69 171 94 0.55 230 18.5 377.364 24.910
4.421 1.978 1.151 8.435 1374 85.8 0.67 160 88 0.55 230 18.5 365.843 24.054
4.271 1.948 1.145 8.191 1373 81.3 0.64 142 79 0.55 230 18.5 343.174 23.020
4.181 1.935 1.143 8.078 1373 78.7 0.62 135 74 0.55 230 18.5 331.341 22.334
4.100 1.918 1.138 7.956 1377 75.4 0.59 125 68 0.55 230 18.5 319.998 21.250
3.956 1.893 1.133 7.752 1379 70.3 0.52 109 59 0.55 230 18.5 300.842 19.612
3.850 1.875 1.13 7.599 1381 66.5 0.49 100 52 0.55 230 18.5 287.003 18.118
3.745 1.851 1.124 7.415 1381 60.0 0.44 80 44 0.55 230 18.5 272.430 16.151
3.620 1.842 1.12 7.364 1380 58.5 0.41 70 39 0.55 230 18.5 261.526 14.912
3.450 1.825 1.12 7.191 1381 52.5 0.4 56 30 0.55 230 18.5 243.376 12.327
Page 111
A9
Rotational Speed – N = 1200 rpm
Rotational Speed – N = 1130 rpm
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
4.931 2.007 1.18 8.435 1203 -- -- 210 105 0.50 265 19.5 408.047 25.732
4.721 1.974 1.174 8.160 1203 -- -- 198 99 0.50 265 19.5 377.914 26.196
4.66 1.957 1.169 8.038 1203 95.50 0.69 196 98 0.50 265 19.5 367.436 26.671
4.593 1.95 1.169 7.966 1207 93.0 0.67 189 93 0.50 265 19.5 358.936 25.910
4.531 1.938 1.169 7.844 1205 93.0 0.67 184 91 0.50 265 19.5 348.650 26.101
4.506 1.927 1.166 7.762 1204 91.0 0.65 178 89 0.50 265 19.5 343.119 25.939
4.425 1.903 1.161 7.568 1204 88.50 0.64 167 84 0.50 265 19.5 328.539 25.568
4.321 1.878 1.157 7.354 1203 85.50 0.63 156 78 0.50 265 19.5 311.737 25.021
4.239 1.856 1.154 7.160 1203 82.50 0.61 147 73 0.50 265 19.5 297.762 24.516
4.191 1.845 1.151 7.079 1203 81.00 0.60 142 70 0.50 265 19.5 291.036 24.052
4.099 1.828 1.145 6.967 1203 77.50 0.56 130 65 0.50 265 19.5 280.135 23.203
3.991 1.796 1.141 6.681 1203 74.00 0.54 117 58 0.50 265 19.5 261.573 22.174
3.92 1.813 1.145 6.814 1204 73.00 0.53 111 55 0.50 265 19.5 262.018 20.991
3.801 1.77 1.135 6.477 1203 69.00 0.50 100 49 0.50 265 19.5 241.513 20.289
3.723 1.775 1.132 6.559 1205 66.50 0.48 94 46 0.50 265 19.5 239.537 19.204
3.701 1.738 1.126 6.242 1205 63.50 0.46 85 43 0.50 265 19.5 226.642 18.973
3.538 1.712 1.12 6.038 1209 57.50 0.42 71 35 0.50 265 19.5 209.579 16.700
3.491 1.703 1.118 5.967 1210 54.50 0.39 65 32 0.50 265 19.5 204.350 15.659
3.257 1.68 1.12 5.712 1215 48.50 0.35 54 26 0.50 265 19.5 182.505 14.246
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
5.060 2.08 1.171 9.272
109 0.51 230 24.8 460.239 23.683
4.870 1.99 1.168 8.384
100 0.51 230 24.8 400.562 24.965
4.680 1.928 1.158 7.854 1128 85.2 0.71 178 91 0.51 230 24.8 360.583 25.237
4.561 1.899 1.154 7.599 1130 82.6 0.17 169 85 0.51 230 24.8 340.005 25.000
4.476 1.869 1.15 7.334 1128 80.1 0.66 158 80 0.51 230 24.8 322.024 24.843
4.355 1.842 1.147 7.089 1128 77.3 0.64 148 74 0.51 230 24.8 302.860 24.434
4.257 1.817 1.142 6.885 1128 74.4 0.61 136 69 0.51 230 24.8 287.526 23.998
4.170 1.803 1.141 6.752 1130 73.3 0.61 131 66 0.51 230 24.8 276.225 23.894
4.112 1.787 1.138 6.620 1130 71.0 0.59 124 63 0.51 230 24.8 267.034 23.592
3.932 1.744 1.131 6.253 1131 65.5 0.54 106 53 0.51 230 24.8 241.181 21.975
3.752 1.713 1.123 6.018 1132 61.5 0.51 91 46 0.50 230 24.8 221.505 20.767
3.580 1.671 1.118 5.641 1137 54.8 0.46 75 38 0.50 230 24.8 198.097 19.183
3.515 1.667 1.115 5.630 1138 53.1 0.44 72 35 0.50 230 24.8 194.148 18.027
3.382 1.647 1.109 5.488 1139 49.1 0.41 60 29 0.50 230 24.8 182.064 15.928
3.277 1.638 1.108 5.406 1153 46 0.39 52 26 0.50 230 24.8 173.789 14.961
3.112 1.616 1.106 5.202 1153 39.8 0.33 40 20 0.50 230 24.8 158.810 12.594
Page 112
A10
Rotational Speed – N = 1050 rpm
Rotational Speed – N = 930 rpm
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A) Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
5.1 1.98 1.18 8.160 1050
96 0.51 230 31.75 408.253 23.515
4.94 1.966 1.177 8.048 1050
93 0.51 230 31.75 390.008 23.846
4.78 1.922 1.17 7.670 1050
86 0.51 230 31.75 359.679 23.910
4.705 1.918 1.169 7.640 1050 75.0 0.74 166 85 0.51 230 31.75 352.623 24.105
4.7 1.926 1.171 7.701 1050
87 0.51 230 31.75 355.070 24.502
4.615 1.906 1.167 7.538 1050 74.0 0.74 162 82 0.51 230 31.75 341.260 24.029
4.561 1.893 1.166 7.415 1050 73.0 0.73 157 79 0.51 230 31.75 331.790 23.810
4.49 1.87 1.162 7.222 1050 71.5 0.71 149 76 0.51 230 31.75 318.089 23.893
4.275 1.813 1.154 6.722 1050 66.5 0.66 129 66 0.51 230 31.75 281.897 23.413
4.078 1.757 1.144 6.253 1052 61.5 0.62 110 56 0.51 230 31.75 250.136 22.388
3.8 1.687 1.131 5.671 1050 54 0.53 84 44 0.51 230 31.75 211.411 20.813
3.65 1.656 1.125 5.416 1058 49.5 0.48 72 37 0.51 230 31.75 193.935 19.079
3.41 1.62 1.118 5.120 1060
29 0.51 230 31.75 171.288 16.931
3.25 1.59 1.109 4.906 1058
23 0.51 230 31.75 156.422 14.704
2.91 1.558 1.104 4.631 1060
15 0.51 230 31.75 132.196 11.347
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
4.94 1.928 1.169 7.742
931
84.0 0.45 230 35.6 375.178 22.389
4.71 1.88 1.156 7.385
931 78.00 0.73 169 77.0 0.45 230 35.6 341.215 22.566
4.635 1.864 1.155 7.232
931 77.50 0.73 167 75.0 0.45 230 35.6 328.825 22.808
4.561 1.839 1.152 7.007
930 76.00 0.71 159 71.0 0.45 230 35.6 313.535 22.645
4.47 1.82 1.152 6.814
930 74.00 0.69 153 69.0 0.45 230 35.6 298.781 23.094
4.354 1.785 1.146 6.518
930 71.50 0.67 142 63.0 0.45 230 35.6 278.393 22.630
4.211 1.741 1.14 6.130
930 68.00 0.63 129 57.0 0.45 230 35.6 253.238 22.508
4.054 1.703 1.136 5.783
930 65 0.6 114 51 0.45 230 35.6 230.004 22.173
3.86 1.662 1.131 5.416
928 60.50 0.56 100 44.0 0.45 230 35.6 205.093 21.454
3.74 1.631 1.126 5.151
929 57.00 0.53 89 40.0 0.45 230 35.6 188.987 21.165
3.375 1.57 1.116 4.631
930 49.00 0.45 65 29.0 0.45 230 35.6 153.320 18.915
3.285 1.531 1.109 4.304
931 45.00 0.42 55 26.0 0.45 230 35.6 138.713 18.744
2.981 1.494 1.104 3.978
940 35.00 0.32 39 17.0 0.45 230 35.6 116.331 14.613
Page 113
A11
Rotational Speed – N = 810 rpm
Rotational Speed – N = 700 rpm
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
5.070 1.943 1.18 7.783 703
58 0.42 150 96.6 387.081 14.984
5.010 1.926 1.175 7.660 703
57 0.42 150 96.6 376.484 15.140
4.870 1.905 1.173 7.466 703
56 0.42 150 96.6 356.705 15.699
4.771 1.854 1.164 7.038 703 51.50 0.79 121 51 0.42 150 96.6 329.403 15.483
4.651 1.826 1.161 6.783 699 50 0.77 117 49 0.42 150 96.6 309.483 15.833
4.571 1.803 1.159 6.569 701 49.2 0.76 111 47 0.42 150 96.6 294.555 15.956
4.560 1.79 1.155 6.477 699 48.9 0.75 110 45 0.42 150 96.6 289.740 15.531
4.510 1.78 1.154 6.385 699 48.3 0.74 107 44 0.42 150 96.6 282.501 15.575
4.388 1.745 1.148 6.089 699 46.5 0.71 100 41 0.42 150 96.6 262.126 15.641
4.201 1.684 1.139 5.559 699 43.9 0.67 89 38 0.42 150 96.6 229.096 16.587
4.080 1.661 1.134 5.375 698 42.2 0.65 82 35 0.42 150 96.6 215.149 16.268
3.870 1.631 1.131 5.100 697 40.7 0.62 76 31 0.42 150 96.6 193.620 16.011
3.760 1.583 1.123 4.692 698 38.3 0.55 66 27 0.42 150 96.6 173.067 15.601
3.570 1.532 1.117 4.233 699 35.2 0.54 56 24 0.42 150 96.6 148.247 16.189
3.510 1.516 1.113 4.111 697 33.3 0.51 54 22 0.42 150 96.6 141.541 15.543
3.400 1.486 1.109 3.845 699 32.0 0.48 46 20 0.42 150 96.6 128.259 15.593
3.230 1.463 1.106 3.641 701 30.0 0.46 41 17 0.42 150 96.6 115.382 14.734
3.020 1.428 1.101 3.335 701 25.8 0.39 32 12 0.42 150 96.6 98.815 12.144
2.920 1.386 1.094 2.978 715 22.5 0.33 23 9 0.42 150 96.6 85.317 10.549
2.527 1.352 1.089 2.683 725 17.4 0.27 16 6 0.42 150 96.6 66.501 9.022
HYDRAULIC PARAMETER ELECTRICAL PARAMETER
Q (l/s) pressure_up (bar)
pressure_down (bar)
H (mwc)
N (rpm)
Us (V)
Is (A)
Ss (VA)
Ps (W)
PF R (Ω)
C (uF)
P_Hyd (W)
global eff (%)
5.11 1.945 1.175 7.854 813
77 0.41 230 57.7 393.714 19.557
5.06 1.932 1.178 7.691 813
76 0.41 230 57.7 381.761 19.908
5.01 1.911 1.171 7.548 813
74 0.41 230 57.7 370.970 19.948
4.89 1.885 1.17 7.293 813
71 0.41 230 57.7 349.852 20.294
4.764 1.873 1.166 7.211 813 73.50 0.75 167 68 0.41 230 57.7 337.024 20.177
4.731 1.866 1.165 7.150 813 73.00 0.75 164 67 0.41 230 57.7 331.849 20.190
4.63 1.839 1.16 6.926 813 71.50 0.73 157 63 0.41 230 57.7 314.572 20.027
4.554 1.809 1.159 6.630 812 69.50 0.71 149 60 0.41 230 57.7 296.194 20.257
4.331 1.749 1.146 6.151 812 65.00 0.67 131 52 0.41 230 57.7 261.321 19.899
4.26 1.719 1.14 5.906 811 62.50 0.65 122 49 0.41 230 57.7 246.807 19.854
4.15 1.697 1.139 5.692 810 61.50 0.63 116 46 0.41 230 57.7 231.714 19.852
3.991 1.659 1.133 5.365 812 58.50 0.59 103 41 0.41 230 57.7 210.057 19.519
3.79 1.616 1.125 5.008 807 55 0.56 91 36 0.41 230 57.7 186.204 19.334
3.6 1.557 1.118 4.478 807 49 0.50 73 30 0.41 230 57.7 158.138 18.971
3.452 1.536 1.114 4.304 807 47 0.47 67 27 0.41 230 57.7 145.765 18.523
3.342 1.516 1.111 4.131 808 45 0.45 62 24 0.41 230 57.7 135.435 17.721
3.14 1.475 1.106 3.764 808 40 0.41 48 19 0.41 230 57.7 115.938 16.388
2.901 1.435 1.099 3.427 807 35 0.35 35 15 0.41 230 57.7 97.534 15.379