Case Study Examples Illustrating Modeling of PV for ...energy.sandia.gov/wp-content/gallery/uploads/...Jeff-Smith.pdf · Case Study Examples Illustrating Modeling of PV for Distribution

Jeff Smith Manager, Power System Studies, EPRI

Workshop on Achieving High Penetrations of PV

Orlando, Florida

9/14/2012

Case Study Examples Illustrating Modeling of PV for Distribution Planning and Analysis

2 © 2012 Electric Power Research Institute, Inc. All rights reserved.

Industry Emphasis on Modeling – Why?

• Rapid-deployment of PV • Higher penetration levels than

ever before • Utilities need to have a handle

on potential PV impacts prior to interconnection

• Much easier to resolve potential issues before penetration levels exceed allowable limits

• Modeling enables the distribution planner to investigate solutions and “work arounds”

ANSI Voltage limit

Increasing PV Output

Local voltage limit

Example of an Existing PV System Causing Unacceptable Voltages Measured Overvoltage Resulting from PV


Right Tool for the Job

Modeling is only as good as the tool used, and the data that goes into it

Power System Modeling and

Analysis

Steady-State

Quasi-Steady State

Short-Circuit

Electromagnetic transient

Stability

Scheduling/ Dispatch

Distributed PV Interconnection

Analysis


Feeder Modeling

Power Delivery – Substation transformer – Primary conductors – Service transformers – Customer service drops

Power Conversion – Residential loads – Commercial loads – Existing PV systems

Power Control – Capacitors – Substation LTC – Line regulators


Solar PV Modeling OpenDSS PV Model

Allows for simulation of wide range of feeder impacts

• Steady-state load flow – Voltage regulation – Thermal loading

• Time series load flow – Interaction with regulation

controls • Short-circuit response

– Increased fault duty • Harmonics

– THD calculations


Solar Resource Modeling - Variability

• Essential to capture time-varying nature of the solar PV

– 15 minute - Hourly – energy related concerns such as peak demand, loading, overvoltage

– < 1 minute – voltage fluctuations, regulation control interaction

• Factors that contribute to variability – Size and layout of the solar facility – Local weather patterns – Centralized vs distributed

• Possible sources of data: – Location-specific measurement data – Proxy data from other PV systems – Mesoscale/satellite based models –

typically 30 minute - hourly interval

Solar Plant Size

Ramping (0-100%)

60 kW < 10 seconds

800 kW < 1 minute

6.5 MW 1.5-2 minutes

Feeder-wide 5-10 minutes

Example Ramp Rates from Existing PV

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

7 9 11 13 15 17 19Hour

Pow

er (p

er-u

nit)

1 min avg60 min avg

PV data “driving” the simulation must be sampled faster than the metric and/or control interaction being simulated


Feeder and PV Model Validation

OpenDSS Feeder and PV Model Validated Against Field Data

RegulatorsMetered Large-Scale PV

Substation

Small variations in voltage due to load fluctuations

Measured PV Production

Field vs Simulated (OpenDSS)


Considers Maximum Range of

PV Output

Static Load Flow - solve load flow - lock all controls - add generation - resolve

Basic Analysis – Voltage Change Test

3-Phase Voltage Profile Plot

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1

1.01

1.02

1.03

1.04

1.05

1.06

0 2 4 6 8 10 12 14 16

Distance from Substation

Per U

nit V

olta

ge

V1V2V3

"HEADROOM""HEADROOM"

3-Phase Voltage Profile Plot

With 1600 kW Uniformly-Distributed Generation

0.92

0.93

0.94

0.95

0.96

0.97

0.98

0.99

1

1.01

1.02

1.03

1.04

1.05

1.06

0 2 4 6 8 10 12 14 16

Distance from Substation (km)

Per U

nit V

olta

geV1V2V3


Advanced Analysis – Time Series Quasi-Dynamic, Quasi-Static, etc

IEEE 8500 Node Test Case

1.025

1.03

1.035

1.04

1.045

04 06 08 10 12 14 16 18 20

volta

ge (p

u)

Hour

Daily Voltage Profile(Average Three Phase)

With PV

Without PV

Mid

-Fee

der

-4

-2

0

2

4

6

8

10

12

14

0 4 8 12 16 20

Tap O

pera

tion

Diffe

rent

ial

Hour

Phase A

Phase B

Phase C

A B C

4% 20% 7%

Increase in Tap Operations Due to PV

2MW PV

2MW PV System located Mid-Feeder Causing Increased Regulator Operation


Spatial- and Time-Based Modeling


Modeling Uncertainty Regarding Future PV Deployment Stochastic Analysis

• Can be used for determining a particular distribution feeder’s hosting capacity for PV by considering large number of possible future PV deployments

• Must consider wide range of – potential spatial interconnections – PV system sizes

• Likelihood of a simulated PV deployment resembling the actual future deployment increases with number of cases considered


Sample Stochastic Results Maximum Feeder Voltage

Blue – primary feeder voltage

Red – secondary customer voltage

Increasing penetration (kW)

Penetration level when violations occur Feeder exposure to violations

Max

imum

Fee

der V

olta

ge

# of

Fee

der N

odes

Vio

late

d

•Each point represents the worst-case voltage violation for a single PV deployment case

•5000 cases shown

Increasing penetration (kW)

ANSI voltage limit


What About Evaluating Smart Inverters?

Smart inverters for providing grid support functions

Volt-Var Control* Volt-Watt Control**

Dynamic Var Control** Volt-Var w/ Hysteresis**

*Currently in OpenDSS **available in OpenDSS Q1 2013

Applying PV Volt-Var ControlMitigating Voltage Issuesg g g

Solar Rooftop PVSolar Rooftop PVCustomer PVVolt-Var Control Solar Rooftop PVSolar Rooftop PV

With volt/var controlCustomer Load Customer PV

Gen

erated Capacitive System

Voltage

Q1

Q3Q2

Volt Var Control

VARs Inductive V1 V2 V3V4

Q4

Primary Voltage

1

1.025

1.05

20% PV20% PV withvolt/var control

0.925

0.95

0.975

1

Vol

tage

(pu)

Baseline – No PV

14© 2012 Electric Power Research Institute, Inc. All rights reserved.

0.90 4 8 12 16 20

Hour

Baseline – No PV


Open Discussions

Contact:

Jeff Smith Manager, Power System Studies EPRI [email protected]


Simulation Platform – OpenDSS

• Open source of EPRI’s Distribution System Simulator

– developed in 1997 – open sourced in 2008 to collaborate

with other research projects

• Used in 100’s of distribution studies • OpenDSS designed from the

beginning to capture – Time-specific impacts and – Location-specific impacts

• Differentiating features

– full multiphase model – numerous solution modes – “dynamic” power flow – system controls – flexible load models

• Needed for analysis of – DG/renewables – energy efficiency – PHEV/EV – non-typical loadshapes

http://sourceforge.net/projects/electricdss

Case Study Examples Illustrating Modeling of PV for ...energy.sandia.gov/wp-content/gallery/uploads/...Jeff-Smith.pdf · Case Study Examples Illustrating Modeling of PV for Distribution

Documents