Cải tiến bộ điều khiển mờ cấp cao cho bài toán điều khiển độ ấm nhà kính

Hội nghị toàn quốc về Điều khiển và Tự động hoá - VCCA-2011

VCCA-2011

Cải tiến bộ điều khiển mờ cấp cao

cho bài toán điều khiển độ ấm nhà kính

A Development on Advanced Fuzzy Based Controller Design for

Humidity Control of Greenhouse

Minh Duc Nguyen1, Viet Boi Chau Luong

2, Tuong Quan Vo

3

1 Department of Mechatronics, Faculty of Mechanical Engineering, HCMUT, Vietnam

(Email: [email protected]) 2 Department of Mechatronics, Faculty of Mechanical Engineering, HCMUT, Vietnam

(Email: [email protected]) 3 Department of Mechatronics, Faculty of Mechanical Engineering, HCMUT, Vietnam

(Email: [email protected])

Tóm tắt Ngày nay, nhà kính không còn là một khái niệm lạ

trong nông nghiệp. Tuy nhiên, điều hòa khí hậu trong

nhà kính còn gặp nhiều khó khăn. Nhiều thông số cần

được điều khiển chẳng hạn như nhiệt độ, độ ẩm tương

đối, độ chứa hơi, lượng khí carbon dioxide trong

không khí ... Trong đó, khó khăn nhất là điều khiển độ

ẩm tương đối. Trong bài báo này đề cập đến bộ điều

khiển mờ thông thường và bộ điều khiển mờ nâng cao

được sử dụng cho máy tăng ẩm và máy hút ẩm. Bộ

điều khiển mờ nâng cao là tự điều chỉnh các thông số

đầu ra dựa trên sai số và đạo hàm của sai số của các

biến điều khiển.

Abstract Nowadays, greenhouse is not longer a strange

conception in agriculture. The plants in a greenhouse

impose their own needs. However, climate control for

protected crops has many difficulties. There are many

parameters are controlled such as temperature,

relative humidity, humidity ratio, carbon dioxide in

the air… Among protecting all of them, relative

humidity is the hardest parameter to control. In this

paper, Conventional Fuzzy Controller (CFC) and

Self-tuning Fuzzy Logic Controller (STFLC) are used

for the humidifier and dehumidifier. The Self-tuning

Fuzzy Logic Controller is adjusted the output scaling

factor on-line by fuzzy rules according to the current

trend of the controlled process. The rule-base for

tuning the output-scaling factor is defined on error

and change of error of the controlled variable.

Keywords Greenhouse, Conventional Fuzzy Controller,

Advanced Fuzzy Controller, Self-tuning Fuzzy

Controller, humidity.

Nomenclature in % indoor relative humidity

o1 % outside relative humidity type 1

Acc m2 surface area of cooling coils

Agap = 0.02m2 area of gaps of wall

Ahc m2 surface area of heating coils

cp kJ/kg.oC specific heat capacity of moist

air

dac kg/kg humidity ratio of air flow

passing over cooling coils

dh kg/kg humidity ratio of supply air flow

by humidifier

din kg/kg humidity ratio of indoor air

do kg/kg humidity ratio of outdoor air

Fo1 m3/s volume flow rate of air blowing

from outside to inside

G kg mass of dry indoor air

Gd kg/s mass flow rate of supply dry air

flow by dehumidifier

Gh kg/s mass flow rate of supply dry air

flow by humidifier

Go kg/s mass flow rate of air blowing

from outside to inside

Gw kg/s mass flow rate of added water

vapor

hcc kW/m2.oC average convective heat-transfer

coefficient of cooling coils

hhc kW/m2.oC average convective heat-transfer

coefficient of heating coils

ih kJ/kg enthalpy of supply air by

humidifier

iin kJ/kg enthalpy of indoor air

iw kJ/kg enthalpy of added water vapor

p = 1 bar pressure of air

phmax bar saturation vapor partial pressure

at indoor temperature

tac oC temperature of air flow passing

over cooling coils

tah oC temperature of air flow passing

over heating coils

tcc oC surface temperature of cooling

coils

th oC temperature of supply air by

humidifier

thc oC surface temperature of heating

775


VCCA-2011

coils

tin oC temperature of indoor air

to oC temperature of outdoor air

to1 oC outside temperature type 1

vwind m/s outside wind speed

1. Introduction We find that the climate in the crop have a great

influence on the growth of plants, capacity growth,

productivity, quality and of tree maintenance

procedures. Environmental control is a central feature

of modern production systems. There is so much

research on the theory and application the related

issues of environmental control Greenhouse has been

done by many researchers (Jones, 1984 [Error!

Reference source not found.]; Gates and Overhuts,

1991 [9]; Stanghellini and van Meurs, 1992 [10];

Taylor, 2000 [11]; Zolnier, 2000 [12]). Most research

has focused on the analysis and control of

environmental conditions inside the glass based on the

concept of energy and material model reasons.

Many kinetic models for greenhouse environmental

developed and documented in the literature, the model

the kinetics of these are non-linear model with

variables mainly air temperature and relative humidity

for (or absolute humidity), concentrations of carbon

dioxide also mentioned. Noise impact on the climate

glass all tastes from solar radiation, outside

temperature (the phenomenon of thermal conductivity

and heat transfer), but the frequency the noise is very

low. Indeed, temperature and humidity are closely

associated with each other through the law of non-

linear thermodynamics. Therefore, we designed a

controller that is strong enough to control system for

greenhouse is difficult.

2. System Modeling Humidifier has two systems: atomizing system and air

handling unit. Water passes through an atomizing

system to become water vapor. This water vapor will

be added to the air.

Following [1], the humidity ratio of supply air flow

by humidifier is

wh in

h

Gd d

G (1)

Following [1, 2], the temperature of supply air flow

by humidifier is

in in w h w h

h in

p

d i i d i it t

c (2)

The air in a dehumidifier first passes over a series of

cooling coils (the evaporator) and then immediately

over a set of heating coils (the condenser) and then

back into the room as dryer air with its temperature

elevated.

Following [1, 2], the humidity ratio of air flow

passing over cooling coils is

1.006

2500.77 1.84

cc cc in cc

in ac

dac

ac

h A t ti t

Gd

t (3)

Following [1, 2], the temperature of air flow passing

over heating coils is

1.006 1.84

hc hc hc ac

ah ac

d ac

h A t tt t

G d (4)

Therefore, the temperature td and the humidity ratio dd

of supply air flow by dehumidifier are

d aht t (5)

d acd d (6)

According law of energy conservation, if we choose

the sampling time is one second, the indoor

temperature and humidity ratio at the step k are

1

h h d d o o

in h d

in

o

t k G k t k G k t k G k

t k G k G k G kt k

G k G k

(7)

1

h h d d o o

in h d

in

o

d k G k d k G k d k G k

d k G k G k G kd k

G k G k

(8)

The indoor relative humidity is

max0.622

inin

h

d p

d pj (9)

3. Controllers Design In Controller, there are two inputs:

The system error is defined as the difference

between the plant output y(k) and the set point

r(k) at the step k is :

e k y k r k (10)

The change rate of error at the step k is

1de k e k e k (11)

And two outputs, as the inputs of the plant:

For humidifier: the mass flow rate of water vapor

Gw (kg/s).

For dehumidifier: the volume flow rate of supply

air Fd (m3/s)

3.1. Conventional Fuzzy Controller (CFC)

The structure of CFC is shown in Fig.1. The input

range for e and de are based on the load disturbance.

The output ranges for Gw and Fd are base on the max

power of humidifier and dehumidifier. The language

variables description of inputs and outputs are shown

in Table 1. The fuzzy membership functions for

inputs and outputs are shown in Fig.2. Table 2 lists 49

language fuzzy rules for the CFC.

776


VCCA-2011

Conventional

Fuzzy

Controller

Plant

du/dt

Gw

Fdder

y

e

Fig.1. Structure of Conventional Fuzzy Controller

Table 1. Language variables description

Inputs Output

NB Negative Big ZE Zero 1

NM Negative Medium VS Very Small 2

NS Negative Small S Small 3

ZE Zero SB Small Big 4

PS Positive Small MB Medium Big 5

PM Positive Medium B Big 6

PB Positive Big VB Very Big 7

-20 -10 0 10 20

0

0.5

1

Error

NB NM NS ZE PS PM PB

-0.5 -0.25 0 0.25 0.5

0

0.5

1

Change of Error


0 0.025

0

0.5

1

Gw

ZE VS S SB MB B VB

0 0.12

0

0.5

1

Fd

ZE VS S SB MB B VB

Fig.2. Membership functions of input and output CFC

Table 2. Rule of Conventional Fuzzy Controller

Gw Fd

Error


Ch

ang

e o

f E

rro

r NB 7 1 6 1 5 1 2 1 1 2 1 4 1 5

NM 7 1 6 1 4 1 2 1 1 2 1 4 1 5

NS 6 1 5 1 3 1 2 1 1 3 1 5 1 6

ZE 6 1 5 1 3 1 1 1 1 3 1 5 1 6

PS 5 1 4 1 2 1 1 2 1 4 1 6 1 6

PM 5 1 4 1 2 1 1 3 1 5 1 6 1 7

PB 4 1 3 1 2 1 1 4 1 6 1 7 1 7

3.2. Advanced Fuzzy Controller (AFC)

Advanced Fuzzy Controller has two fuzzy controllers.

They are Direct Fuzzy Logic Controller (DFLC) and

Self-tuning Fuzzy Logic Controller (STFLC). The

structure of AFC is shown in Fig.3.

Direct Fuzzy

Logic

Controller

Plant

du/dt

GwN

FdNder

y

eGe

Gde

eN

deN

GuHum

GuDehum

Gw

Fd

Self-tuning

Fuzzy

Logic

Controller

Fig.3. Structure of Advanced Fuzzy Controller

DFLC is most similar to CFC but the ranges of inputs

and outputs are different. In DFLC, the inputs (eN,

deN) and outputs (GwN, FdN) are normalized based on

Eqs.(12-15).

max 20N

e ee

e (12)

max 0.5N

de dede

de (13)

min

max min

0

0.025 0 0.025

w w w wwN

w w

G G G GG

G G (14)

min

max min

0

0.12 0 0.12

d d d ddN

d d

F F F FF

F F (15)

The rule for DFLC is the same with CFC. The fuzzy

membership functions for inputs and outputs are

shown in Fig.4.

-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

0

0.5

1

Error


-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

0

0.5

1

Change of Error


0 0.15 0.3 0.45 0.6 0.75 0.9 1

0

0.5

1

Gw

ZE VS S SB MB B VB

0 0.15 0.3 0.45 0.6 0.75 0.9 1

0

0.5

1

Fd

ZE VS S SB MB B VB

Fig.4. Membership functions of input and output of DFLC

The rule base STFLC is developed to tune the

DFLC’s inputs and outputs gains. The inputs are e

and de and outputs are Ge, Gde, GuHum, GuDehum. Its

inputs are the same with CFC’s, both range and

membership functions. The ranges for outputs are

based on the ranges for inputs and outputs of CFC.

The membership functions of inputs and outputs are

777


VCCA-2011

shown in Fig.5. The rule of STFLC is shown in Table

3. Table 3. Rule of STFLC for Ge and Gde

Ge Gde Error

GuHum GuDehum NB NM NS ZE PS PM PB

Chan

ge

of

Err

or

NB 1 4 2 3 5 3 7 7 5 1 4 1 3 1

1 4 2 3 5 3 7 7 5 1 4 1 3 1

NM 2 3 5 2 6 4 6 6 6 1 5 1 4 1

2 3 5 2 6 4 6 6 6 1 5 1 4 1

NS 5 2 6 3 4 5 5 5 4 2 6 1 5 1

5 2 6 3 4 5 5 5 4 2 6 1 5 1

ZE 7 2 6 3 5 5 7 4 5 5 6 3 7 2

7 2 6 3 5 5 7 4 5 5 6 3 7 2

PS 5 1 6 1 4 2 5 5 4 5 6 3 5 3

5 1 6 1 4 2 5 5 4 5 6 3 5 3

PM 4 1 5 1 6 1 6 6 6 4 5 2 2 3

4 1 5 1 6 1 6 6 6 4 5 2 2 3

PB 3 1 4 1 5 1 7 7 5 3 2 3 1 4

3 1 4 1 5 1 7 7 5 3 2 3 1 4

0 0.2

0

0.5

1

Ge

ZE VS S SB MB B VB

0 4

0

0.5

1

Gde

ZE VS S SB MB B VB

0 0.03

0

0.5

1

GuHum

ZE VS S SB MB B VB

0 0.16

0

0.5

1

GuDeHum

ZE VS S SB MB B VB

Fig.5. Membership functions of input and output STFLC

4. Simulation Results Using the mathematical model of the plant our

proposed approaches has been tested. From the initial

condition tin(0) = 27oC, in(0) = 65%. The target is to

follow a control reference set to ref1 = 40% and ref2

= 80%.

We test three fuzzy controllers:

Fuzzy1: Conventional Fuzzy Controller

Fuzzy2: Advanced Fuzzy Controller with self-

tuning Gde

Fuzzy3: Advanced Fuzzy Controller with self-

tuning Ge, Gde, GuHum, GuDehum

First, the controllers are tested without disturbance.

The simulation results for both references are

represented in Fig.6-7. In Fig.6, at the beginning,

error is negative big. So the dehumidifier runs with

max power. But there are some differences when error

is negative small. The settling time of Fuzzy3 is the

least. In Fig.7, the errors of Fuzzy2 and Fuzzy3 are

close to zero. The error of Fuzzy1 is about 1.75.

Second, we consider the response of the system

within disturbance. Disturbance is the effect of

outside air. There are two kinds of disturbances:

normal disturbance and unpredictable disturbance.

Normal disturbance is the change of outside

temperature, relative humidity and wind speed in a

day (following Table 4). The equations of them are

shown in below:

1

1

1 1

27.5 7.5sin

78.5 7.5sin

0.02 2.8 sin

o t

o

o gap wind w

t w t

w t

F A v w t

(16)

0 100 200 300 400 500 600 700 800 900 1000 1100 1200

40

42

44

46

48

50

5254

56

58

60

62

64

6668

Time (s)

Rela

tive H

um

idity(%

)

Fuzzy1

Fuzzy2

Fuzzy3

Fig.6. Response of Ref = 40 % without disturbance

0 100 200 300 400 500 600 700 800 900 1000 1100 120064

66

68

70

72

74

76

78

80

82

84

86

Time (s)

Rela

tive H

um

idity(%

)

Fuzzy1

Fuzzy2

Fuzzy3

Fig.7. Response of Ref = 80 % without disturbance

Unpredictable disturbance is the suddenly change of

temperature to2(oC), relative humidity o2 (%) and

volume flow rate of air Fo2 (m3/s) when someone

opens door or the weather becomes heavy.

Disturbance signals are in the form given in Fig.8.

In Fig.9-10, Fuzzy1 and Fuzzy2 take more time to

response than Fuzzy3.

We evaluate the quality of the controllers through

three parameters: Settling time

Error

Sum of square error

2

1

n

i

real i ref i

SSEn

with n = number of samples

778


VCCA-2011

0 240 480 720 960 1200

15

20

25

30

Time (s)

t o2(o

C)

0 240 480 720 960 1200

80

95

Time (s)

Phi o

2(%

)

0 50 290 530 770 1010 1200

0

5

10

Time (s)

Fo2(m

3/s

)

Fig.8. Unpredictable Disturbance Signals

Table 4. Relative Humidity and Temperature of Ho Chi

Minh City in year

Month Average RH

(%)

Max Average

Temperature

(oC)

Min Average

Temperature

(oC)

1 73.8 32 21

2 71.1 33 22

3 71 34 23

4 73.7 34 24

5 80.7 33 25

6 83.7 32 24

7 84.2 31 25

8 84.5 32 24

9 86 31 23

10 85.2 31 23

11 81.7 30 22

12 77.8 31 22

0 100 200 300 400 500 600 700 800 900 1000 1100 1200

40

42

44

46

48

50

5254

56

58

60

62

64

6668

Time (s)

Rela

tive H

um

idity(%

)

Fuzzy1

Fuzzy2

Fuzzy3

Fig.9. Response of Ref = 40 % within disturbance

0 100 200 300 400 500 600 700 800 900 1000 1100 120064

66

68

70

72

74

76

78

80

82

84

86

Time (s)

Rela

tive H

um

idity(%

)

Fuzzy1

Fuzzy2

Fuzzy3

Fig.10. Response of Ref = 80 % within disturbance

Table 5. Performance characteristics of system with three

Fuzzy Controllers

Settling time (s)

Error (%)

SSE

Fuzzy1 Fuzzy2 Fuzzy3

Wit

ho

ut

Dis

turb

ance

Ref = 40 %

213 175 107

0.5 0.1 0.15

5.0772 4.7607 4.8755

Ref = 80 %

140 100 62

1.10 0.10 0.03

2.4961 1.3167 1.0541

Dis

turb

ance

Ref = 40 %

264 234 155

1.00 0.52 0.63

26.5230 22.0187 21.1630

Ref = 80 %

130 150 120

1.10 0.45 0.20

3.1697 2.0932 1.3749

5. Conclusion In this paper, we simulated the greenhouse humidity

system and the fuzzy controllers. Conventional Fuzzy

Controller and Advanced Fuzzy Controller are both

satisfactory the requirement (error < 2%). But the

performance (Table 5) of Advanced Fuzzy Controller

is better: the setting time is smaller and the setting

error is less. We can choose Advanced Fuzzy

Controller for the fulfillment of complex task of

adaptive set point tracking and disturbance rejection

of a greenhouse humidity system.

References

[1] Hoàng Đình Tín, Lê Chí Hiệp, Nhiệt động lực học kĩ

thuật, Nhà xuất bản Khoa học và Kĩ thuật, 2000.

[2] Shan K. Wang, Handbook of air conditioning and

refrigeration 2nd, McGraw-Hill

[3] Bùi Hải, Tính toán thiết kế hệ thống điều hòa không

khí theo phương pháp mới, Nhà xuất bản Khoa học và

Kĩ thuật, 2005.

[4] Lương Việt Bội Châu, Nguyễn Minh Đức và Võ

Tường Quân, Nghiên cứu thiết kế bộ điều khiển máy

phun sương tăng độ ẩm nhà kính dung Truth Table và

Fuzzy, The 2nd Symposium on Mechanical

Engineering for Young Scienntists (SMEYS 2011)

[5] Farzan Rashidi, Aminollah Kermanpour, Design a

Nonlinear Fuzzy PID Controller for Control for

Nonlinear HVAC Systems, Expert System with

Applications 36 (2009) pp. 4566-4573

[6] Servet Soyguder, Mehmet Karakose, Hasan Alli,

Design and simulation of self-tuning PID-type fuzzy

adaptive control for an expert HVAC system

[7] G.D. Pasgianos, K.G. Arvanitis, P. Polycarpou, N.

Sigrimis, A nonlinear feedback technique for

greenhouse environmental control, Computers and

Electronics in Agriculture 40 (2003) pp. 153-177

[8] Mansour Souissi, Mohamaed Chaabane and Ahmed

Toumi, Greenhouse Climate Discrete Control with

Guaranteed H Performance, International Journal of

Computation Cognition, Vol 3, No.4, December 2005

[9] Gates, R.S., Overhults, D.G. – Field evaluation of

integrated environmental controllers. ASAE Paper

No. 91-4037 (1991). ASAE Publishers, Saint Joseph,

MI.

[10] Stanghellini, C., van Meurs, W.T.M., Environmental

control of greenhouse crop transpiration. J.Agric.

Eng. Res. 51 (1992) 297-311.

[11] Taylor, C.J., Chotai, A., Young, P.C., State space

control system design based on non-minimal state

variable feedback: further generalizations and

unification results. Int. J. Control 73 (2000), 1329-

1345.

[12] Zolnier, S., Gates, R.S., Buxton, J., Mach, C.,

Psychrometric and ventilation constraints for vapor

pressure deficit control. Comput. Electron. Agric. 26 (2000) 343-359.

779

Cải tiến bộ điều khiển mờ cấp cao cho bài toán điều khiển độ ấm nhà kính

Documents