1 Fuzzy Logic Workshop Design of Fuzzy Controller for Temperature Chamber  www.aimagin.com

2 Syllabus What is the system in control ? System IdentificationMathematical model and Choice of sampling period. FiO Boards and RapidSTM32 Blockset for Applied on Temperature Control (HIL Test) Matlab Fuzzy Logic Toolbox Finite State Machine Stand-Alone System Implementation ทฤษฎีระบบควบคุม เป็นสาขาหนึ่งของคณิตศาสตร์และวิศวกรรมศาสตร์ ในที่นี้ การควบคุมหมายถึง การควบคุมระบบพลศาสตร์ ให้มีค่าเอาต์พุตที่ต้องการ โดยการป้อนค่าอินพุตที่เหมาะสมให้กับระบบ

3 What is a control system ?Control Signal (input) Output SISO Heater (input) Temperature (Output) In most systems there will be an input and an output.  This block diagram represents that.  (Control system designers and engineers use block diagrams to represent systems.  Get used to them.)  Signals flow from the input, through the system and produce an output. Heater FAN (2 input) Temperature (Output) MISO

4 Controller Sensor Temperature (Output) Set Point Error Heater dErrorFAN Sensor

5 How to Identify the SystemAnalyze the input-output data pairs to fit the parameters in the used model (structure) Unknown System X Y ? ? ? Experimental determination of system model. There are two methods of system identification: Parametric Identification: The input-output model coefficients are estimated to “fit” the input-output data. Frequency-Domain (non-parametric): The Bode diagram [G(jw) vs. w in log-log scale] is estimated directly form the input-output data. The input can either be a sweeping sinusoidal or random signal. Unknown

6 Equivalent DC motor

7 How to Identify the SystemUnknown System X Y two methods 1.Parametric identification 2.Non-parametric identification. Experimental determination of system model. There are two methods of system identification: Parametric Identification: The input-output model coefficients are estimated to “fit” the input-output data. Frequency-Domain (non-parametric): The Bode diagram [G(jw) vs. w in log-log scale] is estimated directly form the input-output data. The input can either be a sweeping sinusoidal or random signal.

8 System identification processPrior knowledge Experiment Design In building a model, the designer has control over three parts of the process Generating the data set ZN Selecting a (set of) model structure (1st or 2nd order) Selecting the criteria (least squares for instance), used to specify the optimal parameter estimates There are many other factors that influence the final model, however, this course will focus on these three factors and the method for (recursive) parameter estimation Data Choose Model Set Choose Criterion of Fit Calculate Model  Validate Model 

9 Unknown System X Y Output Input

10 Model validation ^ u(t) y(t) Model q ^ u(t) u(t) Plant q y(t) d(t)

11 Prepare plant for system identification

12 Temperature Control Module

13 Initial preparations Connect CN9(GND) to GND Fio BoardConnect Temperature sensor from CN2 to Analog input Select PWM Output pins (B6, B7, B8, etc.) form Fio to CN3,CN4 and CN6 Connect 12 VDC to connector J1

14 FiO Board Std Temp. Board

15 FiO Board Lite Temp. Board

16 ADC Configuration

17 PWM Output pins

18 Temperature sensor modulerefer on P.8 Temperature sensor module

19 refer on P.65 Then

20 refer on P.66

21 Heater and Fan control modules

22 refer on P.66

23 10msec = 100Hz FAN2 FAN1 10% Duty Cycle HALOGEN BULB 40% Duty Cycle

24 refer on P.10 Design of Display

25 refer on P.11

26 Let’s try Send the command for control heater and fan on Target from HOST

27 refer on P.16 Design of Display

28 Let’s try Send the temperature from Target to HOST and display on the Matlab Scope

29 Filter Design

30 Filter design First order Low pass filter

31

32 Continuous C2D Discrete Define cutoff frequency Define sample period

33 ‘ jω’ is substituted for ‘s’

34 Now, since we know that the cutoff frequency, ω , occurs at Magnitude=0.707, this can be substituted into above to get Solve for wc

35 we define cutoff frequency at 20 Hz, then

36 we define cutoff frequency at 5 Hz, then

37 Fast Fourier Transform (FFT)

38

39 System identificationIntroduction

40 SYSTEM IDENTIFICATIONThe System Identification Problem is to estimate a model of a system based on input-output data. System disturbance (not observed) v(t) y(t) u(t) output (observed) input (observed) continuous

41 Typical to be used as input for identificationStep PRBS Sinusoidal

42 Parametric ID of step responseFirst order process with dead time Most common industrial process model Response to a control step applied

43 refer on P.17 Unknown System X Y

44 refer on P.19

45 Import data to system identification toolbox

46 Matlab Command for open system identification toolbox

47 refer on P.17

48 Unknown System X Y Y X

49 temp = simout(:,2) ตัดข้อมูล 2000 ถึง 6000 duty = simout(2000:1:6000,1) tem = temp(2000:1:6000)

50

51 Choice of sampling period

52

53 Applied in many appliances.Researchers have developed a theory for the same concept as the FUZZY set stand out. Applied in many appliances. Air conditioning Washing machine Rice cooker. And more… Neuro Fuzzy Rice Cooker Fuzzy Washing Machine

54 Fuzzy Set Theory 38.7°C 38°C 40.1°C 41.4°C 42°C 39.3°C 38.7°C 38°CConventional (Boolean) Set Theory: 38.7°C 38°C “Strong Fever” 40.1°C 41.4°C Fuzzy Set Theory: 42°C 39.3°C 38.7°C 38°C 37.2°C 40.1°C 41.4°C 42°C 39.3°C “Strong Fever” “More-or-Less” Rather Than “Either-Or” ! 37.2°C

55 Fuzzy Set Definition Discrete Definition:µSF(35°C) = 0 µSF(38°C) = 0.1 µSF(41°C) = 0.9 µSF(36°C) = 0 µSF(39°C) = 0.35 µSF(42°C) = 1 µSF(37°C) = 0 µSF(40°C) = 0.65 µSF(43°C) = 1 Continuous Definition:

56 Linguistic Variable A Linguistic Variable Defines a Concept of Our Everyday Language! 76 77 78 79 80 81 82 83 84 Low temp Normal High temperature Rise temperature 1 0.7 0.2

57 Fuzzy control provides a formal methodology for representing, manipulating, and implementing a human’s heuristic knowledge about how to control a system. Plant Fuzzification Defuzzification Inference Mechanism Knowledge Base Reference input Output Command variable

58 Basic Elements of a Fuzzy Logic SystemCommand Variable (Linguistic Value) Measured Variable (Linguistic Value) 2. Fuzzy-Inference Linguistic Value Numerical Value 1.Fuzzufication 3.Defuzzification Measured Variable (Numerical Value) Plant Command Variable (Numerical Value)

59

60 Choice of Sampling IntervalAnother important aspect in sampled data control systems is the choice of sampling intervals.

61 Plant testing on Target

62 Plant testing on Host

63 2 1

64 1

65

66

67

68

69

70

71 Process Value Set Point

72 dError 0 Error + Error 0 dError - dError + Error - dError 0

73 CASE: 1 Error 0 dError 0

74 CASE: 2 dError - Error 0

75 CASE: 3 dError + Error 0

76 CASE: 4 Error - dError 0

77 CASE: 5 dError - Error -

78 CASE: 6 dError + Error -

79 CASE: 7 Error + dError -

80 CASE: 8 dError 0 Error +

81 CASE: 9 Error + dError +

82 Fuzzy Control Rules Z Z Z Z Z PS PS PM PL Zero Positive Small Z PSPositive Medium Positive Large PM PL

83 Fuzzy Control Rules Zero Positive Small Z PS Positive MediumPositive Large PM PL

84

85

86

87

88

89

90