Introduction to Neural Networks and Fuzzy Logic Lecture

Introduction to Neural Networks and Fuzzy Logic Lecture 13 Dr. -Ing. Erwin Sitompul President University http: //zitompul. wordpress. com 2 0 1 7 President University Erwin Sitompul NNFL 13/1

Fuzzy Logic Fuzzy Control Solution: Homework 12 President University Erwin Sitompul NNFL 13/2

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) n Rule 1: 2: 3: 4: IF IF Rule 5: IF FC with 5 Rules level is error is okay, zero, THEN valve is no change. level is THEN valve is open fast. error is low, positive, THEN valve is open fast. level is THEN valve is close fast. error is high, negative, THEN valve is close fast. level is rate israte negative, error is okay zero AND error is positive, THEN valve is open slow. level is rate israte positive, error is okay zero AND error is negative, THEN valve is close slow. error = reference – level rate of error = – rate of level President University Erwin Sitompul NNFL 13/3

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) 1 st Set of Membership Functions positive negative positive fa st ow ge en en sl op 1 op ch o n cl os e sl fa se cl o zero 1 – 4 – 0. 5 0 0. 5 4 Rate of level error [cm/s] 5 an 0 4 Level error [cm] st – 5 – 4 zero 1 ow negative – 30 – 20 – 10 0 10 20 30 Valve control signal [%/s] President University Erwin Sitompul NNFL 13/4

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) President University Erwin Sitompul NNFL 13/5

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) 2 nd Set of Membership Functions positive negative 5 positive fa st ow ge en en sl op 1 op ch o n cl os e sl fa se cl o zero 1 – 4 – 0. 5 0 0. 5 4 Rate of level error [cm/s] an -1 0 1 Level error [cm] st – 5 zero 1 ow negative – 30 – 20 – 10 0 10 20 30 Valve control signal [%/s] President University Erwin Sitompul NNFL 13/6

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) President University Erwin Sitompul NNFL 13/7

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) 3 rd Set of Membership Functions positive negative positive fa st ow en op en sl an ch o 1 op cl os e sl fa se cl o zero 1 – 4 – 0. 5 0 0. 5 4 Rate of level error [cm/s] ge 10 n –? 0 ? Level error [cm] st – 10 zero 1 ow negative – 30 – 20 – 10 0 10 20 30 Valve control signal [%/s] President University Erwin Sitompul NNFL 13/8

Fuzzy Logic Fuzzy Control Solution: Homework 12 (Cont. ) President University Erwin Sitompul NNFL 13/9

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – e u y Fuzzy P Controller President University Erwin Sitompul NNFL 13/10

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – e y u Fuzzy PD Controller President University Erwin Sitompul NNFL 13/11

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – e y u Fuzzy PID Controller • Weakness: too many rules President University Erwin Sitompul NNFL 13/12

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – e y Du Fuzzy PD+I Controller President University Erwin Sitompul NNFL 13/13

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – u e y ++ Du Fuzzy PD+I Controller President University Erwin Sitompul NNFL 13/14

Fuzzy Logic Fuzzy Control PID-like Fuzzy Controllers r + – u e y ++ Fuzzy PD+I Controller President University Erwin Sitompul NNFL 13/15

Fuzzy Logic Fuzzy Control End of the Lecture President University Erwin Sitompul NNFL 13/16