Fuzzy Logic Based Shunt Active Power Filter

A Fuzzy powered DC link Capacitor voltage regulating algorithm

About Project

This Project tries to successfully replicate the process used in a Scholarly Article. The Shunt Active Power Filter uses Fuzzy logic to determine the output current (from shunt) which helps reduce filter the harmonics in a system.

Introduction

What is Harmonics in Electrical System?

In power systems, a harmonic is a voltage or current occurs at a multiple of the fundamental frequency. They are integral multiples of a fundamental frequency. It is often regarded as noise in the power line.

In the above image notice how current is distorted due to voltage harmonics? Well, we are trying to remove exactly this.

Why it is necessary to eliminate Harmonics in the Power System?

It is well known that Current and the voltage harmonics are directly proportional to the noisy power transfer to the Load. Various household equipments are responsible for the harmonics in the power system. The power system harmonics often manifolds the load current. Equipments like lights, LEDs, Television face the adverse consequence if power system harmonics is allowed to exist.

To overcome this power system harmonics, one need to reconstruct the power connection to introduce harmonics filters in the power system.

Why Active Filter?

Active harmonic filters use a neat method where the filter injects compensating harmonics into the system nullifying the effect. The development in this field has become pretty advanced. The following are some of the widely used active filtering methods:

• PWM method

• Voltage source inverting method

• Sampling and Control strategy

• Fuzzy Based Shunt Active filter strategy (most recent and most effective)

Plan Of Attack

Understand the Role of Fuzzy in SAPF

• Used as DC-Link Capacitor Voltage Algorithm

• Capacitor Voltage, Vc is taken as 14v, ( refer ijsert page6 )

• A voltage is measured after the shunt and before Fuzzy, Vm

• Resistor of 0.1 Ohm is taken due to impedence

Crafting Fuzzy Sets

Input Set

There are two sets: Set1, Set2

Set1 : Set of errors (E) are collected between Vc and Vm at a specific point of time t, given by:

                                   E = Vc - Vm

Set2 : Set of errors described by (dE):

                                dE = E(t) - E(t-1)
                                
                                        or
                                        
                                dE = E - d(E)/dt

Output Set

The output set is a crisp value (percentage) of the current produced with effective magnitude: Ic. The range of Ic can be easily calculated from Vc assuming the shunt has a resistance of 0.1 Ohm (caused by impedance).

Since the E range is from 7 to 14, Vc is 14V, Ic can be calculated by:

                                Ic = x ∈ {(Vc-Vm) / R}

Fuzzy Set Ranges

E Range:

The experimentally permitted range is : E = {7 to 14}

dE Range:

The expereimentally feasible range is : dE = {-0.4 to 0.5}

Ic Range:

The range acquired by calculation from above : Ic = {0 to 70}

All ranges are being refered from ijsert-page3

Chosing Fuzzy Inference System

Using SKFuzzy control.ControlSystem which defaults to Mamdani Inference System with defuzzification using centroid

Coding

• Defining universe
• Initializing Antecedents, Consequents fuzzy variables
• Populate the universe with membership functions
• Rule Writing
• Control System + Simulation

Experimental Results

Successfully performed fuzzification and defuzzification using SkFuzzy module.

Below are the observations from the experiment conducted:

• Membership Graph for Input Set E

• Membership Graph for Input Set dE

• Membership Graph for Output Set Ic

• Final Output for E = 12, dE = -0.1 came out to be 45.99Amps

Resources

• Moscow Scientific Reviews

• Ijesrt Scholarly Article

• MDPI-RES Electronics

Contributors

Pratheek & Sourav