top of page
Writer's picturelms editor

Neural Network Control of Shunt Active Filter in MATLAB

Neural Network Control of Shunt Active Filter in MATLAB

This video explain the working of shunt active filter in MATLAB. The shunt active filter control by neural network concept. the Comparisons of Neural network control with PI control are analyzed in term of total harmonic distortion (THD) of grid current. THD with PI control is 1.83 % and THD with Neural network control is 1.73 %.



Neural Network Control of Shunt Active Filter in MATLAB

Neural networks have been found to be very effective in many applications, including control of electrical systems. One such application is the control of shunt active filters. In this article, we will explore how neural networks can be used to control shunt active filters in MATLAB.

Introduction

Shunt active filters are used to mitigate the effects of harmonics in electrical systems. These filters are composed of a power electronic converter and a control system. The control system is responsible for controlling the converter in order to maintain a specified current waveform. Traditional control methods for shunt active filters can be complex and difficult to implement. However, neural networks have been shown to be effective in controlling these filters.

Shunt Active Filters

Before we dive into the details of neural network control, let's first take a look at how shunt active filters work. Shunt active filters are connected in parallel with the load in order to eliminate harmonic distortion. The power electronic converter in the filter is controlled by a reference current, which is generated by the control system. The control system measures the load current and compares it to the reference current. Based on this comparison, the control system adjusts the converter to maintain the desired current waveform.

Traditional Control Methods

There are several traditional control methods that can be used for shunt active filters, such as proportional-integral (PI) control and hysteresis control. PI control is a common control method that uses feedback to adjust the output of the converter. Hysteresis control, on the other hand, uses a binary decision-making process to determine whether the converter should be turned on or off.

While these traditional control methods can be effective, they can also be complex and difficult to implement. This is where neural networks come in.

Neural Network Control

Neural network control involves using a neural network to generate the reference current for the shunt active filter. The neural network is trained using a set of input-output pairs. The input to the neural network is the load current, while the output is the reference current.

The neural network is trained using a backpropagation algorithm. This algorithm adjusts the weights and biases of the neural network in order to minimize the error between the desired output and the actual output.

Implementation in MATLAB

MATLAB is a powerful tool for implementing neural network control of shunt active filters. MATLAB provides a Neural Network Toolbox that includes several neural network architectures, training algorithms, and performance functions.

To implement neural network control in MATLAB, we first need to define the neural network architecture. This involves specifying the number of layers, the number of neurons in each layer, and the activation function for each neuron.

Once the architecture is defined, we can train the neural network using the backpropagation algorithm. We need to provide a set of input-output pairs for training. The input is the load current, while the output is the reference current. We also need to specify the training parameters, such as the learning rate and the number of epochs.

After the neural network is trained, we can use it to generate the reference current for the shunt active filter. We simply feed the load current into the neural network and obtain the output, which is the reference current.

Conclusion

Neural network control is an effective method for controlling shunt active filters. It offers several advantages over traditional control methods, such as ease of implementation and improved performance. MATLAB provides a powerful tool for implementing neural network control of shunt active filters.

FAQs

  1. What is a shunt active filter? A shunt active filter is a device that is used to eliminate harmonic distortion in electrical systems.

  2. What is neural network control? Neural network control involves using a neural network to generate the reference current for the shunt active filter

28 views0 comments

Comments


bottom of page