top of page

Fault classification location and detection in power system using ANFIS

Fault classification location and detection in power system using ANFIS

Introduction:

Welcome to our discussion on fault deduction, location classification, and face science, which involves the adaptive neuro-fuzzy inference system (ANFIS). In this blog post, we will explore the integration of neural network properties with the fuzzy inference system to create an efficient system known as ANFIS. The primary focus is on using a 9-phase concept to detect faults, classify them, and pinpoint their location in power systems.


System Overview:

The developed model comprises three buses: Bus 1, Bus 2, and Bus 3. Bus 1 and Bus 2 have source feeders rated at 11 kilovolts and 30 MVA. Bus 3 is connected to Bus 1 via a distribution line with a 10-kilometer length. Faults can be introduced at various points in the system, such as underground faults, doubling faults, and more. Transformers are used to convert the 11-kilovolt input to 400 volts for distribution to customers.

Data Collection:

To train the adaptive neuro-fuzzy system, data is collected for normal and fault conditions. The input data includes the root mean square (RMS) voltage and current of Bus 1, as well as zero-sequence voltage and current. The target data consists of four binary numbers, where the first two bits are used for fault detection and classification, and the last two bits indicate the fault location.

Program Execution:

A program is written to collect data for different fault conditions and fault locations, varying transmission line lengths. The data is used to train the ANFIS model for fault detection, classification, and location.

Model Training:

Two ANFIS models are developed—one for fault classification and another for fault location. The classification model uses a fuzzy clustering concept, while the location model employs grid partitioning and training. The trained models are stored for later use.

Model Implementation:

The fault detection and classification model takes RMS voltage and current, along with zero-sequence voltage and current, as inputs and outputs a binary classification. The fault location model takes RMS voltage as input and outputs the precise fault location.

Simulation Results:

The models are tested with various fault scenarios, such as ABC ground faults, ABC triple faults, and more. The fault detection, classification, and location results demonstrate the accuracy of the ANFIS system.

Conclusion:

The adaptive neuro-fuzzy inference system proves to be an effective approach for fault detection, classification, and location in power systems. With a 90-91% accuracy rate, the system demonstrates its reliability in identifying and locating faults. This blog post provides an overview of the methodology, data collection, model training, and simulation results.

14 views0 comments

Comentarios


bottom of page