Fault classification location and detection in power system using neural network
Welcome to LMS Solution! In today's discussion, we will delve into fault detection, classification, and location finding in a three-phase three-bus system using neural networks. This model has been developed to effectively identify faults, classify them, and pinpoint their locations in a power system with three buses.
System Overview
The system comprises three buses: Bus 1 and Bus 2 act as the source, while Bus 3 serves as a load mass. A transmission line connects Bus 1 and Bus 3, where faults are introduced for analysis. The fault detection involves measuring voltage and current in RMS (root mean square) and zero sequence current and voltage from Bus 1.
Data Collection for Fault Conditions
To facilitate fault detection, various fault models are created, such as line-ground fault, double-line ground fault, double-line fault, and triple-line fault. These models are executed with different line lengths (10 km, 5 km, and 1 km) to simulate various fault scenarios.
The collected data includes voltage and current values (VABC) and is categorized into normal and faulty cases. Target values are assigned based on fault type and location, creating a binary representation. The program then collects the data for training the neural network.
Neural Network Training
The input data (X data) includes RMS voltage and current for Bus 1 and zero sequence voltage and current. Target data (D data) comprises binary representations of fault types and locations. The neural network is trained using these datasets to match the collected input and target data.
After training, the model's performance is evaluated using the R-value, which should be close to 1 for a successful match. The trained neural network is then exported for use in Simulink.
Simulink Implementation
In Simulink, the neural network is integrated into a fault detection and classification system. The output is decoded to interpret fault types and locations. The simulation is carried out with various fault scenarios, and the results are analyzed for accuracy.
Simulation Results
The model is tested with AB ground fault scenarios at different line lengths (4 km, 6 km, 9 km), and the fault detection, classification, and location results are observed. The system effectively identifies faults, classifies them (e.g., AB ground fault), and accurately locates the fault within the simulated power system.
Conclusion
This video demonstrates the successful implementation of fault detection, classification, and location finding using a neural network in a three-phase three-bus system. The Simulink simulation provides valuable insights into the model's effectiveness in real-world scenarios.
Comments