A fuzzy neural network-based automatic fault diagnosis method for permanent magnet synchronous generators

Area of Science:

• Electrical Engineering
• Artificial Intelligence
• Control Systems

Background:

• Automatic fault diagnosis is crucial for industrial machinery reliability.
• Permanent magnet synchronous generators (PMSGs) require effective diagnostic methods.
• Existing methods may face challenges with complex structures and parameters.

Purpose of the Study:

• To propose a fuzzy neural network-based automatic fault diagnosis method for PMSGs.
• To enhance the accuracy and robustness of fault diagnosis systems.
• To address real-time computation and fault tolerance in control systems.

Main Methods:

• Combining fuzzy decision theory with deep learning for PMSG fault diagnosis.
• Utilizing particle swarm optimization to tune network parameters.
• Employing fuzzy C-means clustering to identify fault data centers.
• Reconstructing the network model using samples near clustering centers.
• Developing a Takagi-Sugeno (T-S) fuzzy neural network diagnosis strategy.

Main Results:

• The proposed T-S fuzzy neural network diagnosis strategy shows significant improvement.
• The fault diagnosis model demonstrates effectiveness and accuracy on PMSG fault data.
• The parallel processing capability of fuzzy neural networks enhances fault tolerance and robustness.
• MATLAB/Simulink simulations verify the system's design and performance.

Conclusions:

• The developed fuzzy neural network method provides an effective and accurate solution for PMSG fault diagnosis.
• The approach offers improved fault tolerance and robustness, suitable for real-time control systems.
• The integration of fuzzy logic and deep learning presents a promising direction for industrial diagnostics.