Application of Physical Model of Secondary System in Fault Prediction of Wind Farm

A multi-layer protocol adapter card based on CAN bus and various buses is proposed to realize the fault prediction and fault monitoring of various wind turbine PLC. The lower machine is responsible for processing the data collected from the Ethernet network in the field through Modbus/TCP, and remote monitoring of the wind turbine through the industrial configuration software. This paper introduces the implementation scheme of each bus interface in multi-protocol adapter. A classical particle swarm optimization method (GDSPSO) under Gaussian perturbation is established. The optimization performance of genetic algorithm is improved by improving the parameter of genetic algorithm to be fixed. The expression of learning factor changing with the number of iterations is introduced to improve the searching ability of the algorithm. The predictive model is built using the learning parameters optimized by GDSPSO. This method has good forecasting performance and high accuracy. The developed pre-processing multi-protocol adapter has a good reference significance for remote monitoring of large-scale wind turbines.