Intelligent torque diagnosis method of nuclear electrical valve actuator based on WOA-ELM algorithm

Electric valve actuator is a kind of equipment that can accept external control commands, drive the motor and reducer to transmit power to the valve, and servo control the valve opening. Based on the time series performance of electric valve faults, the fault classifier models adopted the whale optimization algorithm (WOA) optimized extreme learning machine (ELM) for achieving intelligent torque information diagnosis. Firstly, the bridge circuit and laser rangefinder is used for the experiment, and the changes of torque, valve position, opening or closing time are considered. Then, the WOA is adopted to optimize ELM and establish diagnostic models to identify and diagnose fault types. Finally, to improve the learning speed and generalization performance of extreme learning machine, WOA is used to optimize the hidden layer neuron threshold and input weight of ELM. The experiments show that the fault diagnosis accuracy of electric valve actuator based on WOA and WOA-ELM is more than 95.71% and 92.8%, respectively. The method proposed in this paper has guided significance of the fault detection and safe operation of electric valves in nuclear facilities.