PID Parameter Tuning Method for Steam Generator Level Control Based on Ant Colony Algorithm

Abstract Steam generator levels in pressurized water reactor nuclear power plants at home and abroad use the classical serial PID control scheme. At present for the steam generator level control system control strategy method research has been fruitful, but when the operating conditions change, the original PID control effect may no longer meet the demand, so the need for PID control parameter calibration of the problem is still very few relevant methods of research. For this reason, this journal proposes a method to train the RBF neural network model based on the historical data of the system, to improve the optimization data, and establish an accurate mathematical model of the system based on the optimized data by the least squares method for the PID parameter tuning method of the subsequent ACO algorithm. The RBF neural network is trained through historical data to obtain the feature model. The optimized data is modelled using F-test to determine the order of the system data, and the corresponding differential equations are used for identification, and the identified differential equations are transformed into transfer functions, through which an accurate mathematical model can be identified in the case that the order of the system structure is not clear. Finally, the ACO algorithm is used to adjust the PID parameters according to the identified mathematical model of the system, and the adjusted PID parameters have a good control effect in different working conditions, and the results show that the optimized data identification transfer function is basically the same as the actual system, and the parameter-optimized control system of the steam generator level control system is better than the original control system in terms of the overshooting control indexes, and it achieves the control of the liquid level inside the steam generator under different working conditions. Steam generator level control under different operating conditions. It is concluded that when the model parameters are changed, the PID parameter control effect of the identification model adjusted by the ACO algorithm is better than the original PID control of the steam generator.