Safety Optimized Coordinated Control Of Dual-Motor Systems Based On Reinforcement Learning

Multi-motor coordinated control systems are widely used in modern industry and are of great significance to improve the flexibility and load capacity of production systems.Among them,dual-motor systems are the common type,and the safe and reliable operation of the entire system is dependent on speed synchronization and torque balancing between two motors.However,the characteristics of large variation of motor system parameters,time-varying load,and two-time-scale dynamic coexistence make it difficult to achieve high-precision coordinated control of dual-motor systems.In addition,in the networked dual-motor control system,the occurrence of cyber attacks directly threatens the safe and stable operation of the system.Therefore,this thesis investigates the coordinated control of dual-motor systems based on reinforcement learning algorithms for dual-PMSM systems,and the main research work is as follows.(1)A reinforcement learning-based optimal coordination control method for rigidly connected dual-motor systems is proposed.Firstly,the mathematical model of the dual-PMSM system is built using the conventional master-slave control structure and PI controllers.Then,by using the theory of output regulation and optimal control,an optimal coordination controller is designed to solve the problem of disturbance suppression under the change of external load of the system.Finally,aiming the uncertainty of the model parameters of the system and the coexistence of fast and slow dynamics,a reinforcement learning algorithm independent of the model parameters is proposed to solve the controller gain.The proposed control approach can enhance the dual-motor system’s tracking and synchronization capabilities,suppress the disruption of unidentified time-varying loads,and prevent the effects brought on by parameter uncertainty.According to simulation results,the dual-motor system’s torque synchronization and speed tracking capability can both be successfully enhanced by the proposed control strategy.(2)To address the situation that the system suffers from FDI cyber attack,a flexible connected dual-motor safety coordination control method is proposed.Firstly,the mathematical model of the dual-PMSM under the cross-connected control structure is established,and the time-scale parameters are extracted to construct a dual time-scale model.Second,the design of the safety coordination controller and the creation of the optimized performance index are based on the LQR optimal control theory.Then,the FDI cyber attack model is established and the system model after the attack is reconstructed.Finally,according to the FDI cyber attack model,a reinforcement learning algorithm using iterative learning of the tampered data is designed to solve the security coordination controller gain.The simulation results verify that the proposed control method can avoid the impact of the FDI cyber attack on the system and improve the synchronization control accuracy of the system at the same time.(3)The networked dual-motor system experimental platform is designed to experimentally verify the proposed control method.Firstly,the overall topology of the networked dual-motor system experimental platform is designed,and the hardware and software components of the experimental platform are described in depth.Then,the experiment implementation procedure is described step-by-step by taking the experiment for safe coordinated control of flexible connected dual-PMSM under FDI cyber attack as an example.Finally,the experimental verification of the design of the control method proposed in the second and third chapters of the article is carried out,and the experimental results are analyzed and compared.The experimental results demonstrate that the proposed control method significantly outperforms the conventional methods in terms of enhancing the dual-motor system’s performance.