Encoderless Cascaded Predictive Control For AC Motor Drives Systems

As a key link in energy conversion,the AC motor drive system plays a significant role in the intelligent reform of traditional energy and the efficient application of renewable energy.The high-performance control strategy is the core and brain of the electric drive system,which directly affects the control accuracy,operational reliability and energy transmission efficiency of the system.With the development of AC electric drive system control strategies,the first and second generation control strategies,which are represented by vector control and direct control respectively,have become the mainstream of industrial applications.As a new high-performance control strategy,predictive control has become a research hotspot in this field because of its model predictability,high dynamics,multi-objective and multi-constraint flexibility,and straightforward structure.However,it still faces challenges in industrial applications.This research focuses on two key scientific challenges:(1)In multi-objective control,predictive control needs to weigh the weighting factors between multiple objects.The complex design of weight coefficients restricts its further development;(2)Predictive control strongly relies on accurate speed feedback signals.In certain industrial scenarios,such as offshore oil and gas mining platforms,the converter needs to drive the motor via a long cable.The speed feedback channel is susceptible to interference,which will affect the system’s control performance.In severe cases,the system may become unstable.Focusing on advanced control strategies of AC electric drive systems,this research first systematically compares and analyzes the control strategies of AC electric drive systems.Motor types include asynchronous motors and permanent magnet motors.Control strategies include vector control,direct control and predictive control.Among predictive control,the predictive torque control and predictive current control are studied.The paper establishes a comprehensive AC electric drive system control system,and further verified that predictive control has the advantages of high dynamics,multi-objective flexibility,and simple and direct structureAiming at the first key scientific problems,this research proposes an improved cascaded predictive control method,which eliminates the complicated design of weight coefficients of predictive control.Each cycle dynamically uses the interaction error between the targets to evaluate the cascade optimization sequence,which improves the defect of the fixed optimization sequence in the traditional cascade predictive control.By adopting the target real-time reference value to define the error evaluation item,the problem of poor adaptability of the original cascaded method under multiple working conditions is overcome.Relevant simulation results and comparative analysis verify the effectiveness of the proposed method.Aiming at the second key scientific problem,this research proposes an encoderless cascaded predictive control based on sliding mode observer.Based on the principle of sliding mode control,a sliding mode observer is designed to estimate the speed signal,which overcomes the problem of the system’s strong dependence on high-precision speed feedback.The proposed method has stronger anti-disturbance.At the same time,the proposed method combines the aforementioned cascaded predictive control strategy,retains the advantage of multi-objective flexibility of predictive control,and enhances the adaptability of the original method under multiple operating conditions.Relevant simulation results and comparative analysis of the control effect of each method under rotation speed disturbance and different working conditions verify the effectiveness of the proposed method.