| As a basic component of intelligent equipment such as robots,CNC machine tools,medical equipment,and electronic manufacturing equipment,the permanent magnet synchronous motor servo system directly determines the comprehensive performance indicators of intelligent equipment.However,due to the increasing comprehensive performance requirements of application objects,it is urgent to expect that the servo system control strategy can not only effectively deal with multiple constraints such as state coupling,parameter time-varying,load disturbance,actuator saturation,and delay,etc.,to achieve closed-loop stability of the system,but also need to ensure that the system obtains the desired metrics on various performances（such as robustness,response speed,and tracking accuracy）.For the above-mentioned multi-constraints and multi-objectives,the traditional servo control method is difficult to meet the requirements of high-speed and high-precision motion control.Therefore,this thesis conducts systematic research on the aspects of fuzzy controller design and fuzzy control gain optimization.And,considering the uncertainty of the servo system structure and parameters,the model-based and model-free multi-objective performance-constrained fuzzy control strategies are proposed according to whether the controlled system model can be obtained.They aim to make the permanent magnet synchronous motor servo system run smoothly while satisfying various performance constraints such as accuracy,response speed,and anti-disturbance by designing appropriate fuzzy control laws.The main work and contributions of the thesis are introduced as follows.Under the influence of time-varying bounded disturbance,actuator saturation,unmodeled dynamics,and other factors,the servo system with an identifiable model has the problem that it is difficult to take into account stability,robustness,and tracking performance.Therefore,a model-based model control method is proposed,i.e.,the adaptive Pareto optimal fuzzy control.First,based on the T-S fuzzy model,a fuzzy state feedback controller and a fuzzy decision maker are constructed to ensure that the system can automatically switch to a stable working state under time-varying tasks.Then,a multi-objective optimizer is designed to pre-tune the control gain,and the Pareto optimal H2/H∞performance is obtained while satisfying the actuator saturation constraints.Finally,to reduce the conservatism of the system,the design conditions of the multi-objective optimizer are relaxed based on membership function correlation analysis and parameterized linear matrix inequalities.Aiming at the stable tracking problem of the servo system when there is no accurate model to rely on,a model-free adaptive fuzzy PI control method is proposed to deal with constraints such as actuator saturation,input delay,and uncertain disturbances to ensure that the closed-loop system is stable while achieving predetermined H∞robust performance.Among them,a composite control structure of prediction-fuzzy-PI is designed,including an anti-saturation PI controller which drives the system to track the target contour,an adaptive fuzzy tuner that schedules the control gain online to deal with the uncertain disturbance of the system,and a predictive function control module which pre-adjusts the control signal to improve the dynamic tracking control performance of the system under input constraints.Meanwhile,the real-time self-tuning input domain of the adaptive fuzzy tuner strengthens the scheduling ability of general fuzzy rules.In addition,with the help of D-decomposition theory,a predetermined setting rule of control gain driven by frequency response data is proposed,which ensures the global asymptotic convergence and H∞anti-interference performance of the servo system.High-performance servo systems are often required to achieve optimal or predetermined values on transient and steady-state performance indicators such as response time,overshoot,or integral error.Given this,a fuzzy control gain optimization algorithm is studied to meet multi-objective performance constraints.For the optimization of the model-based system,a competitive multi-objective bat algorithm is proposed.Specifically,a competitive bat algorithm with fast convergence ability is designed based on the pairwise competition strategy,and combined with the genetic algorithm,the designed tribe competition mechanism achieves a dynamic balance between the population diversity and the convergence.For the optimization of model-free systems,a data-driven constrained bat algorithm is proposed.Specifically,a gradient-based depth-first search strategy is designed to deal with the strict constraints related to the safety of the servo system.Then a boundary constraint processing mechanism and an improved bat algorithm integrated with theεconstraint handling method are proposed to rapidly explore optimal solutions satisfying predetermined performance constraints in a safe feasible region.The experimental results show that the proposed algorithms have significant advantages compared with the representative algorithms in the industry,and are more practical in dealing with multi-objective control gain optimization problems.Based on the self-developed servo driver and the host computer servo tuning platform,the permanent magnet synchronous servo system experiment platforms with a wide range of application scenarios have been built,including a single-inertia rigid connection servo system with an identifiable model,and a double-inertia flexible connection system with uncertain models.Then,the real-time motion control experiments are carried out on the key technology of the proposed fuzzy control strategy for universal tasks and performance objectives.The experimental results show the effectiveness,superiority,and applicability of the proposed strategy. |
PDF Full Text Request