Research On Speed Controller For PMSM Servo System

As one of the three core components of the industrial robot,servo motors and their driving system need to be further optimized.In this paper,some speed control strategies for high performance permanent magnet synchronous motor(PMSM)servo system are studied.PI controller is mostly used in servo control system,is difficult to meet the requirements of tracking performance and robustness for PMSM,which is a multi-variable,nonlinear controlled object with many uncertainties both interior and exterior.Many nonideal elements and filters are considered when designing the traditional PI controller for speed loop,leading to a complicated relationship between parameters of the controller and the performance of the control system.To solve this problem,a compound PI controller designed with the equation of state is proposed in this paper.When we design a traditional controller,the structure of the controller is selected first,and then the performance of the system using the controller is analysed.However,the design method of PI controller proposed by this paper is just the opposite.The attenuation law of tracking error is designed first,based on which the structure of the controller is deduced.The compound PI controller proposed in this paper has a simple structure,and is adapative to varies input.Parameters of the controller have a simple relationship with the performance of the control system,thus are easy to set.By using the compound PI controller,the speed control system is corrected to be a two-order system with overshoot in step response even when the damping ratio is unit.Futhermore,the compound PI control system is an one-degree-of-freedom control system,which can not adjust the tracking performance and disturbance rejection property independently.The speed control system has good tracking performance for step response when using IP controller,i.e.,there is no overshoot in step response,but has also a poor tracking performance for continuously varying input.To solve the problem,a variable structure PI controller is proposed in this paper,acts as an IP controller when the input is a step signal and a PI controller when the input is a continuously varying signal.The secret to get a variable structure PI controller is the use of input differential feedforward and clamping integrator.In order to improve the tracking performance of conventional IP controller,an optimal IP controller is proposed at first.The optimal IP controller helps the control system obtain fastest response with zero overshoot in step response when the damping ratio is unit.However,the optimal IP control system still has a poor tracking performance for continuously varying input.Similarly,a PI controller with different reference gain is propsed.The controller is equivalent to IP controller when reference gain is zero and PI controller when the reference gain is one.As PI control system have steady-error-free tracking performance for continuously varying input under the condition of no disturbance,a PI controller based on arranged transition process is proposed to eliminate overshoot in step response.Conventional 2DOF controller is designed through transfer function without clear physical meaning.To solve the prolem,internal model control with clear physical meaning is used to design disturbance observer(DOB).In this paper,a P+DOB controller widely used in servo control field is analyzed according to internal model control method and is compared with PI controller.Nonmatrix form full-order state observer and reduced-order state observer are deduced by using the design method based on state equation.Observer constructs a model of the system under study and forces the output of the model tracking that of the system,thus controller is also used in observer.The design process of observer is simplified by using the design process of controller for control system.The tracking performance of conventional 2DOF control system is not affected when regulating disturbance rejection ability,which will be affected when adjusting the tracking performance,i.e.,the two characteristics are not full decoupled,resulting in a complicated parameter setting process.To solve the problem,full decoupled P+RSO 2DOF controller and P+FSO 2DOF controller are presented.With full decoupled 2DOF controller,the bandwidth of observer only affects disturbance rejection ability and the proportional coefficient influences tracking performance respectively.The two propertys of the system can be regulated independently,thus simplifies the parameter setting process.Although 2DOF controller can improve the disturbance rejection ability,in other words,achieve the stability robustness of the system,it does not improve the parameter robustness of the system.Some articles analyzed the influence of parameter varying on dynamic performance of 2DOF control system,which is not comprehensive.In this paper,the effect of state coefficient and control gain is analyzed through state equation,and it is founded that state coefficient variation has much less influence on dynamic performances of the system than control gain variation.Hence on account of the first element leading to speedloop control gain variation,that is inertia variation,an inertia identification method is utilized based on orthogonal principle in the paper,and then adjusting the parameters of PI controller and P+ESO controller in real time according to the identificated inertia,which can improve the adaptivity to inertia variation of the two control systems.