Research On Speed Adaptive Control Of Servo System

In the field of industrial robots and CNC machine where servo systems are widely used,servo motors often work under the condition of variable load and frequent speed change.Therefore,the moment of inertia and load torque of the servo system will change significantly.Temperature rise caused by severe operating conditions will affect the value of inductance,permanent magnet flux linkage and resistance of PMSM.When the motor parameters change,the controller parameters need to be re-tuned according to the changed motor parameters,so that the motor has good dynamic and static performance.the parameters of speed controller are related to the inertia and PM flux,so the identification of inertia and flux is the key to realize the adaptive speed control of servo system.Model reference adaptive method(MRAS)is widely used in inertia identification due to its simple algorithm and easy implementation.The traditional model reference adaptive method(Landau discrete time method)has the problem that it can not take into account both the speed and accuracy of identification,In order to obtain more accurate inertia identification values,only the adaptive coefficients can be reduced,but the identification speed will be greatly reduced.In this paper,a landau continuous time(MRAS)inertia identification method based on continuous model is proposed,the inertia identification algorithm based on continuous model is used to illustrate the problem that traditional methods are susceptible to noise.To solve this problem,the same first-order low-pass filter is used to filter the speed and torque current at the same time.The introduction of the filter is equivalent to increasing the degree of freedom of adjustment,and the selection of adaptive coefficient and filter time constant is qualitatively analyzed.Finally,a larger adaptive coefficient and a smaller filter time constant can be selected to greatly improve the accuracy of inertia identification on the premise of ensuring identification speed,so as to quickly and accurately track the change of the actual inertia.Moreover,this method is applicable to the rectangle wave,triangle wave and sine wave speed,and has wide applicability.Generally,the back EMF method is used to measure the PM flux linkage.Although this method has high measurement accuracy,it belongs to the off-line identification method and can not track flux changes quickly.In order to solve the problem of state equation under-rank in on-line identification,most scholars have proposed to apply disturbance current i_d<0,but this will affect the operation state of the motor,it also affects the identification accuracy of flux linkage,and the methods proposed by most scholars are limited to the flux linkage identification of SPMSM.In the process of identification,the non-linearity of the inverters should also be considered and compensated,which is cumbersome.In this paper,a POPE-based parameter identification method is proposed.The accurate estimation of flux linkage can be realized by injecting the same size and opposite direction angle into the electrical angle and measuring the corresponding d-axis voltage and angular velocity under the control of i_d=0,and the identification of flux linkage is not affected by the non-linear factors of inverters(VSI),nor by the resistance and inductance of PMSM.The whole process of flux identification only needs to adjust a convergence factor to fit the flux curve,and the method is simple.Finally,simulation and experimentation were performed on a SPMSM and a IPMSM.Experimental results show that this method is suitable for both SPMSM and IPMSM.The identified flux can update the torque constant and further improve the accuracy of inertia identification.The identification method proposed in this paper can quickly track the change of inertia and flux,and then the identification value can be used to modify the parameters of the speed controller to realize the speed adaptive control of the servo system.