The Study On Parameter Self-tuning For Full-digital DC Servo System

In this paper, the physical platform of a full-digital DC servo system is established,and made into a product. The author takes part in the hardware design, and designs thesoftware independently. In product testing and engineering applications, the author foundthat the set of regulator parameters, fixed in the system, can not work well in allengineering conditions. But for different motors or under different power supply, the fixedparameters will probably lead to an abnormal performance or the disability to reach theideal performance. So it is needed to be tuned by people. In order to reduce manpowercost and improve the degree of automation, the research and development of parameterself-tuning become a necessity.In order to implement the function of parameter self-tuning, regulator parameterdesign method and motor parameters identification method are mainly researched in thispaper. In research of regulator parameter design method, the mathematical models of eachpart are established. And current loop, speed loop and position loop control framework arebuilt orderly from inside to the outside on base of this. The formulas of regulatorparameter tuning are obtained according to the engineering design method. Because thephotoelectric encoder of the speed sensor has the feature of a small detection delay, it iseasy to cause current oscillation when the engineering design method is used to design theparameter of speed regulator. Two kinds of solutions are proposed in this paper, and theexperiments prove the effect is good. In the study on motor parameter identification,Landau discrete-time recursive algorithm, which based on the adaptive thought and thePopov hyperstable theory, is used, and the experiments prove the results are accurate. Atlast the regulator parameters are worked out by the use of motor parameter identificationand the formulas of regulator parameter tuning. A lot of experiments are done in theexperimental platform, and it is proved that the self-tuning parameters can bring asatisfactory control performance.