Method for Implementing Sensorless Feedback for Internal Drive Control Loop Using Kalman Filtering Algorithm

The paper discusses the implementation of a three-circuit control system for the angular position of the drive output shaft. An armature current sensor, an angular velocity sensor of the DC motor output shaft, and an angular position sensor of the drive output shaft are used as feedback sensors. As a rule, the introduction of all of these sensors into a product that has already been put into operation cannot be carried out, in addition, one or another sensor of the internal drive control loop may fail. The three-circuit control system, in turn, has significant advantages when controlling the angular position of the drive output shaft in relation to the single-circuit control system. The paper shows that three-loop control can be implemented even in the absence of feedback sensors in the internal control loops: instead of measuring the controlled variables, in this case they are evaluated. The purpose of the study is construction of a three-loop control system in the absence of sensors of internal control loops using the assessment of controlled variables by a discrete Kalman filtering algorithm. Materials and methods. To implement the proposed algorithm, a mathematical model of the drive has been developed, taking into account each control loop and all controller parameters. Modeling of a system with feedback sensors or evaluation of variables of internal control loops is carried out in the MATLAB Simulink software environment. Research results. Simulation shows that three-loop control of the angular position of the drive output shaft with one physical feedback sensor of the external control loop can be achieved by estimating the variables of the internal control loops of the drive with a discrete Kalman filter. In this case, an increase in the overshoot of the angular position is observed compared to the option of constructing a system with sensors in feedback circuits. To reduce overshoot in the drive transfer function, the damping coefficient is increased and its most optimal value is selected. The quality of regulation in the resulting system does not deteriorate. Conclusions. The implementation of sensorless feedback is possible, but it requires additional computational costs, which makes the computer more expensive. Also, inaccurate correspondence of estimates of control variables of internal loops to their actual value leads to the need to correct the calculated parameters of the regulator of the external control loop of the drive.