The combination of computer simulation of a system and computer based data collection to characterize a real system is useful as a research and application tool. The use of a computer for feedback for the speed control of an electric motor is a system to which this combination may be usefully applied.
Control theory and modeling have been applied to direct current electric motors. Previous work investigating controlling a direct current electric motor with a personal computer utilizing common control techniques is the basis and starting point for this project. The direct current electric motor is the target subject and a personal computer is the feedback control mechanism. In this work, modeling, simulation and analysis of the data were handled through MATLAB software. SIMULINK, which is a subset of MATLAB, was used for simulations.
The initial objective was to model the response of the motor to a step voltage in an attempt to bring the motor to a specified or set point speed of rotation. A standard motor model and manufacturer’s specifications for the Pittman 2800 motor were used in the simulation. It was found that the response predicted by the simulation was much faster than the actual response of the motor. Expecting that the motor was loading the DC amplifier providing the driving voltage, a first order transfer function was included in the model. With adjustment of the time constant to .25 sec, agreement between the simulation prediction and motor response was reasonable. Motor responses with proportional and integral gains in the 1 to 10 range could then be modeled with the simulation. For larger gain values, it was not possible to simulate the motor response and obtain reasonable agreement for both overshoot and rise time. Inclusion of derivative control was found to have little effect and to provide no convincing improvement for the conditions tested.
Improvement in the simulation and motor feedback were obtained using a three-level gain-scheduling adaptive control scheme. In using proportional and integral feedback, three levels of proportional gain were available and were applied, depending on the difference between motor speed and set point. Improvement in reaching an initial set-point was obtained with this approach.
Load variations were applied to the motor by attaching a second identical motor, connected by coupling the shafts. By adding resistances of different value to the second motor’s terminals, load variations could be applied. Under these conditions, it was demonstrated that with gain-scheduling adaptive control, the return to set-point could be achieved more ideally, with less overshoot, than with just proportional and integral control optimized for the motor driving an unloaded second motor.