Performance analysis of pi controller under harsh external disturbances on brushed direct current motor

Automatic control systems are one of the fundamental components of modern engineering applications and are widely used in many areas, ranging from unmanned land, air, and sea vehicles to robotic systems, industrial production lines, and CNC machines. The reliable, stable, and efficient operation of these systems is directly dependent on the success of the control mechanisms they incorporate. Electric motors are the primary source of motion in these systems; the precise control of parameters such as speed and position is a key factor in system performance. This study addresses the speed control problem of a brushed direct current (DC) motor and presents experimental applications on the system. Due to their simple structure, reasonable cost, and ease of control, DC motors are widely preferred in low and medium power applications. Within the scope of the thesis, a Proportional–Integral (PI) type controller was designed and implemented for motor speed control. The fundamental objective of the PI control structure is to ensure that the motor reaches the specified reference speed in the shortest possible time with minimum error and to obtain a stable response in the system. However, this performance depends directly on the correct determination of the controller gains. To this end, a heuristic algorithm called Particle Swarm Optimization (PSO) was utilized. This algorithm, which searches for the most suitable result in the solution space based on swarm behavior in nature and can evaluate multiple solution candidates in parallel, was used to effectively adjust the gains (K_p and K_i). The study consists of two main experimental stages. In the first stage, the goal was to derive a mathematical model representing the dynamic structure of the DA motor used. The motor speed responses corresponding to the step inputs were analyzed, and a model reflecting the characteristic parameters of the system was successfully obtained using the PSO algorithm. The validity of the model was tested by comparing it with system responses under a steady-state reference signal, and it was observed that the model realistically reflected system behavior in terms of criteria such as rise time, overshoot amount, settling time, and steady-state error. In the second stage, the developed PI control structure was implemented in real time on the Altera Cyclone IV FPGA platform. Feedback signals provided by the magnetic encoder were digitally processed, and the system's speed control was performed based on this data. To evaluate the effects of different PI gain combinations on the system response, the reference speed (ω_Ref) and actual speed (ω_Actual) data obtained in each case were compared using Python-based analyses. Comparative evaluations were performed using performance metrics such as rise time, overshoot, settling time, and tracking performance for various gain combinations, and objective performance tables were created. In addition, the effects of control gains on armature current peak values were examined in detail, and the relationship between system performance criteria and gains was visualized using heat maps and three-dimensional surface graphs. Furthermore, multi-objective PSO analyses considering current limits under different load conditions were performed, and detailed evaluations were made for the operating range between 0% and 50%. The search process of the PSO algorithm was also explained graphically, and the optimal gains found in the optimization process were supported by experimental results. Furthermore, the best gain set obtained with PSO has been evaluated by comparing it with manually tuned and literature-recommended controllers. The results obtained reveal that the PI control structure optimized with the PSO algorithm provides high accuracy and stability in the speed control of the brushed DC motor. The system successfully met the desired performance level in terms of both modeling and control, and the effects of the gains on the system dynamics were experimentally verified. In this context, the developed method was evaluated as an applicable solution for similar motor control applications and produced results that contribute to the literature in the field of control engineering.