Real time trajectory control of 3 phase aa servo motor based on intelligent methods

In this thesis, a Space Vector Pulse Width Modulation (SVPWM) based Field-Oriented Control (FOC) method was applied for speed and position control of a three-phase AC servo motor. The FOC principle is based on separating the flux and torque components of the three-phase stator current. This allows the three-phase current to be controlled separately using two current components: one generating flux and the other generating torque, thus enabling the control of the Permanent Magnet Synchronous Motor (PMSM) as if it were a separately excited DC machine. The SVPWM switching technique allows for more precise and efficient control of the motor within the FOC method. By optimizing the three-phase voltage or current vectors, it brings the output signal closer to a sinusoidal waveform. This increases the motor's torque production, improves efficiency, and reduces harmonic distortion. To develop a vector-controlled synchronous motor drive system powered by a three-phase inverter, a novel switched inductor-based transformerless boost DC-DC converter was designed, and a new Artificial Neural Network (ANN) model was used for the switching technique. The boosted DC voltage was then converted to AC through a three-phase H-bridge. Conventional boost converters cannot meet the required voltage boost performance for real-time applications. To achieve a higher voltage gain, converters are often operated at high duty cycles, which can lead to reverse recovery issues in diodes. Moreover, high duty cycles can impose significant voltage stresses on semiconductor switches and diodes. Therefore, there is a need for alternative converters that can meet load requirements and provide high performance. The proposed topology was compared with other designs and demonstrated superior performance with characteristics such as high gain, high efficiency, low voltage ripples, reduced switching losses, and ease of control. The proposed DC-DC converter enhanced motor performance and control precision by reducing voltage fluctuations and power variations, allowing the servo motor to move more accurately and steadily along its trajectory. The low switching losses contributed to more efficient energy conversion, improving system performance and resulting in cleaner and more stable speed and position control signals, allowing the motor to respond better to changes in trajectory. In the switching technique of the proposed boost converter, an Artificial Neural Network (ANN) was used. A dataset was created using the input and output data of the PI controller to train the ANN. The ANN was successfully trained, and a model was created with a single hidden layer and 25 neurons. This structure is quite simple in terms of both hardware and software, offering a system design that can be easily implemented. The DC-DC design and ANN switching technique directly affected the speed and position control of the PMSM. Compared to the PI controller, the settling time for speed-torque convergence with ANN was 0.36 seconds, while it was 0.55 seconds with PI. The simulation results showed that the settling time to the reference speed and position values was shorter and the system response speed was better with ANN compared to PI. The results demonstrate that using the ANN switching technique in the DC-DC boost converter is a favorable approach. In real-time modeling, the Texas Instruments TMDSDOCK28335 DSP based experiment kit was used. The system's performance was analyzed, and the model's performance was presented by successfully tracking the reference inputs with the model's output. Considering all the results, it was shown that in the case of speed and position changes, the system's response speed, stator current, and torque tend to quickly stabilize according to the control model's operating characteristics, yielding results compatible with PMSM performance.