Adaptive neuro-fuzzy logic based sensorless brushless DC motor control

In this thesis, an adaptive neuro-fuzzy logic based sensorless control system for brushless DC motor is studied. Starting by proposing a fuzzy logic based speed controller then followed by using neural networks to form an adaptive neuro-fuzzy inference system (ANFIS). In the next step, ANFIS controller is applied for controlling the speed of sensorless brushless DC motor drive. By using sensorless control method, less parts and maintenance of the motor drive are needed beside decreasing the overall cost of the system. Furthermore, a conventional control method using "Proportional-Integral (PI)" controller is also studied for comparing the performance of the different proposed control methods. For speed control of brushless DC motor, fuzzy logic and ANFIS based controllers have a robust and best performance since they are independent of the mathematical model of the motor which is complex, multivariable, non-linear, and subjected to many variations. For the proposed motor control system, PI, fuzzy logic based and ANFIS based speed controllers have been designed and simulated with a complete drive model. Moreover, motor drive system is made to run under different conditions regarding the load and speed and then the performance of every controller has been analyzed. In the last part of the thesis, all proposed controllers are implemented on an embedded system for physical application and again the performance of every controller is studied.