Sensorless Vector Control Of Three-Phase Induction Motor to Crush Sugar Cane

This paper presents an adaptive neuro-fuzzy inference system sensorless speed estimation for IM drives is designed and simulated, together with fuzzy-PI optimized by GA to crush sugarcane. In previous years, the identification and monitoring of this highly nonlinear dynamic plant was the focus of research efforts in the field of IM control. Speed sensors are needed by existing vector control methods for field orientation and control, but their installation makes the drive system bulky, unreliable, and expensive, and installing them might not be possible in some applications, like high-speed drives or motor drives in hostile environments. In these scenarios, speed is determined using stator quantities that are simple to measure. To address the speed sensing issue, numerous speed sensorless approaches have been developed. Developed speed estimation algorithms require a lot of computing time and/or depend on various parameters. The proposed estimation method in this thesis uses ANFIS to get the speed signal. A fuzzy PI controller is used in place of the traditional PI controller. A hybrid learning technique is utilized to train the ANFIS, which is used as an estimator. When the motor drive is operating in a closed loop at different values of speeds and loads for speed observation, the data used for training is obtained using conventional FOC simulations. The use of MATLAB®2020a is used to model the entire drive system. Last but not least, the drive results have been examined for both steady-state and dynamic settings, including sensitivity to motor parameter uncertainty, tracking the drive's speed set points, quick torque response, low-speed behavior, and step response with speed reversal. For the transient response, speed tracking, and low-speed operation, the simulation result error between actual and estimated speed has been less than 0.1%, 0.5%, and 0.2%, respectively. According to simulation data, the overshot and settling time are reduced by 0.0504% and 0.14 seconds at full, respectively, when employing fuzzy-PI compared to PI controllers for the reference speed of 60 rad/sec. The fuzzy-PI controller also produces a robust response.