Temperatures estimation system of electrical machines on wireless sensor networks
Yi Huang · DepositOnce · 2021
This dissertation proposes a model-based software method to develop a temperatures estimation system for an asynchronous machine, which is implemented in wireless sensor networks (WSN). The system can estimate the temperatures of the stator winding, the rotor cage and the stator core. Firstly, a physical model of an asynchronous machine is built and validated in Dymola. The electrical, mechanical and the thermal behaviors performed well in the Dymola simulation model. Based on the physical model, an efficient and reliable thermal model for tracking the temperatures of the stator winding, the rotor cage and the stator core is built using Dymola. All the thermal parameters of the asynchronous machine are identified. One of the most difficult tasks is to identify the reference stator core losses and reference friction losses, which can be determined by a no-load test and load test on the test bench. The conductance values are calculated by the losses and temperatures at the steady state of the machine. The best-fit capacitances are found by using Genopt, an optimization program. Two different algorithms are used for the temperatures estimation. A 4th-order Kalman filter (KF) algorithm and a 9th-order extended Kalman filter (EKF) are first implemented based on the state-space equations in MATLAB/SIMULINK. The Model-in-the-Loop (MiL) method is used to verify the algorithms. The physical model in Dymola and the algorithms are connected together in the simulation using SIMULINK. After the verification of the algorithm, both are implemented in a wireless sensor network (WSN), which is based on the IEEE1451 standard using Contiki OS. To estimate the respective temperatures of the stator winding, the rotor cage and the stator core of an asynchronous machine, KF and EKF algorithms are implemented into the resource restricted embedded system. Finally, under different experiment conditions, the temperatures estimation system in WSN are tested on the test bench. The real-time WSN temperature estimation system is independent from the control algorithm and functional under any load condition, as long as the current of the stator is a nonzero system and measured with very high accuracy.