Parameters estimation of an inverter-fed induction machine using optimization algorithms and an experimental testbench

The use of electrical machines in the industry and in the mobility sectors is widespread. Consequently, an accurate estimation of the parameters and losses of the machines is essential. The induction machine is the most reliable and commonly used machine in the industrial field. Therefore, this research focuses on analysing it to estimate its parameters and quantify the losses, divided by their individual contributions. Traditionally, specific tests, such as no-load and blocked rotor tests, are required for this purpose. However, the method presented in this thesis exploits a different configuration based on a back-to-back model using two identical machines, which are fed reciprocally by two inverters. This approach eliminates the need of the traditional test. The two machines are mechanically coupled and can operate as either a motor or a generator, depending on the frequency imposed by the inverters. The parameters calculation relies on the voltages and currents measured from the machines and the resolution of the lumped-parameter circuit. The use of the inverters enables operation under different conditions (PWM and square wave were analysed) to explore various operating points with different frequencies, voltages and harmonic contents. The parameters estimation is performed using two optimization algorithms: Particle Swarm Optimization (PSO) and Interior Point (IP). They calculate the machine parameters starting from the sensors measurement of voltages and currents of the machines. To validate the algorithms and the minimizing functions, simulations were conducted using MATLAB/Simulink. An experimental testbench was developed to obtain experimental results for the parameters and the losses. The losses analysis focused on identify individual contributions, including Joule losses, iron losses and mechanical losses. The results obtained from both simulations and experimental tests demonstrate the effectiveness of the proposed method in accurately estimating the parameters and losses of induction machines, offering a practical and efficient alternative to traditional testing methods.

The use of electrical machines in the industry and in the mobility sectors is widespread. Consequently, an accurate estimation of the parameters and losses of the machines is essential. The induction machine is the most reliable and commonly used machine in the industrial field. Therefore, this research focuses on analysing it to estimate its parameters and quantify the losses, divided by their individual contributions. Traditionally, specific tests, such as no-load and blocked rotor tests, are required for this purpose. However, the method presented in this thesis exploits a different configuration based on a back-to-back model using two identical machines, which are fed reciprocally by two inverters. This approach eliminates the need of the traditional test. The two machines are mechanically coupled and can operate as either a motor or a generator, depending on the frequency imposed by the inverters. The use of the inverters enables operation under different conditions (PWM and square wave were analysed) to explore various operating points with different frequencies, voltages and harmonic contents. The parameters estimation is performed using two optimization algorithms: Particle Swarm Optimization (PSO) and Interior Point (IP). They calculate the machine parameters starting from the sensors measurement of voltages and currents of the machines and the resolution of the lumped-parameter circuit.. To validate the algorithms and the minimizing functions, simulations were conducted using MATLAB/Simulink. An experimental testbench was developed to obtain experimental results for the parameters and the losses. The losses analysis focused on identify individual contributions, including Joule losses, iron losses and mechanical losses. The results obtained from both simulations and experimental tests demonstrate the effectiveness of the proposed method in accurately estimating the parameters and losses of induction machines, offering a practical and efficient alternative to traditional testing methods.