Минимизация и перераспределение коммутационных потерь в инверторе напряжения при использовании алгоритма широтно-импульсной модуляции с прогнозированием

Abstract

The article deals with testing a combined PWM method for a voltage source inverter, which is aimed at minimizing the switching losses and distributing them among the power converter modules. To reduce switching losses in the inverter, duty cycles are calculated for several possible PWM shaping methods, and a PWM method in which the phase with the largest value of current flowing through it is not switched is chosen. Another advantage of this method is its ability to control the thermal operating conditions of the inverter’s IGBT modules. This is especially important for designs in which the power IGBT modules are installed on a common heat sink, because with such configuration, the modules of a three-phase converter operate under different thermal conditions. To avoid overheating of a particular IGBT module, an algorithm is proposed using which losses can be redistributed among the modules. Since it is difficult to establish an exact mathematical relationship between the desired behavior and the controlled parameters, a prediction method is used. The total losses anticipated for each possible PWM method are calculated, and the objective function that takes into account several variable parameters is determined. By varying the objective function’s weighting coefficients, different system operation modes can be specified. Three system operation modes defined by three objective function configurations are analyzed. In its first configuration, the objective function makes it possible to select the PWM method in which the total losses in the system are kept to a minimum. In its second configuration, the objective function allows the user to calculate the minimum losses in the inverter’s most heavily heated leg. In its third version, the objective function is a combination of the first two configurations and is aimed at simultaneously reducing the total losses in the system and losses in its most heavily heated module. The performance characteristics of each of these three objective functions and of the standard vector PWM are tested on a heat sink thermal model. The presented simulation results show that application of the first objective function helps to decrease the total losses in the system as compared with those in the case of using the standard vector PWM; application of the second objective function makes it possible to decrease losses in one phase, whereas the temperature in the other legs tends to grow; and the third objective function was found to be the most effective one, because its application made it possible to achieve the minimal losses in the system and a lower temperature of the most heavily heated module.