Study On The Solving Quickly Of Harmonic Elimination PWM Model

With the rapid development of Power Electronic Technology, Inverter, which is the main circuit structure of the power electronic equipments, has been used on large scale. These devices have the merits of high efficiency and energy saving. The damage of harmonic brought by inverter get more and more seriously to power network, electric equipments and communication network et al. People pay more and more attention to the research of harmonic elimination technology. Selective harmonic elimination technique, which is the core control technique of modern inverter, is studied in this dissertation in detail. SHET can eliminate selected harmonics and control the fundamental component according to the requirements, and has some other advantages, such as lower switching frequency, no given lower order harmonic, high voltage gain, and high quality of outputs and so on.The mathematical model of selected harmonic elimination technique is a nonlinear equation. For a long time, it is a main factor to hamper the real-time application of SHET. The research is carried on aiming to quickening convergence speed. The automatic-changing Homotopy algorithm is proposed in the dissertation. The algorithm, which can change the solving method based on fundamental component, has the advantages of large convergence domain, high convergence rate and high precision.Many trigonometric functions are contained in the mathematical model, during solving equation, the calculated speed and the precision are affected by the calculations of trigonometric functions. In order to overcome the shortcoming brought from the calculation of the trigonometric function, the approach of numerical equivalent, which remove the trigonometric function from the nonlinear equations of SHEPWM, is presented in this dissertation. So it is available to avoid the calculation of trigonometric function during solving equations, therefore, reduce the time ofsolving and improve precision. In addition, several predicting initial value according to different control models, which affects the computed rate and convergence, are given, these formulas accelerate the convergence speed on large scale. And several examples are given to further verify the validity of the proposed methods.Direct-omitting minimum pulse is presented without optimizing, the goals that hardware circuit is convenient to be made and switching loss is decreased and switching frequency are reduced are attained.Finally, several simulations and experimentations conducted to demonstrate visually the validity and feasibility of the foregoing proposed method.