Research On The DC-biased Magnetization Model And Algorithm Of Magnetic Controlled Reactor

Along with industrial scale continuous developing in our country, the electricity demand are growing, green power supply has become one of the contemporary trend. As an important kind of reactive power compensation device, magnetic controlled reactor has been widely used in electric power industry, for which not only reduce line loss, also improve power quality. During normal work, It can produce alternating magnetic field and dc magnetic field, under this two magnetic field, DC bias magnetic hysteresis phenomenon will produce, working with the distortted current. While, It’s performance is up to the hysteresis characteristics and the core loss characteristics by a large extent. As the project design require more and more high accuracy, a good DC-biased magnetization model is need to set up. From the perspective of the magnetic hysteresis characteristics, the text set up a DC-biased magnetization model of MCR. And Combine with the electromagnetic coupling principle, establish a mathematical model to study the work current characteristics under DC-biased.First of all, The text expounds the basic working mechanism of MCR to lay a foundation for the following study on DC-biased characteristics; And Use the ANSOFT software to establish a 2-D finite element simulation model of MCR, put the DC-biased magnetization curve of silicon steel into the simulation of magnetic field, which provided by manufacturer. Get the magnetic induction intensity distribution of MCR under DC-biased; also, make an analysis of difference on DC-biased between MCR and transformer.Then, make an inctroduce for classical hysteresis model of ferromagnetic materials which are commonly, including Precisach model, J-A hysteresis model, also a improved model which based on Precisach hysteresis model and so on. Specificly explore the impact on the hysteresis loop of the J-A model’s parameters, at last, make a summarize.Whether Precisach model, J-A model or the improved model based on Precisach model, whose magnetic hysteresis loops are symmetrical, when comes to the small hysteresis phenomenon, They can not described. Obviously, these classcical hysteresis models are not applicable to MCR. To recive an accurate describe on the DC bias characteristics of MCR, a new different approach must be provided. The text based on 400 v /100 kvar MCR, which provided by Guangzhou Junfa electric co. LTD, using relevant record instrument to get experimental samples of voltage and current waveforms and it’s values under different triggering Angle. Then, organize them into samples.And basd on these samples, this text proposes a DC bias magnetic hysteresis model which based on BP neural network, at last, get the hysteresis loops under different bias. The DC-biased magnetization model can well describe the DC bias of MCR. What’s more, this model not only avoid the mathematical expression complexity of the traditional hysteresis model, also improve the accuracy .which is more suitable for engineering application.Finally, based on the B-H curve from the DC-biased magnetization hysteresis model of BP neural network, combined with the electromagnetic coupling principle, establish the mathematical model of MCR. Through C # programming, get the winding current and other branch current, compare them with the experimental waveforms, turn’s out basic consistent. It shows that the mathematical model of MCR is right. For which provide an effective method for the DC-biased research of MCR.