后向离散状态事件驱动电力电子仿真方法

摘要

为解决电力电子系统中的刚性状态方程数值解算困难问题,在后向量化状态系统(BQSS)算法的基础上提出一种后向离散状态事件驱动(BDSED)仿真方法.BDSED方法是隐式的,其关键在于如何选择一组合适的量化函数值组合,使得据此计算出的导数向量能够让各状态变量向其对应的量化函数值趋近.由于在每一步计算中,每一个状态变量下一时刻的量化函数都有两种取值,所以对于复杂高维系统,枚举的方式并不可行;同时各状态变量的量化函数值的确定存在互相耦合、相互制约的关系,导致问题更加困难.为解决该问题,提出一种基于有限状态机的实现方案,并在带非理想器件模型的变换器电路上进行了算例验证.仿真结果表明,基于有限状态机实现的BDSED能高效地选取量化函数值组合,在解算变换器等刚性系统时仿真效率明显优于DSED方法和传统的时间离散刚性解法.%A backward discrete state event driven (BDSED) simulation method based on the backward quantized state system (BQSS) algorithm is proposed in this paper, to solve the difficult problem of rigid state equations in power electronics systems. BDSED method is implicit, and the key point is to select the correct combination of quantization values, which allows the state variables to approach their quantization values. Each state variable has two possible quantization values at each step, thus enumeration is infeasible in complex systems. What's more, taken into account the constraint and mutual coupling between state variables when determining their quantization values, the question becomes quite difficult. Thus, an implementation approach of BDSED based on finite state machine is proposed in this paper. The simulation algorithm is tested in a converter circuit with non-ideal device models. Simulation results show that BDSED based on finite state machine can efficiently select the combination of quantization values, and its simulation efficiency is obviously higher than DSED as well as traditional time-slicing ODE methods when solving stiff systems such as power converters.