实际大系统工程中,可靠性评估软件计算耗时长,难以满足快速且精确的评估要求。为此,该文将TRIZ理论引入电力系统可靠性评估算法的改进中。首先根据TRIZ理论,应用矛盾矩阵找出序贯仿真算法计算时间与精度之间矛盾的解决方案,再根据矛盾解决原则筛选出塔存储技术及拉丁超立方采样(Latin hypercube sampling,LHS)技术改进基于蒙特卡洛模拟的传统算法。改进后的算法在缩短故障状态搜索时间的同时能更全面覆盖系统元件状态持续时间的概率分布。在IEEE-RTS 79测试系统上进行的算例分析结果表明,改进后的算法收敛更快,能在较短的计算时间内获得高精度的解,且具有良好的稳定性。
In the reliability evaluation of practical large engineering system by the system reliability evaluation software, the time for the computation is too long, so it is hard to meet the demand on fast and accurate calculation. For this reason, The TRIZ theory is led into the improvement of power system reliability evaluation algorithm. Firstly, according to the TRIZ theory, the contradiction matrix is applied to find out the method to resolve the conflicts between the calculation speed and the obtained accuracy by sequential simulation algorithm; secondly, according to the principle of resolving conflicts the tower storage technology and Latin hypercube sampling (LHS) are screened out to improve the Monte Carlo simulation based traditional algorithm. The improved algorithm can shorten the time to search the fault state and completely cover the probabilistic distribution of state duration of system components: Simulation results of IEEE RTS-79 show that the improved algorithm is stable and it converges faster, besides, using the improved algorithm the solution with high accuracy can be obtained by shorter calculation time.