欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
提出一种基于模型的配电网故障诊断方案．该方案首先根据配电网原理模型的仿真数据和实际观测值存在的差异得到极小冲突集，然后由离散二进制粒子群优化算法推出可能的故障元件和故障形式，最后由贝叶斯方法确定概率最高的诊断结论。通过实际建模、编程和实验证明了该方案的可靠性和有效性。仿真结果表明，与HS—Tree、BooleanAlgebra方法、遗传算法等算法相比，离散二进制粒子群算法搜索效率更高，可节约1／3。1／2的搜索时间，并且可以避免当问题规模较大时出现内存溢出问题。
英文摘要：
A scheme of model-based fault diagnosis is proposed for distribution network,which obtains the minimal conflict sets according to the difference between the simulative data of distribution network principle model and the actual observed data,applies the discrete BPSO（Binary Particle Swarm Optimization） algorithm to reason out the possible fault elements and fault forms,and adopts the Bayesian method to finally determine the diagnosis conclusion with the highest probability. Practical modelling,programming and experiment verify its reliability and effectiveness. Simulative results show that,compared with HS-Tree algorithm,Boolean Algebra algorithm and genetic algorithm,the discrete binary particle swarm optimization algorithm has higher search efficiency,saving 1/3--1/2 of search time and avoiding memory overflow.
同期刊论文项目
同项目期刊论文
A New Classification Method for Transient Power Quality Combining Spectral Kurtosis with Neural Netw
期刊信息
位置：