摘要
在网络的最优化设计和容差分析中,灵敏度分析具有重要意义.常规的灵敏度计算方法都是纯数值计算,它们的特点是逐点进行完整的数值求解,因此不可避免的存在冗余项对消问题、数据存储量大和计算误差等问题.以图论知识为基础,应用连通图中的k 树树支导纳乘积,给出网络函数及其灵敏度分析的符号表达式.该方法较之以数值计算为基础的增量网络法和伴随网络法,具有计算精度高、计算速度快、易于计算机实现等优点,算例表明该算法理论的正确性和可行性.
The sensitivity analysis is important in electric network optimum design and tolerance analysis. The general method of sensitivity analysis is pure numerical calculation, their characteristics are completly numerical analysis one by one. Those methods inevitably exist computational complexity, such as redundant polynomials eliminating, needing more data memory capacity and calculation error. Based on network graph theory, using admittance product of k-tree branches in connected graph, the symbolic expression of network function and its sensitivity are presented. Comparing with incremental network method, the method is more fast and effective and is prone to program. It is proved by the example that the method has validity and feasibility.
出处
《长沙电力学院学报(自然科学版)》
2004年第3期38-41,53,共5页
JOurnal of Changsha University of electric Power:Natural Science
基金
国家自然科学基金资助项目(50207001)
关键词
灵敏度分析
网络函数
k-树
符号计算
sensitivity analysis
network function
k-tree
symbolic calculation