在可控串联电容补偿控制中，针对电力系统模型及运行参数的不确定性，应用神经网络和模糊控制理论设计了非线性控制器，从线性最优控制器设计值中提取出经验规则，将模糊化后的变量引入神经网络结构，采用变尺度算法对网络参数优化，解模糊后得到反馈增益矩阵，仿真结果表明控制器对静态稳定和暂态稳定均有良好效果。

本文结合相似日方法和人工神经网络方法的优点，提出一种基于相似日和径向基函数网络的（SD&RBF）新方法。该方法具有较好的稳定性和精度，减少了普通相似日方法对人的经验的依赖，仍然保持了应用简便，实用有效的特点，开发了96点短期负荷预测软件，在陕西电网已经成功运行一年多，给出了几种方法的实际应用结果并进行比较 统计表明新方法的平均均方根误差较普通相似日方法从3.153%降低到2.954%综合方法进一步降低到2.854%。

Dynamic equivalents of distribution systems, including the dynamics of electric loads and on-line tape changers, are needed in voltage stability studies. This paper addresses a method to dynamic equivalent of a radical distribution system for voltage stability studies. A new equivalent model is proposed. The procedure and detailed formulae for separately aggregation of dynamic loads and network with OLTCs are derived. Case studies show the accuracy of the equivalent model and the availability of the method.

Final validation of a dynamic equivalent (DE) is to obtain the equivalent parameters from field tests and apply them to practical stability studies. A microcomputer-based equipment is developed for online measurement and identification, with which several valuable field results have been captured. The results are then applied to dynamic response simulation and stability limit calculation.