中文摘要：

针对我国大部分抽水蓄能电站无法开展水泵水轮机真机试验的现状,为了能够开展抽水蓄能电站的厂内优化调度和水火电站联合经济运行,提升抽水蓄能电站的动态效益,从研究抽水蓄能电站的水泵水轮机模型试验资料、引水管的设计资料以及制造厂提供的发电机效率特性曲线等原始资料出发,首先,获取了模糊动力特性数据;其次,结合抽水蓄能电站计算机监控系统的历史数据库,采用挖掘数据对模糊动力特性数据进行基于＂指数衰减＂方法的点对点静态修正;再次,采用基于可拓神经网络的训练方法,基于＂同趋不相交和机组段水头线性等间距＂的原则,对动力特性数据进行再次动态修正,并把两种方法集成编译为计算机软件模块,嵌入式开发了＂抽水蓄能电站水泵水轮机动力特性分析系统V3.0＂。实践应用表明,修正后的抽水蓄能电站动力特性方程符合工程应用实际,取得了很好的实际效果,这为下一步开展抽水蓄能电站的厂内优化调度和水火电站联合经济运行调用奠定了基础。

英文摘要：

Considering the shortage of real machine test data in design of pump-turbine for most of the pumped storage power stations in China, this paper presents a study of plant optimal operation and joint economic operation of hydropower and thermal power stations based on model test data of pump turbine, design data of pipe system, and efficiency characteristic curve of generator provided by the manufacturer to enhance the dynamic benefits of pumped storage stations. First, data of fuzzy dynamic characteristics were measured and a point-to-point modification on fixed dynamic characteristics of fuzzy data was made with an ＂exponential decay＂ static method in combination with the historical databases recorded by the computer monitoring system of pump-storage station. Then, the data of dynamic characteristics were corrected dynamically via an extension neural network training method based on a principle called ＂disjoint and linear equidistant head of unit＂. These two methods were coded in integrated modules that are embedded in a software system. Thus we have developed the system V3.0 for analysis of the dynamic characteristics of pump-turbine in pumped storage power station. Application shows that the characteristic equations of pump-storage station derived using the previously-mentioned modification and correction better describe the real cases and their solution produces satisfactory results. The present work would lay a basis for further study on plant optimization of pump-storage station and economic operation of joint hydro-thermal power stations.