初次建立基于神经网络方法的大体积混凝土裂缝控制 ES的框架结构 ,对框架的设计思路、总体结构、知识库的构建、神经网络应用模式的具体实现等展开了描述 ,重点对神经网络聚类方法在神经网络知识库和自学习功能中的应用进行了研究 ,并且通过计算机模拟进行了可行性验证 ,从中得出了一些有益的结论和新的理论设想 The frame structure of mass concrete crack control ES based on neural network method was firstly established and the design idea, overall structure, construction of knowledge base and concrete realization of neural network application mode were described. The emphasis was put on neural network clustering method The application of neural network knowledge base and self-learning function has been studied, and the feasibility verification has been carried out by computer simulation, from which some useful conclusions and new theoretical ideas