基于知识库的锅炉“四管”爆漏预报专家系统

Expert System of Boiler's Four-Tube Crack Prediction Based on Knowledge Base

从分析锅炉"四管"爆漏的特点出发,提出了基于知识库的锅炉"四管"爆漏预报专家系统。文中给出了系统总体结构模型和面向对象模型,并根据系统要求建立了数据库;积累案例形成案例库;获取专家经验及领域知识建立了规则库;依据知识层次结构建立了模型库;4库一体化形成专家系统。文中介绍了基于安全、基于规则和基于模型的预报推理方法,比较了各种预报推理方法的优、缺点,并提出了基于案例、规则和模型的混合推理,其优点在于能充分发挥各种推理方法的优势,克服各自存在的不足,能最大限度地找到问题的解。文中还讨论了贴近度的定义、表示和计算以及规则表示等问题。最后,给出了系统的功能和实现方法。应用证明:基于知识库的锅炉"四管"爆漏预报专家系统可以得到与专家同样的预报结果。

Expert system of boiler's four-tube crack prediction based on knowledge base in posed according to analysis of characteristic of four-tube crack. General structure of system and object-oriented model are discussed in the paper. Database (DB) is set up according to system demand, case base (CB) founded through accumulating cases; rule base (KB) established after accepting expert experience and knowledge of this field, and model base (MB) built according to the structure of knowledge. The four bases are recombined in the expert system. The predictive methods based on case, on rule and on model wer described. Compared with each other, advantage and disadvantage of these reasoning methods are obtained. The mixed reasoning based on case, rule and model is posed to overcome their shortcoming, retain fully their advantage and find the answer to the problem as possible. The definition, representation and calculation of the degree of similarity and rule representation are discussed. The function of system and way of realization are described in the paper. It is proved by application that expert system of four-tube crack prediction based on knowledge base can achieve the same predictive result as the expert. Figs 3 and refs 8.