【摘要】 自1980年以来,“专家系统”作为人工智能中一个活跃的分枝已被引入土木工程领域,开始对房屋结构的破损状态进行诊断,但至今尚未进入实用阶段。其主要困难是专家知识的获取与表达。在以往的研究工作中,可能过高地估计了从事人工智能研究的“知识工程师”的作用,而相对地低估了房屋诊断决策方法上的难度。本文根据从局部到整体和从个别到一般的认识原则,提出一个对钢筋混凝土结构进行破损评估的语言决策模型。根据这个模型建立了一个对钢筋混凝土梁进行诊断的原型系统、该系统用LISP语言写成,并成功地在IBM-PC微机上运行。通过多次考核证明,该系统可以得出和专家评估基本一致的结果。更多还原

【Abstract】 Since 1980, the expert system methodology has been used in devoloping some linguistic decision models for assessing the damage of building structures. But it is still troublesome. Perhaps, the sufficient experience in the civil engineering and the strict mathematical training are lacking for our knowledge engineers. In this paper, a linguistic decision model, which follows the real way of the experienced civil engineer, has been built for assessing the damage of a reinforced concrete (R.C.)beam.The methodology involves two steps.In the first step, information is collected from a human expert regarding allthe major factors which may contridute to the damage of a structure and this information is arranged in the form of a graph.The graph involves both primary factors whose linguistic values can be ascertanied from physical measurements (such as the breadth of a crack, etc.) and nonprimary factors (suchas the corrosion state, etc.) which can not be directly controled from meas urement. Next,the information is collected about th edependence of each factor on other primary or nonprimary factors. This intormation is expressed in the form of IF-THEN rules and decision tables. Finally, an algorithm is devel oped which will yeild the linguistic value of the state of damage, of the R. C. structure from the observed values of the various primary factors. The decision model is written in the LISP language. The use of the model is illstratedby several practical examples.更多还原