人工智能系统发动机故障诊断

作为人工智能研究中面向实际应用的课题，本系统综合安２４飞机发动机故障诊断的专家经验和各种典型实例，用基于规则的演绎推理系统，实现了对飞机发动机故障的经验性诊断。并利用人机对话方式，提供推理报告，进行自学习（机器学习）。系统模型由知识库、推理机制、自学习机制及解释机制四个模块构成，它们相对独立，能进行平行处理。有利于调试、修改及扩充新知识。同时本系统还有效地解决了栈空间的问题。其调试结果表明：与ＩＮＳＩＧＨＴ２相比，响应时间短，速度快，功能强。并有较好的用户界面。具有一定的实用性。

Artificial Intelligence System for Airplane Engine Trouble Diagnosis

In order to put to practical use the study of artificial intelligence, this paper presents an artificial intelligence system for trouble diagnosis of aircraft engines by incorporating some experts' experience in AN-24 engine trouble diagnosis and some typical examples. This system carries out the experimental diagnosis of airplane engine trouble with a deduction-inference system based on the regulations, produces an inference report in a manner of man-to-computer talk and, in addition, can be used for self-study. The system model consists of a knowledge room and inference, self-study and explanation mechanisms. These four sections are relatively independent, which makes parallel operations possible and facilitaes debugging, revision and the addition of new knowledge.Furthermore, the system solves the memory problem effectively. The debugging result shows that in comparison with INSIGHT 2, this system has shorter responding time, higher speed and better function, thus catering to the needs of more customers.