基于混合推理的枪械故障诊断

目的 提高枪械研制过程中的故障诊断效率与正确率，针对现场排除难度大、单一诊断方法准确率低且操作过程复杂等问题，建立RBR-CBR混合推理模型，实现枪械故障的快速准确诊断。方法 针对枪械故障因果关系复杂、随机性大和迭代更新快等特点，提出以RBR为主、CBR为辅的推理模式。采用RETE算法进行RBR推理，提高规则匹配的效率；采用可拓方法，建立枪械故障实例的物元模型；对数值型和区间型不同组合形式的局部相似度进行准确表征；基于最近邻近法进行实例相似度分析。结果 基于RBR和CBR的混合推理方法，开发了枪械故障诊断专家系统。此专家系统可有效指导研制过程中枪械故障问题的诊断和排除。结论 利用混合推理方法，在保证枪械故障诊断速度的前提下，提高了故障诊断准确率，证明了基于RBR和CBR的混合推理方法的有效性和此专家系统的高效性。

Firearm Fault Diagnosis Based on RBR and CBR Hybrid Inference Methods

In order to improve the efficiency and accuracy of firearm fault diagnosis, the RBR-CBR hybrid inference model is established for such problems as difficult manual site elimination, low accuracy of single diagnostic method and complex operation process, achieving rapid and accurate diagnosis of gun faults. Aiming at the characteristics of complex causal relationship, large randomness and fast iterative update of firearm faults, a reasoning mode with RBR as the core and CBR as the supplement is proposed. RBR inference is carried out by RETE algorithm to improve the efficiency of rule matching; the matter-element model of gun failure examples is established by extended method; the local similarity of different combinations of numerical and interval types is accurately characterized; and the similarity of examples is analyzed based on the nearest neighboring method. Based on the hybrid inference methods of RBR and CBR, an expert system for firearm fault diagnosis is developed. This expert system effectively guides maintenance support personnel in firearm troubleshooting. The hybrid inference method is used to improve the accuracy of fault diagnosis under the premise of ensuring the speed of firearm fault diagnosis, and the effectiveness of the RBR and CBR-based hybrid reasoning method and the high efficiency of this expert system.are verified.