【作者】 高峰；

【导师】 乔建忠；

【作者基本信息】 东北大学 ， 计算机应用技术， 2014， 硕士

【摘要】 近几十年我国经济发展迅猛,楼房的平均高度也在不断增加,电梯越来越成为人们生活和工作中必不可少的工具。电梯的可靠性和安全性时刻在威胁着人们的人身和财产安全,因此必须做到及时发现和处理电梯故障。要保证电梯的可靠性需要做到以下两点：在电梯正式出厂安装使用之前对电梯进行可靠性测试；及时发现和处理电梯运行过程中出现的故障。根据当前对电梯故障诊断的需要,本文研究了专家系统和支持向量机用于故障诊断的相关理论,在专家系统的推理机中融入支持向量机,利用支持向量机的数据挖掘能力,从故障数据中分析出电梯故障原因。通过专家系统建立知识库并进行管理维护,使用规则对电梯故障进行推理诊断并用解释器做出相应的解释说明。系统发挥了两者的优势,提高诊断能力。本文的主要研究内容如下：首先本文介绍了故障诊断技术,对故障诊断方法进行了分析与比较,介绍了几种典型的故障诊断专家系统,并分析了电梯的常见故障。其次建立电梯故障诊断专家系统模型,选取产生式规则表示法、正向推理分别对知识进行表示和推理。再次研究了电梯故障诊断算法,包括支持向量机的基本原理及其关键点；选用径向基核函数,对其参数的选择采用交叉验证和网格搜索的方法,用“一对一”的方法来处理故障诊断的多分类问题,建立了支持向量机故障诊断多分类器模型。最后本文采用了专家系统与支持向量机相结合的方法,建立了电梯故障诊断专家系统。实现了专家系统知识的获取、知识的表示、知识库、推理机及解释器。更多还原

【Abstract】 Chinese economy is developing rapidly in recent decades, and the average height of the buildings is also increasing. The elevator has become indispensable important tool for people to live and work. The reliability and safety of the elevator threats people’s life and property at every moment, so it’s necessary to ensure the elevator fault detection and treatment on time. To ensure the reliability of the elevator two things need to be done:one is to test the reliability of the elevator before the elevator officially leaves the factory to be installed; the other is timely detection and treatment when the operation of the elevator fails.According to the current needs of the elevator fault diagnosis, this thesis studies the theory of expert system and support vector machine. The support vector machine is integrated into the expert system to achieve precise reasoning, using the support vector machine’s data mining capability to analyze the elevator fault reasons from the fault data. The expert system establishes, manages and maintains the knowledge base, and uses rules to reasoning the elevator faults and makes appropriate explanations by the interpreter. and provides the explanation mechanism. The system gives full play to their advantages to improve the diagnostic ability. The main contents of this thesis are as follows:Firstly, this thesis introduces the technology of fault diagnosis, analyzes and compares the fault diagnosis methods, and introduces several typical fault diagnosis expert system, and analyzes the common faults of the elevator. Secondly, this thesis builds an elevator fault diagnosis expert system model, selects the production rule representation, forward reasoning to represents and reasonings knowledge respectively. Thirdly, this thesis researches elevator fault diagnosis algorithm, including the basic principle of support vector machine and several key points; builds the support vector machine fault diagnosis model with Radial Basis Function uses Cross-validation and Grid-search to select the parameters, and using "one-versus-one" to deal with fault diagnosis. Finally, this thesis adopts the method of combining expert system and support vector machine to build fault diagnosis expert system for elevators, which implements the expert system knowledge acquisition, knowledge representation, knowledge base, reasoning machine and the interpreter.更多还原