基于全息电功率图形分析的机电液系统在线故障诊断方法研究

项目名称： 基于全息电功率图形分析的机电液系统在线故障诊断方法研究

项目编号： No.51305326

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 机械、仪表工业

项目作者： 张征凯

作者单位： 西安建筑科技大学

项目金额： 20万元

中文摘要： 针对大型机电装备在极端工况下表现出的性能退化﹑可靠性降低以及故障形成机理研究缺乏问题，以提高设备运行安全性、可靠性、故障诊断与运营维护技术水平为目的， 融合三相电参量提供的幅值、相位、相间和相序信息创建全息电功率图形，运用全息电功率 图形分析系统耦合特性及其故障演化机理。揭示多过程耦合（机电耦合→机液耦合→界面耦 合）中，工况载荷与流体耦合因素对系统全局动态性能与运行状态的影响规律，建立系统耦合特性及运行状态影响电功率流平稳与非平稳流动的数字特征，构建电功率图形及表征电功率流数字特征的图示化方法；阐明电功率的时频特性以及工频和高次谐波电功率图形在系统不同耦合特性及运行状态下的能量转换、合成、传递和存储的数学原理与物理意义；利用电功率时频特征提取系统故障演化特征信息，建立基于电功率图形分析的机电液系统耦合状态评价与运行状态识别理论与方法，为设备节能控制以及故障机理和预示技术研究提供新理论和新方法。

中文关键词： 电功率图形；机电液系统；在线故障诊断；特征提取；人工智能技术

英文摘要： In view of lack of research about degradation of performance, decrease of reliability, and fault forming mechanism under extremely harsh condition of large-scale mechatronics equipment, holographic-electric-power graph fused by amplitude, phase, time, and phase sequence of three phase electric parameters is utilized to research coupling characteristics and mechanism of fault evolution of mechanical-electrical- hydraulic system based on the purpose of improving the security, reliability, fault diagnosis, and operation maintenance level of facility. The effect of load conditions and coupling factors on the dynamic performance and running condition of system is revealed during multi-coupling (electric-mechanical coupling→Hydraulic - mechanical coupling→interfacial coupling) procedure. The graph description way and character extraction of electric power flow both in stable and unstable condition are searched. Time-frequency features of electric power are analysed. Mathematic theories and physical implications of basic and harmonic electric power graphs are illustrated which can show effective energy transformation, transfer, synthesis and storage under different dynamic couple characteristics and running conditions of system. Characteristic information of fault evolution can be extracted from time-frequency characte

英文关键词： Electrical graphics；mechanical electrical and hydraulic system；on-line fault diagnosis；feature extraction；artificial intelligence