项目名称: 基于模型-传感器信息融合的典型液压设备故障预测方法研究
项目编号: No.51275524
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 机械、仪表工业
项目作者: 李洪儒
作者单位: 中国人民解放军军械工程学院
项目金额: 80万元
中文摘要: 液压设备广泛应用于航空航天、武器装备等重要领域,通过对故障预测与健康管理(PHM)理论的研究与应用,可有效提高液压设备的运行可靠性。本项目针对故障预测准确度亟待提高这一突出问题,拟将基于模型的故障预测方法与基于传感器(数据)的故障预测方法结合起来,应用信息融合技术建立基于模型-传感器的故障预测新方法,提高故障预测的置信度水平,进一步推动液压设备PHM系统的研究与应用。具体研究内容包括:①应用自适应随机共振理论,研究强噪声背景下液压设备潜在故障微弱特征信号检测新方法;②组合应用小波分析、经验模态分解和数学形态学等技术,研究液压设备潜在故障特征提取新方法;③通过性能劣化试验,研究典型液压设备的一般劣化规律,建立其性能劣化模型;④改进D-S证据理论中的可信度函数和组合规则,以利于对模型信息与传感器数据进行有效融合,并进一步建立基于模型-传感器的典型液压设备故障预测新方法。
中文关键词: 液压设备;微弱信号检测;退化特征提取;信息融合;故障预测
英文摘要: Widely used in aerospace and weapons, hydraulic equipment's operational reliability can be effectively enhanced by research and application of prognostics and health management(PHM). To meet the urgent need for fault prognostics accuracy enhancement, this project plans to establish one new fault prognostics method based on model-sensor, utilizing information fusion technology to integrate the both methods based on model and on sensor. The new method can enhance the confidence level of fault prognostics and promote the research and application of hydraulic equipment PHM system. The project research includes: 1)New method on hydraulic equipment incipient fault weak characteristic signal detection in strong noise background, utilizing adaptive random resonance theory;2)New method on hydraulic equipment incipient fault feature extraction, utilizing wavelet analysis, empirical mode decomposition and mathematical morphology;3)Research on universal deterioration law of typical hydraulic equipment to establish its performance deterioration model, by performance deteriotation tests;4)Improvement method on probability assignment and combination rules in D-S evidence theory to effectively fuse model information and sensor data, and new method on fault prognostics for typical hydraulic equipment based on model-sensor.
英文关键词: hydraulic equipment;weak signal detection;feature extraction;information fusion;fault prognostics