基于隐马尔科夫模型的装载机载荷谱编制方法研究

发布时间：2018-04-24 14:38

本文选题：工程车辆 + 载荷谱　；参考：《吉林大学》2017年硕士论文

【摘要】：工程机械作为装备制造业的重要组成部分,所处的作业环境复杂多变,其关键零部件的可靠性和耐久性受多方面因素的影响,而这也一直是制约其质量提升及品质发展的关键。载荷谱是实现工程车辆服役性能测试、合理正向设计、室内模拟加载试验、寿命预测和疲劳分析等的重要参考。实现实测载荷的统计分析,外推方法的合理选择以及外推方法的优化,使得外推得到的长期载荷谱的准确性和真实性得到保障是载荷谱研究领域的重点同时也是难点。为提升国产装载机的自主研发水平,本文依托项目“工程机械节能减排共性技术研究”(项目号:SQ2015BA0400198),结合隐马尔科夫模型理论,以某型装载机为例,对在载荷谱编制中具有不可忽略影响的人为因素进行了研究。同时,对由于分段作业产生的非平稳载荷提出了处理方法。本文的主要研究内容如下:1.本文以装载机为研究对象,对装载机载荷谱的测试方案进行了详细的阐述。包含试验样机的情况、测试场地以及物料准备、测试方案的制定、采集子样大小的确定、采集测点以及传感器的安装等。并以实测装载机举升缸压力数据为例,完成了趋势项消除、数字化滤波、奇异点检测等数据预处理。2.针对工程车辆操作时人为操作频繁,载荷特性受操作方式影响大的问题,本文综合多方面影响因素,将驾驶员的操作类型分为普通型、粗暴型、温和型三类。由于隐马尔科夫模型在模式识别及故障诊断领域良好的表现将其应用于驾驶员操作类型的识别上,首先采用Baum-Welch算法对载荷数据进行训练,然后进行对数似然率的匹配,识别驾驶员操作类型。针对三种不同类型的操作方式及载荷特点,分别选择雨流矩阵外推、POT外推、参数估计外推。使驾驶员操作的人为因素在载荷谱外推方法的选择中得到充分的考虑,实现外推方法的合理选择。3.针对工程车辆由于工况复杂多变以及循环作业方式造成的载荷整体不平稳的特点,采用分段方法实现测试载荷的平稳化处理,对各段载荷的平稳性进行检验。利用隐马尔科夫模型对装载机工作模式切换时产生的过渡载荷进行处理,将采用分段方法处理时被忽略的转折点信息包含进载荷处理的过程中,并将两种方法处理得到的全雨流矩阵以及期望雨流矩阵进行对比,验证了方法的可行性。在编谱过程中包含更多对疲劳具有贡献的信息,有利于提高编谱的精度。通过MATLAB开发了一款“工程车辆非平稳载荷处理分析软件V1.0”。为研发工作者提出了非平稳载荷处理的方法,完成了载荷处理的便捷化和规范化。本文针对装载机载荷特点,基于隐马尔科夫模型对载荷谱的编制方法的合理选择以及优化进行了研究。实现了驾驶员操作特性的识别并提出了非平稳特性载荷的处理方法。实现提高载荷谱编制精度,为工程机械零部件的可靠性分析设计提供准确依据的目的。
[Abstract]:As an important part of the equipment manufacturing industry, construction machinery is in a complex and changeable working environment. The reliability and durability of its key parts are affected by many factors, and this has been the key to its quality improvement and quality development. Load spectrum is an important reference for engineering vehicle service performance test, reasonable forward design, indoor simulated loading test, life prediction and fatigue analysis. To realize the statistical analysis of the measured load, the reasonable selection of extrapolation method and the optimization of extrapolation method, the accuracy and authenticity of the extrapolated long-term load spectrum are guaranteed, which is the emphasis and difficulty in the field of load spectrum research. In order to improve the level of independent research and development of domestic loaders, this paper relies on the project of "study on common technologies of energy saving and emission reduction of construction machinery" (project number: SQ2015BA0400198m), combining with the theory of hidden Markov model, taking a certain type of loader as an example. The human factors which can not be ignored in the compilation of load spectrum are studied. At the same time, the non-stationary load caused by subsection operation is treated. The main contents of this paper are as follows: 1. In this paper, the load spectrum test scheme of loader is described in detail. Including the situation of the test prototype, test site and material preparation, the formulation of the test plan, the determination of the size of the sampling sub-sample, the acquisition of measuring points and the installation of sensors, etc. Taking the measured data of lift cylinder pressure of loader as an example, the preprocessing of data, such as eliminating the trend term, digitizing filtering and detecting singularity, is completed. In view of the problem that the manual operation is frequent and the load characteristic is greatly influenced by the operation mode, the operating types of the driver are classified into three types: common type, rough type and mild type. Because of its good performance in the field of pattern recognition and fault diagnosis, Hidden Markov Model is applied to the identification of driver's operation type. First, the load data is trained by Baum-Welch algorithm, and then the logarithmic likelihood ratio is matched. Identify the driver's type of operation. According to three different operation modes and load characteristics, rain flow matrix extrapolation and parameter estimation extrapolation are selected respectively. The artificial factors of driver operation are fully considered in the selection of load spectrum extrapolation method, and the reasonable selection of extrapolation method is realized. In view of the complex and changeable working conditions of the construction vehicle and the unsteady load caused by the cyclic operation mode, a piecewise method is adopted to realize the stabilization of the test load, and the smoothness of the load in each segment is tested. Using Hidden Markov Model to deal with the transient load generated when the loader's working mode is switched, the information of the turning point which is ignored by the piecewise method is included in the process of load processing. The feasibility of the method is verified by comparing the total rain flow matrix and the expected rain flow matrix obtained by the two methods. More contributions to fatigue are included in the process of spectrum compilation, which is helpful to improve the accuracy of spectrum compilation. A non-stationary load processing software V1. 0 for engineering vehicles is developed by MATLAB. The method of nonstationary load processing is proposed for R & D workers, and the convenience and standardization of load processing are accomplished. Based on hidden Markov model, the reasonable selection and optimization of load spectrum are studied in this paper. The identification of driver's operating characteristics is realized and the method of dealing with non-stationary load is put forward. The purpose of improving the precision of load spectrum compilation and providing accurate basis for reliability analysis and design of construction machinery parts.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH243

【相似文献】