机械设备波动运行状态参数的预测方法研究

发布时间：2019-06-24 12:41

【摘要】：本文基于国家自然科学基金资助项目—非线性旋转机械转子系统的突变故障预测研究（50975105）撰写。随着生产技术的高速发展，电力、机械等行业的设备日趋大型化和精密化，设备的健康状况对企业的安全、经济生产具有重要意义。预测技术可以从设备历史状态参数的发展趋势中挖掘其变化规律，对潜在的故障隐患进行预报，为合理安排维修计划提供技术支持，从而保证设备的安全运行。本文首先针对机械设备振动状态数据通常具有波动性的问题，提出了一种利用马尔科夫方法对灰色预测结果修正的预测模型。首先利用灰色等维新息GM(1,1)模型对样本数据进行灰预测，根据状态实测数据与其灰预测结果之间的误差百分比划分马尔科夫状态区间，建立马尔科夫状态转移概率矩阵。在对设备状态进行预测的时候，利用马尔科夫状态转移概率矩阵和当前状态的误差百分比状态向量计算得到马尔科夫修正值，对灰色预测结果进行修正，实现对波动状态参数的预测；此外，本文从寻求新方法的角度出发，通过对蚁群算法的学习，建立了基于蚁群算法的信号重构波动运行状态参数预测模型，首先对波动性数据使用信号重构处理之后让其在一定范围内波动，然后采用蚁群算法中信息素的思想来对数据进行预测；最后，本文介绍了作者在研究生期间的主要科研项目-国内某汽车制造厂商的发动机工厂设备点检与维修信息管理系统，并介绍在该系统中如何实现对设备的故障预测。通过对潜油泵的实例证明，首先本文所建立的灰色-马尔科夫波动运行参数预测模型，不论在数据预测精度还是在对波动性数据的趋势预测上都有不错的效果；其次基于蚁群算法的信号重构波动运行状态参数预测模型作为一种新的预测模型具有很高的预测精度，，而且对数据趋势的预测也具有较好的结果；最后通过在系统中建立预测模型，将预测技术用到实际的企业中，为工程师对设备状态的判断提供更多的信息，使得判断结果更加具有科学性，更加具有可信性。
[Abstract]:This paper is based on the sudden fault prediction of nonlinear rotating machinery rotor system, which is supported by the National Natural Science Foundation of China (50975105). With the rapid development of production technology, the equipment in power, machinery and other industries is becoming larger and more refined. The health of the equipment is of great significance to the safety and economic production of enterprises. The prediction technology can excavate the change law from the development trend of the historical state parameters of the equipment, predict the potential hidden trouble, and provide technical support for the reasonable arrangement of the maintenance plan, so as to ensure the safe operation of the equipment. In this paper, a prediction model based on Markov method is proposed to modify the grey prediction results in order to solve the problem that the vibration state data of mechanical equipment are usually fluctuating. Firstly, the grey equal dimension innovation GM (1, 1) model is used to predict the sample data. According to the error percentage between the measured state data and the grey prediction results, the Markov state interval is divided, and the Markov state transition probability matrix is established. When predicting the state of the equipment, the Markov correction value is obtained by using the Markov state transition probability matrix and the error percentage state vector of the current state, and the grey prediction results are modified to realize the prediction of the fluctuation state parameters. In addition, from the point of view of finding new methods, through the study of ant colony algorithm, a prediction model of wave operating state parameters based on ant colony algorithm is established. Firstly, the volatility data is reconstructed and then fluctuated in a certain range, and then the idea of pheromone in ant colony algorithm is used to predict the data. Finally, this paper introduces the equipment spot inspection and maintenance information management system of an automobile manufacturer in China, which is the main scientific research project of the author during the graduate period, and introduces how to realize the fault prediction of the equipment in the system. Through the example of submersible oil pump, it is proved that the grey-Markov fluctuation operation parameter prediction model established in this paper has a good effect both in the accuracy of data prediction and in the trend prediction of volatility data. Secondly, the signal reconstruction fluctuation state parameter prediction model based on ant colony algorithm has high prediction accuracy as a new prediction model, and also has good results for the prediction of data trends. Finally, by establishing the prediction model in the system, the prediction technology is applied to the actual enterprise, which provides more information for the engineer to judge the state of the equipment, and makes the judgment result more scientific and credible.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TH17

【参考文献】