基于RKGM-AR模型的船舶柴油机热力参数趋势预测研究

发布时间：2018-02-28 15:32

本文关键词： 船舶柴油机灰色关联分析组合预测排气温度预警　出处：《大连海事大学》2014年硕士论文　论文类型：学位论文

【摘要】：船舶柴油机作为船舶的“心脏”,其健康状态不仅会影响航行安全,还会影响船舶公司的成本和收益。为了解决船舶柴油机在健康维护和管理过程中出现的欠维修和过维修问题,现在柴油机的维修方式已由定时维修向故障预测与健康管理方式转变。而柴油机热力参数的趋势预测分析为这一转变提供了技术支持,实现了对柴油机的故障状态进行预报。本文以某实习船主机1号气缸排气温度为主要预测对象,提出一种组合预测模型对其进行趋势预测分析,以实现对柴油机进行故障预报。首先,对柴油机常规的热力参数进行分析研究,阐述了灰色关联分析方法的用途和计算原理,并采用灰色关联分析法对柴油机典型热力参数进行聚类分析,得到排气温度的关联参数。其次,分析几种常用预测方法的优劣,提出组合预测模型是将来的发展趋势,并建立经四阶龙格库塔法改进的灰预测模型与时间序列AR模型相结合的组合预测模型,分别发挥了上述两种预测模型的优势。再次,通过对排气温度的报警限值和预警等级界定的计算方法进行研究,实现了预警功能,并以排气温度作为主序列,各缸平均排气温度、扫气温度、主轴承出口滑油温度、气缸冷却水出口温度作为辅序列,分别选择柴油机排气温度在平稳变化和上升变化时上述五个参数的样本数据,采用联合预测的方法对排气温度进行趋势预测分析。最后,将组合预测模型应用于实船,并将实船排气温度的预测值与实测值进行比较和误差分析,以验证预测模型的有效性。
[Abstract]:As the heart of the ship, the health status of the marine diesel engine will not only affect the safety of navigation, but also the cost and income of the shipping company, in order to solve the problem of undermaintenance and overmaintenance of the marine diesel engine in the process of health maintenance and management. Now the maintenance mode of diesel engine has been changed from regular maintenance to fault prediction and health management, and the trend prediction analysis of diesel engine thermal parameters provides technical support for this change and realizes the prediction of diesel engine fault state. In this paper, the exhaust temperature of the main engine No. 1 of a practical ship is taken as the main prediction object, and a combined forecasting model is put forward to forecast the trend of the engine in order to realize the fault prediction of the diesel engine. Firstly, the conventional thermodynamic parameters of diesel engine are analyzed, the application and calculation principle of grey correlation analysis method are expounded, and the typical thermodynamic parameters of diesel engine are analyzed by cluster analysis. Correlation parameters of exhaust temperature are obtained. Secondly, the advantages and disadvantages of several commonly used forecasting methods are analyzed, and the combined forecasting model is proposed as the development trend in the future, and the combined prediction model which combines the grey prediction model with the AR model of time series improved by the fourth order Runge-Kutta method is established. The advantages of the above two prediction models are brought into play respectively. Thirdly, by studying the alarm limit value of exhaust temperature and the calculation method of warning grade, the function of early warning is realized, and the exhaust temperature is taken as the main sequence, the average exhaust temperature of each cylinder, the scavenging temperature, the oil temperature at the outlet of the main bearing, and the oil temperature at the outlet of the main bearing. The outlet temperature of cylinder cooling water is taken as the auxiliary sequence, and the sample data of the five parameters mentioned above are selected respectively when the exhaust temperature of diesel engine changes smoothly and rising, and the trend of exhaust temperature is predicted and analyzed by using the method of joint prediction. Finally, the combined prediction model is applied to the real ship, and the prediction value of the exhaust temperature of the ship is compared with the measured value and the error analysis is carried out to verify the validity of the prediction model.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U664.121

【参考文献】