机坪地面空调机组运行状态监测的关键技术研究
发布时间：2019-02-26 14:01
【摘要】：针对机坪地面空调间歇故障引起的使用效能低、维修滞后等问题,近年来通过预测来实时监测设备的运行状态,达到对设备的提前维修。内容涉及数据挖掘算法中的关联Apriori算法的改进,及其改进算法与聚类k-means算法相结合的间歇故障预测方法,并基于此实现了延误维修预测。首先对关联Apriori算法进行了改进。其中针对关联Apriori算法频繁扫描事务数据库低效的问题,通过实时构造间歇故障数组并对其对应项累加求和的方法来提高运行效率。仿真表明:改进后的算法的效率要明显由于原算法。然后基于改进后的AS-Apriori算法进行二次关联,再与聚类k-means算法相结合进行间歇故障预测。并且在初始条件更严格和数据集扩大了10倍的同时,对于处理数据类型和变量的不同,得到两种故障预测结合方法(第二种是第一种的改进方法),并且通过仿真得到:地面空调故障预测第二种结合方法更适合在实际现场海量故障数据的操作。最后,利用延误维修预测估计出永久故障临界区以安排合理维修,主要通过正态分布模型对间歇故障的维修延误堆积预测出永久故障的临界区。仿真表明:预测的维修波及延误累加概率呈线性分布,即可预测性高的间歇故障更便于预先维护管理,减少永久故障的形成。
[Abstract]:Aiming at the problems of low efficiency and delayed maintenance caused by intermittent fault of apron ground air conditioning, in recent years, real-time monitoring of the running state of the equipment has been carried out through prediction, so as to achieve the advance maintenance of the equipment. This paper deals with the improvement of the associated Apriori algorithm in data mining algorithm and the intermittent fault prediction method based on the combination of the improved algorithm and the clustering k-means algorithm. Based on this, the delayed maintenance prediction is realized. Firstly, the associated Apriori algorithm is improved. In order to solve the problem that the associated Apriori algorithm scans transaction database frequently, the efficiency is improved by constructing the intermittent fault array in real-time and adding the corresponding terms to it. Simulation results show that the efficiency of the improved algorithm is obviously due to the original algorithm. Then the improved AS-Apriori algorithm is used to carry out the quadratic correlation, and then combined with the clustering k-means algorithm, the intermittent fault prediction is carried out. And while the initial conditions are stricter and the data set is 10 times larger, for the different data types and variables, two combined fault prediction methods (the second is the first improved method) are obtained. And the simulation results show that the second combination method is more suitable for the operation of mass fault data on the ground air conditioning system. Finally, the critical area of permanent fault is estimated by using the prediction of delay maintenance to arrange reasonable maintenance. The critical region of permanent fault is predicted by normal distribution model for the accumulation of maintenance delay of intermittent fault. Simulation results show that the predicted probability of maintenance and delay accumulation is linearly distributed, that is to say, intermittent faults with high predictability are easier to maintain and manage in advance and reduce the formation of permanent faults.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V351.3;TP311.13
[Abstract]:Aiming at the problems of low efficiency and delayed maintenance caused by intermittent fault of apron ground air conditioning, in recent years, real-time monitoring of the running state of the equipment has been carried out through prediction, so as to achieve the advance maintenance of the equipment. This paper deals with the improvement of the associated Apriori algorithm in data mining algorithm and the intermittent fault prediction method based on the combination of the improved algorithm and the clustering k-means algorithm. Based on this, the delayed maintenance prediction is realized. Firstly, the associated Apriori algorithm is improved. In order to solve the problem that the associated Apriori algorithm scans transaction database frequently, the efficiency is improved by constructing the intermittent fault array in real-time and adding the corresponding terms to it. Simulation results show that the efficiency of the improved algorithm is obviously due to the original algorithm. Then the improved AS-Apriori algorithm is used to carry out the quadratic correlation, and then combined with the clustering k-means algorithm, the intermittent fault prediction is carried out. And while the initial conditions are stricter and the data set is 10 times larger, for the different data types and variables, two combined fault prediction methods (the second is the first improved method) are obtained. And the simulation results show that the second combination method is more suitable for the operation of mass fault data on the ground air conditioning system. Finally, the critical area of permanent fault is estimated by using the prediction of delay maintenance to arrange reasonable maintenance. The critical region of permanent fault is predicted by normal distribution model for the accumulation of maintenance delay of intermittent fault. Simulation results show that the predicted probability of maintenance and delay accumulation is linearly distributed, that is to say, intermittent faults with high predictability are easier to maintain and manage in advance and reduce the formation of permanent faults.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V351.3;TP311.13
【参考文献】
相关期刊论文 前10条
1 陈维兴;曲睿;孙毅刚;;基于改进Apriori算法的地面空调间歇故障预测[J];计算机应用;2016年12期
2 徐加民;王冲;张磊;孙东华;;航空安全系统维护与故障维修技术分析[J];数字技术与应用;2016年11期
3 杨洪富;贾晓亮;任寿伟;;基于数据驱动的航空发动机故障诊断与预测方法综述[J];航空精密制造技术;2016年05期
4 安正;;煤矿机电设备故障诊断及维修技术探析[J];机械管理开发;2016年10期
5 陈亮;;电力系统维修技术的故障排除方法[J];电子技术与软件工程;2016年18期
6 牛猛;;Apriori算法的研究与实现[J];赤峰学院学报(自然科学版);2016年08期
7 李向新;;浅谈航空装备的绿色维修[J];科技视界;2016年09期
8 杨若庸;;民用航空器维修故障原因及质量改进措施[J];企业技术开发;2016年08期
9 孙毅刚;曲睿;陈维兴;王慧敏;;面向数据挖掘的静态电源综合故障诊断研究[J];计算机测量与控制;2015年10期
10 张海峰;胡明华;;航空公司短期航班计划编排模型及算法[J];南京航空航天大学学报;2015年04期
相关博士学位论文 前3条
1 朱霄s，
本文编号：2430840
本文链接：https://www.wllwen.com/shoufeilunwen/xixikjs/2430840.html
