摘要
现代社会，产品质量就是企业的生命线，汽车制造业更是如此，如何及时掌握汽车0部件的质量成为汽车生产厂家深为关注的问题。当汽车厂家得到准确的汽车部件维修数据时已经是汽车出厂多年，厂家不能及时掌握产品的质量也就不能对产品做出指导改进。本课题的研究在1定的程度上可以弥补汽车0部件维修数据向厂家决策层反馈的延时。为厂家高层及时掌握产品质量情况提供参考，从而推动产品质量以及售后服务质量的提高。
本文的研究从奥迪-A6轿车某1维修站点在某段时间内的维修数据表出发，通过对现有的某0件的维修数据表预测未来某个时间此0件的千车故障数。主要解决了横向数据与纵向数据的合理性分析，数据拟合与残差分析以及对未来千车故障数的预测。维修数据表主要包含哪个批次生产的轿车（即生产月份）、售出时间、维修时间、维修部位、损坏原因及程度、维修费用等等。通过这样的数据可以全面了解所有部件的质量情况，若从不同的需求角度出发科学整理数据库中的数据，可得到不同用途的信息，从而实现不同的管理目的。通过对数据分析和预测，就能逐步确定此种0件总的寿命，而更好的指导厂家对质量的改进，也能指导0件维修站对某些0件做提前准备，以免在紧急状况下，因缺货而导致信誉下降，更重要的是给用户带来了损失。
关键词：反馈的延时 数据拟合 残差分析 维修数据表
Abstract
In modern society, the quality of products is the lifeblood of enterprises, and it is the same truth with automobile manufacturing, so manufacturers attach great importance to how to grasp of the quality of a motor vehicle’s parts and components timely. The manufacturers of motor vehicle parts get the accurate identification data of maintenance vehicle only many years after the identification of the vehicles. The manufacturers can not make guiding improvement on products now that they can not control the quality of the products timely. The thesis can compensate, to some extent, the delay of feedback to policymakers on the maintenance data of the automotive components and it can provide the top manufacturers with timely information about the quality of products, thereby promoting the quality of products and enhancing the quality of after-sale service. The thesis studies from the maintenance data tables of Audi -A6 cars in a repair site within a certain period of time, and it forecasts the potential breakdown of one thousand cars in the future through the maintenance data sheet of existing parts. This thesis focuses on analyzing the horizontal and vertical resolution of the main data reasonably, data fitting, residual analysis and the projections of the potential breakdown of one thousand cars in the future and it also illustrates the main batch production cars (namely, the production month), sold time, maintenance time, maintenance parts, causes and extent of damage, repair costs, etc in the maintenance of data tables. The managers can master comprehensively the quality of all the components through such data, and they can attain different information to achieve different management purposes if they collate the data scientifically from different perspectives of demand. Through data analysis and projections, we can gradually establish the total life of such components, and provide better guidance for manufacturers to improve the quality, at the same time, prepare parts depots for some parts in advance, in case the shortage of parts in emergency situations causes the declining of credibility, and more importantly, a loss of the users.
Key words: the delay of feedback data fitting residual analysis maintenance data table
