本文通过对新车型开发流程的调查，以过往车型的品质不良数据为基础样本，结合
威布尔分布理论和品质管理中的具体课题，建立了严谨有效的数学模型，利用 Minitab
软件对该模型中的参数进行量化，提出了一种实用且有效的汽车不良品质预测方法。
首先，主要介绍了汽车行业品质预测方法的研究现状，针对正态分布和曲线拟合方
法在整车品质预测应用中的不足和局限性，提出了一种新的预测方法——威布尔分布函
数预测方法，并对该方法的基础理论进行了详细阐述。
其次，对威布尔分布函数预测方法的应用流程进行了研究。通过对原车型过往数据
的整理，得到关键失效零件清单，应用威布尔分布对关键零件的失效数据进行分析，从
关键失效零件的样本数据中估计零件件的重要寿命特征，应用该重要寿命特征参数，结
合保修政策的变动，对这些零件在保修期限内的累计不良率进行预测，并考虑了对策效
果对关键零件累计不良率的影响，然后将关键零件的累计不良率叠加，从而预测出新车
型上市 1 年内销售的车辆在保修期限内的累计失效率。
最后，对威布尔分布函数预测方法的有效性进行了评价。结果表明：此预测方法所
得出的品质目标曲线与实际市场发生的品质不良曲线相吻合，证明该方法可以有效预测
出新车的品质状况，且该方法理论依据和数学建模过程严密，操作过程易于理解，有较
强的可操作性。
关键词：汽车；品质预测；威布尔分布 ；累积失效率
Abstract
With the consumers demanding on the quality of the automobile increased, the market
share and profits of the automobile enterprise was guaranteed by the quality management and
cost control in the intense market environments. Product failure research has been seen as a
quality management goal setting important topic. Meanwhile, with the development of quality
management quantifying the trend and the warranty policy changing, develop new products in
the future scientific and effective quality management target gradually becoming the
mainstream of the major car companies research direction.
In this article, a predict method was introduced for quality fail on automobile which was
based on the investigation of the process for the development of new models, and the passed
vehicle of the data samples, combining with the weibull distribution theory and specific topics
in quality management, the rigorous and effective mathematical model is established, using
Minitab software to quantify the parameters in the model, a practical and effective quality
prediction method was proposed in this paper
First, mainly introduced the research status of the current automotive industry quality
prediction method, the shortage and limitation of the normal distribution and the curve fitting
method applied in automobile enterprise was pointed out ,a new forecasting method called A
Predict Method of Quality Fail on Automobile was proposed and studied .
Second, the application of weibull distribution function prediction method process was
studied. Based on the original model data in the past automobile, the key failure parts list was
selected, the weibull distribution method was analyzed on the failure data of these key parts
to get the important life characteristic parameter, the accumulative failure rate of these key
parts was predicted combining with warranty policy changing and the characteristic parameter,
also the countermeasure effect of these key parts was considered. the cumulative failure rate
of these key parts was added to predict the new vehicle accumulative failure rate within 1 year
sales during the new warranty period.
Finally, the effectiveness of the prediction method was evaluated. the actual market
results was satisfied with the predicted quality target curve, which was proved that the method
could effectively predict the quality of the new vehicle , the prediction method was based on
the theoretical basis and mathematical modeling process, which is easy to understand the
operation process.
Keywords: Automobile; quality fail prediction; Weibull distribution; Accumulated failure rate
==>>