基于灰色模型的手机灰色故障率的研究及应用

摘要

The prediction of failure rate is generally divided into individual-based prediction and overall-oriented prediction.The prediction of the overall failure rate is accompanied by a large number of individual randomness which leads to difficult to predict.The article creatively defined grey fault rate,grey model fault count,grey model total count and the corresponding formulas.By using the gray model to counteract the random error, the influence of individual randomness on the overall failure rate prediction was solved.In addition, by using the conventional bath curve characteristics in the failure rate curve,the bottom plateau of the bath curve was defined as the failure rate characteristic value.The number of new mobile phone failure was predicted through the old and new phone failure rate eigenvalue ratio and the actual cell phone handset fault data.We got accurate forecast of about 2%deviation in three years.%故障率的预测一般分为针对个体的预测和针对整体的预测,整体故障率预测同时伴随着其内部大量个体的随机性导致预测困难.结合灰色模型和故障率定义,创新性地定义了灰色故障率、灰色故障数量和灰色样本数量及相应的计算公式.利用灰色模型针对随机误差的抵消作用,解决了个体故障随机性对整体故障率预测的影响.另外,利用故障率曲线中常规的澡盆曲线特征,定义澡盆曲线的底部平稳段取值为故障率特征值,并借助新旧手机的故障率特征值比值及旧手机的实际手机故障数据来预测新手机故障数量,取得了三年偏差2%左右的精确预测结果.