路面使用性能预测与统计分析

提出用重构背景值的GM(1，1)模型对路面使用性能试验数据进行预测，新的背景值计算公式的显著特点是使GM(1，1)模型具有对建模结果进行优化的能力，能获得最佳的拟合和预测精度。将获得的预测序列进行统计分析，为路面性能预测评价提供参考依据。实例计算结果说明，将重构背景值的GM(1，1)模型用于路面使用性能试验数据预测有很高的精度，为有效缩短试验时间提供了一个值得探讨的方法。同时从统计分析中可知，对数正态分布模型更适合于弯沉指标。

Prediction and statistical characteristic analyses of pavement performance

The restructure background value GM(1,1) model is proposed to predict the test data of index of pavement performance. An obvious feature of the new calculating formula of background value is that it makes GM(1,1) model have the ability to optimize the modeling results, the best fitting and forecasting precision can be obtained with it. The reference to evaluate pavement performance can be provided by statistical analysis of forecasting sequence. The calculation result shows that a high accuracy could be obtained by applying the restructure background value GM(1,1) model to prediction of the test data, and a valuable method is proposed for shortening the test time of indexes of pavement performance. In the statistical analyses, the result shows that the lognormal distribution model is more suitable for the index of deflection.