由于疲劳寿命试验具有难度较大 ,其实验数据具有分散性强的特点 ,如果直接用灰色预测模型 (GM (1 ,1 ) )作预测 ,其预测值与真实结果的误差较大。在此提出一种试验仿真的新方法。首先将数据处理成对数序列 ,以降低其数据分散性 ;然后再用灰色预测模型进行预测。实验表明 ,经过这样处理后 ,由试验数据所得到的累加序列具有更好的规律性。这时再用GM(1 ,1 )模型可以得到较好的预测结果 Because the fatigue life test is more difficult and its experimental data has the characteristics of strong dispersion, if the gray prediction model (GM (1, 1)) is used as the prediction, the error between the predicted value and the true result is larger. Here a new method of experimental simulation is proposed. The data is first processed into a logarithmic series to reduce its data scatter; then the gray prediction model is used to predict. Experiments show that after such processing, the accumulative sequence obtained from the experimental data has a better regularity. In this case, the GM (1, 1) model is used to get a better prediction result