基于BP神经网络的冻土路基沉降预测

摘要

通过采用BP神经网络对铁路建设期间冻土温度变化及冻土路基沉降进行了评价和预测,该方法减少了人为因素的干扰,有很强大的非线性映射能力,可以相对精确进行较长时间、较大范围的冻土路基沉降预测,弥补了现场实测数据的不足,可推广应用。%This paper using BP neural network made evaluation and prediction to temperature changes and roadbed settlement of frozen soil during the period of railroading,this method reduced the interference of human factors,had very strong nonlinear mapping ability,could be relatively accurate for a long time,wide range of frozen soil roadbed settlement prediction,made up the lack of field test data,could be widely used.