论文部分内容阅读
人工神经网络具有较强的非线性映射能力。本文介绍了神经网络BP算法的一些改进措施。这些措施可以提高BP算法的学习收敛速度,同时也可以提高BP网络性能的稳定性。为避免软土路基沉降传统计算方法中各种人为因素的干扰,本方法利用实测资料直接建模。基于改进的BP神经网络模型,建立了可依据现场量测信息对软基路堤沉降量随时间而发展的过程进行动态预报的分析方法。本文所建立的BP算法模型比较独特,利用该模型预测软土路基沉降精度高,预测结果的稳定性好。
Artificial neural network has strong nonlinear mapping ability. This paper introduces some improvement measures of BP neural network algorithm. These measures can improve learning convergence speed of BP algorithm and improve the stability of BP network performance at the same time. In order to avoid the interference of various human factors in the traditional calculation method of soft soil subgrade settlement, this method uses the measured data to directly model. Based on the improved BP neural network model, an analytical method based on field measurement information is established to dynamically predict the subsidence of soft ground embankment with time. The BP algorithm model established in this paper is rather unique. Using this model to predict the settlement settlement of soft soil roadbed is high and the stability of prediction results is good.