基于BP神经网络的改良盐渍土强度预测研究  

Study on the prediction of the strength of improved saline soil based on BP neural network

作　　者：李治斌 袁立月 丁黔 张淼鑫 王红梅[1] LI Zhibin;YUAN Liyue;DING Qian;ZHANG Miaoxin;WANG Hongmei(College of Construction Engineering,Heilongjiang University,Harbin 150080,China;College of Water Resources and Electric Power,Heilongjiang University,Harbin 150080,China)

机构地区：[1]黑龙江大学建筑工程学院,黑龙江哈尔滨150080 [2]黑龙江大学水利电力学院,黑龙江哈尔滨150080

出　　处：《水利科学与寒区工程》2021年第5期51-54,共4页Hydro Science And Cold Zone Engineering

基　　金：冻土工程国家重点实验室开放基金(SKLFSE201802,SKLFSE201919)。

摘　　要：为了研究固化盐渍土的力学性能,在实验室内配制成人工盐渍土,分析固化剂掺量、温度和加载速率3个影响因素,通过改良土的无侧限抗压试验,得到了48组不同影响因素下人工盐渍土样的强度值。根据所得数据结果,利用BP神经网络建立改良盐渍土的单轴抗压强度预测模型,预测结果误差均小于0.1,且均方根误差为0.027,确定系数为0.985,表明该模型具有良好的预测效果。In order to study the mechanical properties of solidified saline soil,the artificial saline soil was prepared in the laboratory,and the unconfined compressive tests of the improved soil were carried out considering the dosage of curing agent,temperature and loading rate,and the strength values of 48 soil samples under different influencing factors were obtained.According to the results of the obtained data,the BP neural network was used to establish the uniaxial compressive strength prediction model of the improved saline soil.The error of the prediction results was less than 0.1,and the root mean square error was 0.027,and the determination coefficient was 0.985.These showed that the model had good prediction effect.

关 键 词：盐渍土改良 固化剂 无侧限抗压强度 BP神经网络 强度预测模型 

分 类 号：S156.41[农业科学—土壤学]