摘要
针对传统TOPSIS法的不足,提出一种膨胀土分类的改进TOPSIS法。选取能充分反映膨胀土胀缩特性的液限、塑性指数、小于2μm胶粒含量与自由膨胀率为指标进行分析。综合考虑样本数据的波动信息与独立信息,采用独立信息数据波动(DIDF)赋权法确定权重。然后将灰色关联分析(GRA)与TOPSIS法融合,结合指标权重计算指标分类标准与待分类样本的贴近度。通过比较样本与各类别贴近度的大小确定所属类别,同时还可对同一类别膨胀土胀缩性进行排序,为工程建设提供更详细的参考依据。最后,选取2个工程实例共3 2个膨胀土样本对所建立模型进行验证。结果表明,该方法分类结果与实际较吻合,准确率平均达9 0.6 3%,能满足工程需要。
Aiming at the issues of traditional TOPSIS method, an improved TOPSIS method for the classification of expansive soils was proposed. Four classification indexes which can reflect the swell-shrink characteristics of expansive soils including liquid limit,plasticity index,clay particles,free swell ratio were chosen. Both wave information and independent information of samples data were considered comprehensively,and a new independent information data fluctuation weighting method was used to calculate the weights. Then,the grey relational analysis and TOPSIS method were integrated,and the closeness degrees of the classification standards of indexes and samples were calculated by combining with the weight values. The classification of expansive soils was specified by comparing the magnitude of closeness degrees.This method can also arrange the order of swell-shrink characteristics in the same category,which can provide more detailed reference for engineering construction. Finally,the classification method of expansive soils was validated with 32 testing examples from 2 engineering projects. The results show that the classification results satisfyingly agree with the measured data at the average accuracy of 90. 63 %,which can satisfy the engineering demand.
作者
赵国彦
李鑫
梁伟章
ZHAO Guoyan LI Xin LIANG Weizhang(School of Resources and Safety Engineering, Central South University, Changsha, Hunan 410083, China)
出处
《中国地质灾害与防治学报》
CSCD
2017年第3期130-136,共7页
The Chinese Journal of Geological Hazard and Control
基金
国家自然科学基金面上项目(51374244)
国家重点基础研究发展计划(973)项目(2010CB732004)
关键词
膨胀土
逼近理想解排序法
灰色关联分析
独立信息数据波动赋权法
分类
expansive soil
technique for order preference by similarity to an ideal solution
grey relational analysis
independent information data fluctuation weighting method
classification
