路基智能压实评价指标研究进展综述

A Review of Research Progress on Intelligent Compaction Measurement Values for Subgrade

摘要:

系统总结了智能压实评价指标（ICMV）迄今为止的主要研究成果，讨论了ICMV的发展历程、计算原理、优势与缺陷、填料适用性以及与原位指标拟合的影响因素。分析表明：ICMV与路基填料之间存在一定的匹配机制，工程上应尽量按照该机制选择ICMV以提升检测效果；拟合分析时，应综合考虑填料类型、ICMV类型和原位指标以确定影响因素，避免回归特征的冗余或缺失；鉴于部分特征对预测值存在非线性影响，不建议使用线性模型对智能压实数据进行预测。针对上述问题，本文在ICMV与填料的匹配关系、拟合因素的重要性、回归算法的选择策略三个方面给出了具体的建议及未来的研究方向，以期作为ICMV后续应用及研究的参考。

Abstract:

The major achievements of intelligent compaction measurement value （ICMV） are systematically summarized in this paper. The following contents of ICMV are discussed： development history， calculation mechanisms， advantages and disadvantages， filler applicability and influencing factors between ICMV and in-situ test values. The analysis indicates that there are the matching mechanisms between ICMV and subgrade materials and ICMV should be selected according to these mechanisms to improve the detection effect. During the fitting analysis， the types of subgrade ICMV and in-situ measurement values should be considered comprehensively to determine the influencing factors， so as to avoid the redundancy or the loss of regression features. Since some features have nonlinear effects on labels， it is not recommended to use linear models to predict the labels of intelligent compacted data. In view of the above problems， specific suggestions， including matching relationship between ICMV and materials， the importance of influence factors and the selection strategies of regression algorithms， are presented here in order to serve as references for the future research and application of ICMV.