Influence of roller-related factors on compaction meter value and its prediction utilizing artificial neural network

Abstract

With the continual innovation of modern technology, the construction of transportation infrastructure is developing toward intelligentization, integration, and automatization. In such a background, Intelligent Compaction (IC) is acknowledged to be novel and promising in compaction industry, but now still imperfect in actual application. The objective of this research is to analyze the influence of roller-related factors on Compaction Meter Value (CMV) representing compaction status of structure and predict CMV based on these factors by Artificial Neural Network (ANN), which is significant for the theory and practice of IC technology. For achieving the goal, a soil compaction project was conducted with IC vibratory roller and the model was established after data processing, architecture design and network training. The analysis results show that as the increase of roller passes, CMV has a slow growth at first followed with a fast increase, and eventually stabilizes. As for vibration force and vibration acceleration, CMV exhibits a generally upward trend with their rises. However, the increment of CMV is relatively lower in comparison with the effect of roller passes. Different from foregoing three factors, CMV increases firstly and drops thereafter with increasing roller speed. The optimal speed rises when soil structure has reached a dense state. The prediction results indicate that the CMV generated by ANN is in good agreement with the measured values. It is reasonable to believe that the ANN-based computational model is a feasible tool with the potential for predicting CMV.