Regression analysis of some factors influencing soil compaction

Experiments were conducted in a laboratory soil bin, at the Regional Research Center of Asian Institute of Technology, to develop compaction models for a silty clay loam soil. The development of the models made use of dimensional analysis techniques. Three independent parameters were investigated: (1) tire variables (section width, diameter, inflation pressure), (2) soil variables (moisture content, initial cone index), and (3) external variables (travel speed. axle load, number of passes). Bulk density and cone index were considered as dependent variables. Results showed that axle load and number of tire passes were the most prominent factors which greatly influence soil compaction. Furthermore, soil moisture content, aspect ratio, and tire inflation pressure also revealed significant effects. The greatest soil compaction occurred during the first three passes of the tire. Soil compaction models were established and were found to provide good predictions. The trend established by the models signifies that general relationships can be established to predict soil compaction related to soil types. Furthermore, the models provided predictions at different soil and machine working conditions. Using the models, assessment of soil compaction can be made to develop a decision support system to establish useful recommendations for appropriate soil management practices and solutions to site-specific soil compaction problems. (C) 2001 Elsevier Science B.V. All rights reserved.