Real-time evaluation of compaction quality based on RF- ACGWO with high robustness and generalization ability
The effective evaluation of the compaction quality is a key issue for the safety of the earth-rock dam. However, the existing prediction models of compaction quality are devoted to improving the prediction accuracy, ignoring the generalization ability and robustness, resulting in the deviation of practical evaluation results, which is inapplicable to the complex construction environment of the project. To address the abovementioned problem, a novel real-time evaluation model for construction unit compaction quality based on random forest optimized by adaptive chaos grey wolf (RF-ACGWO) was proposed, in which, the RF was used for compaction quality prediction, while the ACGWO was used to overcome the deficiency of low efficiency and accuracy for traditional RF parameter selection, and to improve the generalization ability and robustness of the model. Furthermore, the meteorological factors on the project site were considered as one of the influence parameters to improve the accuracy of the model. After embedding the 3D rolling monitoring system with the proposed method, the real-time evaluation, guidance and feedback on the project site can be realized. The application of a large-scale and real-life hydraulic engineering demonstrates that the RF-ACGWO has the highest accuracy of 0.838, the best generalization ability of 0.793 and the most stable robustness comparing with conventional MLR, BPNN and RF method.