Productivity Prediction and Analysis Method of Large Trailing Suction Hopper Dredger Based on Construction Big Data

Abstract

Trailing suction hopper dredgers (TSHD) are the most widely used type of dredgers in dredging engineering construction. Accurate and efficient productivity prediction of dredgers is of great significance for controlling dredging costs and optimizing dredging operations. Based on machine learning and artificial intelligence, this paper proposes a feature selection method based on the Lasso-Maximum Information Coefficient (MIC), uses methods such as Savitzky-Golay (S-G) filtering for data preprocessing, and then selects different models for prediction. To avoid the limitations of a single model, we assign weights according to the predicted goodness of fit of each model and obtain a weight combination model (WCM) with better generalization performance. By comparing multiple error metrics, we find that the optimization effect is obvious. The method effectively predicts the construction productivity of the TSHD and can provide meaningful guidance for the construction control of the TSHD, which has important engineering significance.