The dredger construction environment is harsh, and the mud concentration meter can be damaged from time to time. To ensure that the dredger can continue construction operations when the mud concentration meter is damaged, the development of a dredger with advantages of low price and simple operation that can be used in emergency situations is essential. The characteristic spare mud concentration meter is particularly critical. In this study, a data-driven soft sensor method is proposed that can predict the mud concentration in real time and can mitigate current marine mud concentration meter malfunctions, which affects continuous construction. This sensor can also replace the mud concentration meter when the construction is stable, thereby extending its service life. The method is applied to two actual construction cases, and the results show that the stacking generalization (SG) model has a good prediction effect in the two cases, and its goodness of fit R² values are as high as 0.9774 and 0.9919, indicating that this method can successfully detect the mud concentration.

Dredging is a basic construction for waterway improvement, harbor basin maintenance, land reclamation, environmental protection dredging, and deep-sea mining. The dredging process of cutter suction dredgers is so complex that the operational data show strong characteristics of dynamic, nonlinearity, and time delay, which make it difficult to predict the productivity accurately via basic principles models. In this paper, we propose a novel integrating PCA-LSTM model to improve the productivity prediction of cutter suction dredger. Firstly, multiple variables are reduced in dimension and selected by PCA method based on the working mechanism of cutter suction dredger. Then the productivity is predicted via mud concentration in long short-term memory network with relevant operational time-series data. Finally, the proposed method is successfully applied to an actual case study in China. Also, it performs well in the cross-validation and comparative study for several important characteristics: (i) it involves the operational parameters based on the mechanism analysis; and (ii) it is a deep-learning-based approach that can deal with operation series data with a special memory mechanism. This study provides a heuristic idea for integrating the data-driven method and supervision of human knowledge for application in practical engineering.

The suction-lifting system of cutter suction dredgers consumes a large amount of energy. Optimizing their performance is of great significance for enhancing the overall efficiency of dredgers. This study proposes the effective specific cutting energy, a new indicator suitable for evaluating the energy consumption of the cutting system of cutter suction dredgers. It reflects the cooperation state between the cutter system and the pump-pipe system and has important reference value for improving construction efficiency. The calculation method of the effective specific cutting energy is given, which is calculated by the cutter motor power, slurry concentration, and slurry flow rate. Based on the machine learning framework, a model framework for predicting the specific cutting energy according to the relevant parameters of the suction-lifting system is constructed. Real ship data from the cutter suction dredger “Changshi 12” are used for experiments. First, eigenvalue screening is carried out based on the dredging knowledge and mechanism, then outliers are removed, and finally data processing is performed using Spearman correlation coefficient and PCA dimensionality reduction techniques. Subsequently, five machine learning algorithms, such as RF and XGBoost, are used in combination with a grid search to find the optimal hyperparameters, and Lasso is used as the meta-learner to integrate the prediction results. The experimental results show that the Random Forest and Stacking models have high prediction accuracy for slurry concentration, cutter motor power, and slurry flow rate, verifying the feasibility of this method.