The Production Prediction of Cutter Suction Dredger Based on PCA

Abstract

In order to ensure the working efficiency of the cutter suction dredger, it is necessary to predict its momentary production accurately. In the actual operation process, the dynamic characteristics of the dredger are very complicated, and many factors affect the momentary production. In order to ensure the smooth output of the dredger, the operator mainly refers the six indicators: the left swing winch speed, the right swing winch speed, the cutter depth, the 1# deck pump speed, the 2# deck pump speed, and the cutter speed to dredging operation. In this paper, Principal Component Analysis (PCA) is performed on the six control factors, and the five main influencing factors are obtained, which are the left swing winch speed, the right swing winch speed, the cutter depth, the 1# deck pump speed, and the cutter speed. Then, according to the average proportion of each factor in the principal component analysis results, the input variables are reduced. The four main influencing factors are the left swing winch speed, the right swing winch speed, the cutter depth, and the cutter speed. In the system simulation modeling, the above five main factors and four main factors are used as the input variables of the GRNN(Generalized Regression Neural Network), and the momentary production is used as the output variable of the GRNN to establish the prediction model of momentary production of the cutter suction dredger. The prediction results show that the neural network model can still maintain good prediction accuracy while reducing the input variables. Therefore, Principal Component Analysis can be used to simplify neural network design and provide real-time momentary production reference for operators on the dredging operation.