Trailing suction hopper dredger operation parameter learning method

Abstract

The invention relates to a method for learning operation parameters of a trailing suction hopper dredger. The technical scheme is as follows: the method comprises the following steps: ; randomly dividing the m effective data groups into a training set, a cross validation set and a test set, 4-8 multi-layer sensor models are established, the loss functions of all the multi-layer sensor models of the training set are obtained by adjusting five hyper-parameters, and the model with the minimum loss function is selected from the loss functions of all the multi-layer sensor models of the cross validation set to serve as an optimal model, namely an implicit model; an input explicit model and an output explicit model are established through an implicit model and a regression model, the multi-objective optimization problem of the expected value c1 of the yield and the expected value c2 of the mixture density is solved through an ant colony algorithm, and an optimal parameter combination is obtained. The model is good in stability, high in accuracy and high in universality, and optimal parameter combination can be achieved under the condition that the expected value c1 of the given yield andthe expected value c2 of the mixture density are met.

Description

technical field [0001] The invention belongs to the technical field of dredger operation parameter learning. In particular, the invention relates to a learning method of operating parameters of a trailing suction dredger. Background technique [0002] The main task of a trailing suction dredger is to dig sediment from the seabed or river bed while sailing and transport it to a designated area. The maneuverability and efficiency of a trailing suction dredger make a trailing suction dredger a large land An indispensable machine for reclamation projects, improving the dredging efficiency of trailing suction dredgers and reducing the workload of operators has become a hot issue in current research. [0003] Braaksma et al. (J.Braaksma, J.B.Klaassens, R.Babuska and C.Keizer.Modelpredictive control for optimizing the over dredging performance of a trailingsuction hopper dredger.Proceedings of the Eighteenth World Dredging Congress, pp.1263-1274, 2007.) dedicated In order to opti...

AI Technical Summary

Problems solved by technology

The model studied by Li et al. is not stable enough and the accuracy is not high. The main research focuses on the model between the controllable parameters and the yield rate. Although it is helpful to predict the yield rate with the given parameters, it cannot meet the requirements. The optimal combination of operating parameters is given in the case of production rate

The studies by Braaksma et al. and Li et al. do not provide a general method for similar engineering

Images