The dredging production of the cutter suction dredger directly reflects the dredging efficiency.In order to increase the production,it’s effective to establish the production model of the cutter suction dredger to predict the production.There are many factors that affect the production of the cutter suction dredger,and each factor affects each other.Therefore,the fitting accuracy of the production mechanism model of the cutter suction dredger established according to the production influencing factors is poor,and the black box model of production is decided to build.Common neural networks,such as BP(Back Propagation),RBF(Radial Basis Function),etc.,have a long training time and ordinary generalization ability,which cannot meet the requirements of production prediction.Therefore,the optimized Extreme Learning Machine(ELM)algorithm is applied to establish a production prediction model in this thesis.Firstly,preprocessing the data collected by the actual ship,and use the processed data as the training and testing data set of the black box model.And,according to the production calculation formula of the cutter suction dredger,the structure of the cutter suction dredger,and the dredging process,eight factors that affect the production are selected.And five principal components are obtained by principal component analysis between eight factors and production.And through the comparison of the prediction results of the RBF model,it can be seen that the principal component analysis can reduce the complexity of the model while ensuring the prediction effect of the model and reduce the training time of the model.Therefore,the input variables of the black box model were reasonably selected through principal component analysis,and the structure of the model was determined.Then,the production prediction model based on optimized extreme learning machine algorithm is established.Regularized parameter is utilized to eliminate structural risks in ELM,and the Regularized Extreme Learning Machine(RELM)is established;Particle Swarm Optimization(PSO)is utilized to perform global optimization on randomly selected parameters in RELM,and the Particle Swarm Optimization Regularized Extreme Learning Machine(PSO-RELM)is established,which can improve the robustness and generalization ability of RELM.The simulation experiment results show that the PSO-RELM model can realize the forecast of the dredging production of the dredger and has a good effect.Finally,Echarts(Enterprise Charts)is utlized to visually display and analyze the dredging information.And the relevant dredging strategy based on the expert’s experience is obtained.Combining the dredging strategy with visualization can realize the visual auxiliary operation of the dredging process to improve the efficiency of dredging. |