Research On Prediction Method Of Ship Energy Consumption Based On Multi-model Fusion

The green and intelligent development of ships is an inevitable requirement for implementing the national development strategies of "dual carbon" and "marine power".Ship intelligent energy efficiency management is one of the effective methods to realize the green and intelligent development of ships.In order to further enhance the level of intelligent energy efficiency management for ships,it is necessary to conduct more accurate and efficient research on ship energy consumption prediction methods and technologies.This thesis focuses on a large ocean bulk carrier as the research object and conducts research on intelligent prediction methods for ship energy consumption.To address the issues of low accuracy and weak adaptability of existing ship energy consumption prediction models in relation to sailing environments,a self-adaptive ship energy consumption prediction method based on intelligent model selection fusion is proposed.A ship energy consumption prediction fusion model is constructed,which considers the influence of sailing environmental factors and exhibits high accuracy and strong adaptability.Comparative analysis of ship energy consumption prediction is performed,thus providing theoretical methods and research directions for ship energy consumption prediction and energy efficiency optimization.The main work of this thesis is as follows:(1)Conducted data mining and analysis of ship energy efficiency data and navigation environmental data.Based on the collected ship energy efficiency data and navigational meteorological data,the spatiotemporal distribution characteristics of ship energy consumption influencing factors were identified,revealing the correlation between ship energy efficiency level and environmental factors,thereby laying the data foundation for ship energy consumption prediction considering the spatiotemporal differences of ship navigation environment and energy efficiency data.(2)Establishment of ship energy consumption prediction models considering multiple influencing factors.Based on the analysis of spatiotemporal distribution characteristics of ship energy efficiency,multiple ship energy consumption prediction models based on machine learning were constructed.By evaluating the model performance indicators and the fitting effects,a multidimensional comparative analysis of the prediction performance of various mainstream single-model energy consumption prediction methods was conducted,thereby laying the foundation for the subsequent development of fusion models.(3)A ship energy consumption prediction method based on multi-model fusion is proposed.The fusion of multiple high-performance single models is carried out,and a ship energy consumption prediction fusion model based on the Stacking framework is constructed.The analysis results indicate that the proposed ship energy consumption prediction fusion model performs better than the individual models in various performance indicators.(4)Verification and analysis of ship energy consumption prediction method based on multi-model fusion.In view of the problems of low model selection efficiency and difficult optimization of the best model parameter combination in model fusion,a ship energy consumption prediction method combining intelligent model selection strategy and adaptive optimization of model parameters is proposed,and the effectiveness of this method is verified through examples.The analysis results show that this method can predict ship energy consumption more accurately.Compared to the improved Stacking fusion model before optimization,the root mean square error(RMSE)is reduced by 50.0% and the mean absolute error(MAE)is reduced by 9.9%.Compared to the best single ET prediction model,the RMSE is reduced by 66.7% and the MAE is reduced by 12.6%.Therefore,it provides a theoretical basis for ship energy efficiency optimization technology and has important practical value for achieving intelligent energy efficiency optimization and control of pollutant gas emissions in ships.