Research On The Energy Management Strategy Of Marine Gas-electric Hybrid Power System Based On Model Predictive Control

Low-carbon,green,and intelligent approaches are the primary directions for the development of ship power systems.As ship power system performance requirements continue to rise,a single power source becomes insufficient to meet the demand.Hybrid power systems provide a solution by leveraging the complementary advantages of multiple power sources and addressing their respective limitations.Among these,the marine gas-electric hybrid power system,which combines a natural gas engine with an electric motor,has garnered significant attention.This system maximizes the efficient and clean nature of natural gas engines while benefiting from the energy storage capability and rapid dynamic response of the electric power system.Compared to traditional marine power systems,the marine gas-electric hybrid power system offers additional mechanical and energy degrees of freedom due to its multi-power source coupling.The system’s utilization of diverse energy sources enables flexible,safe,and reliable operations.However,there are challenges such as the lack of systematic design in system matching,simplistic algorithms,and inadequate theoretical depth in energy management strategies,which hinder the technical advancement and widespread adoption of this system.To address these issues,comprehensive research is urgently needed in capacity parameter matching,energy management strategy enhancement,and real-time applications.The objective is to determine the optimal power system configuration and energy management strategy that enhance ship dynamics while minimizing costs and emissions.The primary focus of this paper is the investigation of the energy distribution strategy and real-time application of the marine gas-electric hybrid power system.To achieve this goal,the paper aims to develop a dedicated test platform and a semi-physical simulation platform.Additionally,experimental verification of the energy management strategy will be conducted.The main research contents of this paper encompass the following aspects.(1)A novel methodology is proposed to address the lack of a systematic design for matching capacity parameters in new marine gas-electric hybrid power systems.This method considers the dynamic characteristics of the natural gas engine(NGE),as well as the low load and low efficiency zones,in order to determine the optimal mixing degree for the natural gas engine and electric motor.By contrast,traditional matching methods fail to account for these factors.The matching results of the capacity parameters are used to establish a test platform for the marine gas-electric hybrid power system.Test data from the platform is then utilized,along with a basic principle model,to develop a dynamic simulation model for the marine gas-electric hybrid power system.The objective of this model is to validate its effectiveness and provide a simulation platform for subsequent verification of the energy management strategy.(2)To address the challenge of hybrid characteristics exhibited by power sources in marine gas-electric hybrid power systems during their simultaneous operation,a hybrid dynamic prediction model is developed based on the principles of hybrid logic dynamic theory.This model,along with the optimization of a performance index function,is utilized to propose an optimal energy allocation strategy that aims to strike a balance between power,economy,and emission performance.The hybrid dynamic prediction model,along with the dynamic simulation model,serves as the controlled objects for performance testing,validating the feasibility and effectiveness of the energy optimal allocation strategy.Notably,simulation results indicate a significant 47.47% reduction in performance when employing the energy optimal allocation strategy in terms of dynamics.(3)For the problem of poor power performance improvement of the optimal energy allocation strategy,a dynamic coordination control strategy based on nonlinear model predictive control is proposed.Firstly,a prediction model for the dynamic coordination process of the NGE is derived.Secondly,in order to overcome the limitations of fixed weights,a variable weight fuzzy regulator based on fuzzy theory is introduced.Lastly,leveraging the principles of nonlinear model predictive control,a dynamic coordination control strategy is developed,incorporating the variable weight fuzzy regulator.Additionally,a cascade control strategy is formed by sequentially integrating the optimal energy distribution strategy.Experimental results demonstrate that the cascade control strategy can enhance power performance by 27.37% while considering the economic and emission performance of the energy optimal distribution strategy.(4)To enhance the real-time performance of the cascade control strategy,this study proposes a mixed-model predictive explicit segmented linear controller with an embedded optimization method.Addressing the computation time issue associated with the energy optimal allocation strategy’s solution of the mixed integer quadratic programming,an upgraded explicit output-feedback-tracking controller is presented based on the theory of segmented linear controllers.Simultaneously,the real-time code generation challenge in the dynamic coordination control strategy is resolved through the utilization of an embedded optimization method.By optimizing the cascaded control strategy using the mixed-model predictive explicit segmented linear controller and embedded optimization method,a real-time implementation strategy suitable for real-time applications is obtained.The hardware-in-the-loop(HIL)platform is utilized to test the dynamic simulation model as the controlled object.Simulation results indicate that the real-time implementation strategy exhibits smaller variations in equivalent gas consumption and NOx+HC emissions compared to the cascade control strategy,albeit with a reduction of 2.68% in dynamics.Considering the real-time factor,the real-time implementation strategy proves effective.(5)In response to the problem that the real-time implementation strategy does not consider the operational stability conditions of the test platform of the ship’s gas-electric hybrid power system,a real-world implementation strategy based on online estimation of NGE torque is proposed.Firstly,considering the unavailability of NGE torque feedback on the test platform,a torque online estimation method based on the sliding mode control algorithm is proposed and verified using the HIL platform.Secondly,to better simulate actual ship operation conditions,machine learning technology is employed to conduct hierarchical cluster analysis and time series averaging on the authentic ship operation data,resulting in the identification of seven distinct loading cycles for experimental testing.Moreover,to comprehend the interplay between weights in the machine implementation strategy,a comprehensive comparison of power,economy,and emission trade-offs is conducted based on the test platform’s evaluation.