Research On Energy Management And Intelligent Decision System For All-Electrical Ship

Recently,All-Electrical Ship(AES)which equipped Integrated Electrical Propulsion System(IEPS)as its hallmark has become the trend for future modern ship.AES has advantages of reducing fuel consumption and maintaining costwhile improving ship survivability and effectiveness.Energy Management System(EMS)is one of the critical control functions in AES,which improves the effectiveness coordinating with other control systems.EMS is critical for improving survivability and mission effectiveness while reducing manning cost.Only a few countries(US and UK)and multinational corporations has mastered the techonolgies of energy management and intelligent decision dedicated for AES and those technoligies are kept secret.Lack of information and successful cases as reference poses challgent for research on EMS dedicated for AES.In this dissertation,some technologies of Energy Management and Intelligent Decision System(EMIDS)for AES are researched.Firstly,EMIDS for AES design is lack of experience and unified standards,a design scheme of EMIDS for AES is studied basing on analyzing and concluding the functional requriments of EMIDS for AES.Taking the shipboard MVDC integrated electric propulsion system as the research plant,a hierarchical control architecture with multi-time scales for EMIDS is proposed to meet the requirements of real-time and autonomous performance in different control layer.A modular modeling method and simulation scheme for design of EMIDS is proposed to decouple the innership between models in complex system for improving the effectiveness of modeling and simulation.Secondly,the systems in AES has significantly different time constants which result in a stiff problem when simulating the all systems in a unified model,the simulation cost will be very expensive.Basing on analyzing the time scales of control strategies in EMIDS,a all-system model oriented for design EMIDS is constructed with neglecting the high frequency switching and fast changing electromagnetic characteristics.Four-quadrant propeller model and ship motion model are presented for simulating different conditions of propulsion system.All modules are integrated into a unified model which represents thewhole system with input-output relationship for EMIDS design.Comparing the proposed model with a detailed mechanism model simulation results under the same condition,the validity and efficiency of the proposed reduce-order model for simulation are proved.Thirdly,in order to ensure the stable operation of the shipboard MVDC IEP system,the DC bus voltage should be maintained in a tolerance range.According to the DC bus voltage deviation degree,a energy coordinated control strategy based on DC bus voltage threshold is proposed and the effects of the strategy is verified by simulation.Aiming at the demand of the parameters optimization for the control strategy requires several simulations,a parameter optimization program is proposed by combining a improved swarm intelligence algorithm and the simulation model.Simulation results show that the obtained optimal parameter set can achieve a better control effect than other parameter set.Fourthly,the problem of shipboard IETS MVDC distribution network reconfiguration is researched.Considering the distribution network of IETS is a DC grid without reactive power,a two-stage distribution network reconfiguration strategy is proposed.In the first stage,a path for supplying sufficient power to the de-energize load should be find fastly with a shortest path search algorithm based on BFS and real power flow distribution on this path are computed;In the second stage,if the path which gived by Stage I has power deficit,a decision which load on this path should be unloaded to maximize the power supply to unaffected loads while meeting the generation capacity constrains on this path should be made.In this stage,a novel Gray Code Cucook Search(GCS)algorithm is proposed to solve the discrete optimal problem.Some fault cases are tested,the simulation results verified the proposed two-stage distribution network reconfiguration strategy has the capability to deal with the issue of load restroration after faults.Finally,resource allocation issue in AES is researched.Considering the resource allocation decision presents a property of markov,a constrained Multi-Agent markov decision process model for resource allocation is constructed,and given agents operating independently,the Multi-Agent markov decision process can be decomposed into seval independent markov decision process,a linear programing method is used to compute the optimal strategy for the Multi-Agent markov decision process.Resource allocation simulations are carried out under cruise and battle and mission change conditions,the simulation results indicate that resource allocation following the optimal strategy providethe demanded resource for the mission vital systems to maintain the vital system work properly in most of simulation time.The achivements of this dissertion which can provide powerful simulation and analysis tool for the design and development of EMIDS for AES posses important meaning in theory and engineering practice.