Research On The Modeling Methods Of Power Engine For Marine Engine Room Simulator

Intelligent ships and low emission are the trends in the world's shipbuilding industry.As a training system to enhance the skill of crew,Marine Engine Room Simulator(MERS)is widely accepted and promoted.MERS,the digital representation of the ship's engine room,has good mechanism and control logic about thermal,mechanical,electrical,and gas system,and a friendly user interface(UI).The network MERS can be interconnected with the intelligent ship and provide simulation and verification platforms to accelerate the research and development progress of the intelligent ship.The power engine in ship is vital,who is the energy source for the propulsion system and the electric system.Research on modeling methods of engine is an important content of MERS.A model calculation framework for data sharing is designed.In the previous MERS,the calculation data from models was enclosed so that it's impossible to allow secondary mining or interconnection.Based on the mastery of both the marine engineering systemand computer programming technology,the interface between the system input and output(IO)and the model calculation is extracted realizing the physical separation and data interconnection.Based on computer unicast and multicast communication technology,the real-time calculation data sharing and multi-system connection are realized,and the interface for data secondary mining is provided.The data updated from the system IO is realized by combining the event triggering mechanism and the polling mechanism.Using the MVVM structure reduces the coupling between the system UI and the data.With the real-time loading mechanism the computer computing and memory resources are reduced.The modeling method of the marine two-stroke engine has been improved.There is usually no distinction between two-stroke engine and four-stroke engine modeling and a uniform component division method is adopted.In fact,during the process of piston's moving up and down,the scavenging air will be sucked into or exhausted from the space formed by the cylinder block and the lower part of the cylinder.In this process,the scavenging air is heated by the cylinder wall before entering the combustion chamber and the temperature rise,meanwhile,the cylinder wall is cooled.The overall performance of the engines considering this structure or not was researched on BOOST.It was found that the model's accuracy was higher when considering this structure.The parameters of the highest accuracy increase are the exhaust manifold temperature,the main engine power,and the scavenging manifold temperature,which are increased by 5.52%,4.36%,and 4.13%,respectively.A modeling method for improving the calculation speed of the two-stroke engine model is presented.The mean value(MV)model can meet the requirements of real-time calculation even in high-speed engines.The zero-dimensional(OD)model can almost meet the requirements of real-time calculation in low-speed engines,but the computing is very large so that it cannot be directly applied in MERS.By adjusting the calculation frequency of OD cylinder model and MV cylinder model,the calculation speed of the engine model is increased.By replacing the pressure curve during the gas exchange process with two straight lines the calculation speed of the OD model is further improved.This model was built and simulated in MATLAB.It is found that under steady-state conditions,the accuracy of the improved OD model is the same as that of the OD model.Under dynamic conditions,the accuracy of the improved OD model is related to the computational frequency.When the frequency takes 5,the maximal relative error of the pressure in the dynamic process is only 0.363%.The solution for model of nitric oxide emission from internal combustion engine(ICE)is improved.In the sketch of 12 combustion products,13 nonlinear equations can be obtained according to the equilibrium constant method to solve 13 unknowns.The commonly used Newton-Raphson iterative method is very sensitive to the initial value and the oil-gas equilibrium ratio resulting in easy diverging and low accuracy.By constructing the 2-norm function of the nonlinear equations,the iterative solution problem of the nonlinear equations is transformed into an optimization problem.The performances of particle swarm optimization,perturbed particle swarm optimization,genetic algorithm and trust region dogleg on this optimization problem are compared and analyzed.It is found that the perturbed particle swarm optimization and the trust region dogleg can achieve the accuracy of 10-10 or more,while others are around 10-1.At the same time,it is found that the trust region dogleg is insensitive to the initial value and has a very high convergence speed,and the calculation time can be as low as 10-2 seconds.An engine modeling method that can simultaneously describe the dual fuel engine's fuel mode,gas mode,and fuel EGR mode is presented.The calculated heat release rate model is only for gaseous fuels or liquid fuels.The Wiebe heat release rate model is an empirically based model that is not sensitive to fuel type,but the validated parameters are only suitable for a specified working condition.The linear relationships of the combustion duration and the combustion rate distribution coefficient varying with the amount of oil supplied or the amount of gas supplied is obtained after the calibration with Wiebe model at loads of 25%,50%,75%,85%,100%,and 110%in the fuel mode and the gas mode respectively.But there are no obvious relationships of the start of combustion,the combustion duration and the combustion rate distribution coefficient varying with the amount of oil supplied and EGR rate after the calibration with Wiebe model at loads of 25%,50%,75%,100%and EGR rate of 10%,20%,30%,40%.The engine model capable of describing both fuel mode,gas mode and fuel EGR mode can be realized by combining linear function and 100%load reference point adjustment at both fuel mode and gas mode,and 2D interpolation at fuel EGR mode.