Research On Combustion Models And Control Strategy For Micro-pilot Ignited Dual-fuel Engine Based On Model-aided Development

Micro pilot diesel induced ignite natural gas dual-fuel engine is one of the important technical means to simultaneously reduce emissions and improve efficiency.Though its complex control dimension and strategy hinder the exploration of the control system and the improvement of technology.Model Assisted Control Development Technology is a new approach,which used the control object model to meet the control system development requirements,it can reduce development time and cost while increasing system reliability.Combining the development requirements of the micro-pilot ignition dual-fuel engine control system and the character analysis of the model algorithm,the quasi-dimensional model algorithm suits MACD technology most.However,the current researches on the dual-fuel engine model focus on the off-line quasi-dimensional algorithm,which cannot achieve a real-time calculation of the model.Moreover,the description of the coupling combustion relationship between the two fuels is relatively simplified.Therefore,a detailed off-line dual-fuel quasi-dimension combustion model that describes diesel and natural gas coupled combustion must be modeled.As well as the real-time on-line quasi-dimensional model must be modeled.They are the foundation of MACD realization.The paper aims to build a MACD technology platform to assist micro-pilot ignition dual-fuel engine control system development,the off-line quasi-dimensional,and the real-time on-line quasi-dimensional model has been built respectively.Through the engine test platform and real-time computing platform,the effectiveness,predictability,and real-time performance of both two models are proved.Based on the accuracy of the model,the macroscopic and microscopic numerical characteristics of the off-line and on-line models are compared,and the differences between the two model algorithms are analyzed fully.Finally,based on the MACD technology platform,the combustion characteristic optimization control strategy for the micro-pilot ignition dual-fuel engine has been developed.The research could provide a solid foundation for micro-pilot ignition dual-fuel engine performance improvement and control system development,provide the platform for the iterative upgrading of subsequent control strategies.Concrete research works organized as follow:In order to increase the calculation accuracy of the off-line quasi-dimensional model,discrete droplet and complex surface flame propagation processes are used to describe the combustion process of diesel and natural gas.The quasi-dimensional combustion process has been optimized into an on-line quasi-dimensional model.The multi-cylinder and multi-region parallel algorithm which adopt array input and the single-step solution has been proposed,so that the mathematical model of each calculation area can be established according to the unified formula and the multi-region running method throughout the whole cycle.By using the diesel fuel interregional transfer model and the natural gas dynamic flame core model,the calculation area and complex equation of the off-line quasi-dimensional model increase with time are optimized.As a result,the complexity of the model and the amount of compiled code has been reduced,and the computing speed is improved.To verifies the validity and predictability of the model,in this paper,A quadrature orthogonal test design method is used.This method divides orthogonal test data into a control group and three validation groups,by computing the validation group test data and calculating the data's error to judge the effectiveness and predictability of the model.Simultaneously,the real-time computing platform,main frequency 3.6 GHz,verifies the on-line model's real-time performance.It obtains the applicable scale of the on-line model and then gives a setting step.The results show that the error,the test data compare to the off-line and on-line model,within6% and 8% respectively,both the models are effective and predictive.Array parallel algorithm and multi-region through-run calculation methods are suitable for multi-cylinder multi-region quasi-dimensional engine model,which can not only save up to 30%computation time but also stabilize the model computation time.At the same time,it proves that the on-line model meets the real-time simulation requirements of the 6K dual-fuel,and it recommends to set the simulate step to 0.05 ms.Further,the macroscopic and microscopic numerical features of models are compared and analyzed respectively.By introducing the fitting degree which compared the off-line and on-line model error distribution.It is proved that the two modeling methods on the macroscopic data reflect similar numerical characteristics and change rules.Then,by comparing the numerical characteristics of the diesel combustion process,natural gas combustion process,and in-cylinder state variable,it is proved that the two models show similar characteristics in microscopic,Although the numerical performance is different,which shows two model algorithms differences.At the same time,the on-line model dynamic numerical characteristics prove that the on-line model can achieve the same control operation characteristic as the dual-fuel engine test platform,and it presents the dynamic response by the engine operation law.The general method and technical route of applying MACD platform werepresented.Based on thisroute,the aim was realizing dual-fuel mode operation in alloperating conditions.Iterative researched the combustion optimization control strategy of the speed and excess air.Proved there were delay and hysteresis effects caused by the control algorithm and the intake and exhaust system in combustion optimization control strategy.By increasing the throttle valve control feedforward and changing the throttle PID calculated output to modified final output value,the delay time can be reduced by 65% and the stability control operation of speed and excess air coefficient can performance well.