Master Cylinder Pressure Estimation Of The Integrated Electro-hydraulic Brake System

Brake-by-wire system(BBW)is the key component required for electric and intelligent vehicles.As a branch of BBW,the integrated electro-hydraulic brake system(IEHB)dominates the research and development of BBW.Master cylinder pressure control is one of the key technologies of IEHB which requires sensors to obtain the master cylinder pressure information.Once the sensor fails,IEHB will not be able to perform normal functions,which seriously threatens the driving safety.Therefore,in order to improve the reliability of IEHB,it is urgent to break through the master cylinder pressure estimation technology.The existing literature lacks comprehensive testing and in-depth research on several important model characteristics used for master cylinder pressure estimation: friction coefficient of brakes,pressure-position character of the hydraulic brake circuit and friction character in IEHB,resulting in low accuracy and poor robustness of the existing estimation algorithms.In this paper,facing the key technical demands of intelligent and electric vehicles,this research focuses on the master cylinder pressure estimation of IEHB.Firstly,the research status of IEHB products and the master cylinder pressure control algorithms at home and abroad is briefly introduced,and the significance of master cylinder pressure estimation is pointed out.After that,the structure and working principle of the IEHB studied in this paper are introduced,and then three methods for estimating the master cylinder pressure are introduced,which are: based on vehicle dynamics,based on the pressure-position relationship of the hydraulic brake circuit,and based on IEHB’s dynamic equation.According to the above three methods,the existing literatures are reviewed in two aspects: related basic models and master cylinder pressure estimation methods,respectively,so as to summarize the shortcomings of the existing research and lead to the content of this paper.For the master cylinder pressure estimation based on vehicle dynamics,the algebraic formula of the friction coefficient of the brakes is first derived based on vehicle longitudinal dynamics.After that,through real vehicle tests under different working conditions,the evolution of the friction coefficient in the braking process under different initial temperatures of the brakes,brake pressures and initial vehicle speeds is studied,and a revised model of the friction coefficient is proposed based on the test results.By adopting this model,the algebraic formula of the master cylinder pressure estimation is derived based on the five-degree-of-freedom vehicle model.Real vehicle tests show that the proposed algorithm for estimating the master cylinder pressure is more accurate than that based on constant friction coefficient and vehicle longitudinal dynamics and can adapt to slopes and steering conditions.For the master cylinder pressure estimation based on the pressure-position relationship,a theoretical pressure-position model is deduced by letting the equivalent bulk elastic modulus represented by the deformation of the hydraulic brake circuit be equal to that of the brake fluid itself.Then,parameter identification and model verification are conducted based on test data.Since the theoretical pressure-position model is not feasible to be directly used for pressure estimation,the pressure-position relationship is tested under multiple conditions,and it’s found that the hysteresis pressure has a certain positive correlation with the rack speed,and a dynamic pressureposition model is proposed based on the test results.By comparison with the commonly used static pressure-position model,the estimated pressure based on the dynamic pressure-position model is more accurate with faster response speed.Further,after analyzing the advantages and disadvantages of the master cylinder pressure estimation algorithm based on vehicle dynamics and pressure-position model,a fusion-based master cylinder pressure estimation algorithm through the recursive least square(RLS)is proposed.First,the vehicle dynamics is adopted to pre-estimate the master cylinder pressure,and then RLS is adopted to update the parameters of the pressure-position model based on the pre-estimated pressure,rack displacement and rack speed.Finally,the updated pressure-position model is used to estimate the master cylinder pressure.Vehicle tests and simulations show that the fusion-based algorithm combines the advantages of single-model-based algorithm,and its stability,accuracy,adaptability to operating conditions,and robustness to brake pad wear are all improved with only small delay.However,this algorithm lacks robustness to the inaccuracy estimation of vehicle parameters.Aiming at the shortcomings of the above algorithm,a master cylinder pressure estimation algorithm based on fusion of IEHB’s dynamics and pressure-position model is proposed.Firstly,the dynamic equation of IEHB is established based on Newton’s law,and then the sliding friction force in the reduction gears of IEHB is theoretically analyzed and tested under multiple conditions.Test results not only show that the sliding friction force is approximately a linear function of the motor torque,but also show that the sliding friction force remains appropriately stable under common operating conditions.Based on the test data,a robust adaptive Karnopp friction model is established,and the pressure is pre-estimated by the friction model and IEHB’s dynamic equation to ensure the robustness of the fusion-based algorithm.Vehicle tests and simulations show that the master cylinder pressure estimation algorithm based on the fusion of IEHB’s dynamics and pressure-position model has comparable accuracy,stronger robustness and portability than that based on vehicle dynamics and pressureposition model.The master cylinder pressure estimation algorithm is integrated into the master cylinder pressure control algorithm.Real vehicle tests show that under step conditions of different amplitudes,the steady-state error of the estimation-based control is mostly less than 5%.;Under normal driving conditions,the accuracy of the estimation-based control is slightly lower than the sensor-based control,but the root mean square error(RMSE)is less than 1 bar,which proves the engineering practicality of the proposed pressure estimation algorithm.