Numerical investigation of performance and multiparameter prediction model of high-pressure fuel filters and cavitation at filtration orifices considering variable fluid properties

The filtration of particulate impurities seriously affects efficiency and life of engines. For the existing common rail direct injection (CRDI) system of engine, a new type of high-pressure fuel laser micro-hole filter was developed, which can effectively filter fuel impurities in the common rail fuel injection system, while ensuring that pressure drop loss of the filter is small. This filter has a large feature size span and a complex structure, and a comprehensive and systematic understanding of its internal flow helps to propose optimization considerations. This paper presents numerical research on high-pressure filtration systems with different parameters considering variable fluid properties, and a particle adhesion model is introduced in the solver to determine the motion trajectory of particles. The main results show that diameters of filter holes with the best filtration capacity increase as the working pressure increases. Under the same conditions, the maximum difference in filtration efficiency of different filters can reach 4.1 times. On the basis of simulation results, multivariate flow prediction models based on artificial neural networks and nonlinear regression methods are established. Average relative errors of the prediction models are <3 %. To study the cavitation characteristics of filter holes with various structural parameters, numerical simulation research considering compressibility and variable properties of fuel is conducted on the cavitation phenomenon in microchannels of micron sized filter holes. The results indicate that the diameter ratio and length-diameter ratio have a significant impact on cavitation of filter holes.