Migration and Deposition Characteristics of Sand Particles on Mesh Filter Surfaces in Water-Saving Irrigation Systems: Evidence from Simulation and Experimental Verification

To explore the characterization of the transport and deposition of sludge particles in the mesh filter of water-saving irrigation systems, taking the inlet flow, particle concentration, and calculation time as experimental factors, a filtration test was carried out by computational fluid dynamics-discrete element method model (CFD-DEM) coupling and physical testing. Through the test results, deposition distribution, average particle size, and geometric standard deviation of the filter cake are analyzed. The cake pressure drop model of the filter cake’s average particle size and the filter cake’s geometric standard deviation are established. The results showed that the deposition of filter particles is triangular during the filtration process, and the axial deposition particles of the filter screen gradually increase. The proportion of small particles (smaller than 0.6 mm) in the cake layer formed by particles is the largest, greater than 70%; the main deposition area of particles is the lower filter screen (90%), and the particles deposited on the upper filter screen are all small particles. The increase in flow significantly affects the distribution of different sediment particles in the inlet and middle sections of the filter cake layer; the average particle size and geometric standard deviation of the filter cake increase with the increase of flow. An analysis of pressure drop was also conducted. The pressure drop of the filter cake decreases with the increase in geometric standard deviation and average particle size. Compared with the test pressure drop, the calculated pressure drop is consistent. The obtained results have important theoretical and practical significance for establishing a screen filter screen and cake pressure drop model, as well as optimizing filter media filtration efficiency and cleaning efficiency, and the relative error does not exceed 10%.