Abstract An in-house compressible three-dimensional (3D) finite volume Computational Fluid Dynamics (CFD) method, with statistical turbulence model and moving grid capabilities, is presented and applied to the suction stroke of a simplex plunger positive displacement pump. The approach utilizes a pressure-based implicit solution algorithm and a mass transfer cavitation model. Fluid-actuated valve dynamics are approximated using a fluid-structure interaction algorithm. The closed valve and early phase of valve opening are approximated by a permeable wall. A circular segment model is introduced, significantly reducing computing time. Experimental validation is performed by time-resolved pressure and flow rate measurements, as well as high-speed visualizations of the valve dynamics. A speed variation is conducted to investigate harmless advanced and erosive distinctive partial cavitation. The simulation reproduces the delay of cavity collapse, observed with increasing speed, and reveals distinctive void patterns related to the chamber pressure time progression. These void patterns are fundamental for understanding the cavitation dynamics and potential erosion risk in the system. Deviations from data remain in the flow phase subsequent to the collapse due to overestimated wave reflection at the inlet boundary in the suction pipe.