Real-time simulation of gap flow in the valve plate/cylinder block pair of axial piston pumps based on a POD-Kriging surrogate model

The efficiency and lifespan of axial piston pumps are significantly influenced by the valve plate/cylinder block pair due to its large contact area and harsh load-bearing conditions. Accurate gap flow simulation of the valve plate/cylinder block pair is crucial for the pump design and maintenance. However, high-fidelity simulation modelling of this friction pair is time-consuming and not suitable for rapid design cycles and real-time condition monitoring. To address this issue, this paper proposes a real-time gap flow simulation method using a proper orthogonal decomposition (POD)-Kriging surrogate model. First, a transient computational fluid dynamics (CFD) model of axial piston pumps is established and validated experimentally. Second, the validated CFD model generates field data of the friction pair under various operating conditions to train the POD-Kriging surrogate model. Finally, the trained surrogate model predicts the gap flow of the friction pair in real time under unseen operating conditions. The experimental results demonstrate that the surrogate model can capture the pressure and velocity fields in the friction pair accurately and rapidly. Compared to the traditional offline CFD simulation, the surrogate model achieves a high accuracy rate of more than 0.95 and reduces the simulation time from 46 min to 43 ms.