PHYSICS-INFORMED NEURAL NETWORK BASED FLOW RIPPLE PREDICTION IN AXIAL PISTON PUMPS FOR FAULT DIAGNOSIS

Norboyev A. T.


Annotatsiya

This paper presents a Physics-Informed Neural Network (PINN) based approach for predicting flow ripple in axial piston pumps for fault diagnosis. Axial piston pumps are widely used in hydraulic systems where their performance reliability is critical for industrial applications. One of the key indicators of pump health is flow ripple, which reflects fluctuations in the output flow rate and is strongly related to internal mechanical conditions and potential faults. However, direct measurement of flow ripple in real industrial environments is challenging, while traditional data-driven methods often require large labeled datasets and lack physical interpretability. To overcome these limitations, a PINN framework is developed by integrating measured pressure ripple data with governing physical laws of hydraulic pipeline dynamics. This hybrid approach enables the reconstruction of unmeasured flow ripple and transforms the problem into a physics-constrained inverse problem. Both numerical simulations and experimental validations are conducted to evaluate the performance of the proposed model. The results demonstrate high prediction accuracy, with relative errors within 0.3%, confirming the robustness of the method. Furthermore, the study shows that the proposed approach can effectively capture fault-related characteristics of pump operation. The framework also proves capable of handling high-frequency inverse problems when extended architectures are applied. Overall, the proposed method provides a reliable and interpretable tool for intelligent fault diagnosis in hydraulic systems.


Adabiyotlar

  1. Plaut, R.H., Virgin, L.N., “Vibration and snap-through of elastica strips,” ASME Journal of Applied Mechanics, 2009.
  2. Wang, C.M., Lam, K.Y., He, X.Q., “Instability of variable arc-length elastica,” Mechanics Research Communications, 1998.
  3. Zhang, X.W., Yang, J.L., “Inverse problem of elastica beam,” Acta Mechanica Sinica, 2005.
  4. Wang, Q., Zou, H.L., Deng, Z.C., “Snap-through of elastica under displacement control,” Mechanics Research Communications, 2020.
  5. Pulngern, T., Sudsanguan, T., “Elastica of curved beams,” International Journal of Non-Linear Mechanics, 2013.
  6. Chucheepsakul, S., Wang, C.M., “Large deflection of beams,” ASME Journal of Engineering Mechanics, 1994.
  7. Lu, N.Y., Lu, C.J., “Stability of shallow arch systems,” International Journal of Non-Linear Mechanics, 2014.
  8. Plaut, R.H., Virgin, L.N., “Elastica vibration and instability,” ASME Journal of Applied Mechanics, 2009.
  9. Chen, J.S., Ro, W.C., “Elastica deformation and stability,” ASME Journal of Applied Mechanics, 2010.
  10. Liakou, A., Detournay, E., “Constrained buckling of elastica,” International Journal of Non-Linear Mechanics, 2018.
  11. Domokos, G., Holmes, P., Royce, B., “Constrained Euler buckling,” Journal of Nonlinear Science, 1997.
  12. Harne, R.L., Wang, K.W., “Vibration energy harvesting systems,” Smart Materials and Structures, 2013.
  13. Bowen, C.R., Kim, H.A., Weaver, P.M., “Piezoelectric energy systems,” Energy & Environmental Science, 2014.
  14. Yang, D., Mosadegh, B., “Soft robotics and elastomeric beams,” Advanced Materials, 2015.
  15. Chen, T., Mueller, J., Shea, K., “Multi-material structural design,” Scientific Reports, 2017.