Digital Twin-Driven Zero-Shot Fault Diagnosis of Axial Piston Pumps Using Fluid-Borne Noise Signals

TL;DR

This work introduces a digital twin–driven zero-shot fault-diagnosis framework for axial piston pumps that uses fluid-borne noise signals to train classifiers without fault data. By calibrating high-fidelity physics models with healthy-state data and generating synthetic fault signals, the approach enables robust fault diagnosis with real-world data, validated by accuracies exceeding 95%. A physics-informed neural network provides a virtual sensor for flow ripples, while Grad-CAM ensures the learned features align with physically meaningful fault signatures. The results demonstrate that calibrated simulations tightly match experimental signals and that Grad-CAM-guided architecture design enhances generalization in simulation-to-reality transfer.

Abstract

Axial piston pumps are crucial components in fluid power systems, where reliable fault diagnosis is essential for ensuring operational safety and efficiency. Traditional data-driven methods require extensive labeled fault data, which is often impractical to obtain, while model-based approaches suffer from parameter uncertainties. This paper proposes a digital twin (DT)-driven zero-shot fault diagnosis framework utilizing fluid-borne noise (FBN) signals. The framework calibrates a high-fidelity DT model using only healthy-state data, generates synthetic fault signals for training deep learning classifiers, and employs a physics-informed neural network (PINN) as a virtual sensor for flow ripple estimation. Gradient-weighted class activation mapping (Grad-CAM) is integrated to visualize the decision-making process of neural networks, revealing that large kernels matching the subsequence length in time-domain inputs and small kernels in time-frequency domain inputs enable higher diagnostic accuracy by focusing on physically meaningful features. Experimental validations demonstrate that training on signals from the calibrated DT model yields diagnostic accuracies exceeding 95\% on real-world benchmarks, while uncalibrated models result in significantly lower performance, highlighting the framework's effectiveness in data-scarce scenarios.

Figures (17)

• Figure 1: The proposed DT-driven zero-shot fault diagnosis framework for axial piston pumps using FBN signal
• Figure 2: Test Rig
• Figure 3: Numerical model and Fault injection scheme
• Figure 4: Pressure ripple comparison under health conditions: experimental data versus simulation results from (a) $\mathcal{M}_{\text{3D-CFD}}$(without ITA calibration) and (b) $\mathcal{M}_{\text{1D-MOC}}$(with ITA calibration).
• Figure 5: Pressure ripple comparison under faulty conditions: experimental data versus simulation results.