Time Delay Study of Ultrasonic Gas Flowmeters Based on VMD-Hilbert Spectrum and Cross-Correlation

Lingcai Kong, Liang Zhang, Hulin Guo, Ning Zhao, Xinhu Xu2024 · DOI

TL;DR

This paper introduces a novel signal processing method for ultrasonic gas flowmeters, combining Variational Mode Decomposition (VMD) and Hilbert Spectrum analysis with Cross-Correlation (CC). This approach significantly improves the accuracy and reliability of time delay measurements, crucial for precise flow rate determination, by effectively filtering noise and enhancing echo signal quality.

What Problem Does This Paper Address?

Ultrasonic gas flowmeters are widely used due to their advantages, but their accuracy heavily relies on precise time delay measurement of the ultrasonic echo signals. Existing methods for estimating this time delay, such as thresholding or feature point methods, can be susceptible to noise and computational complexity. Furthermore, ultrasonic gas flowmeters are particularly prone to signal attenuation and fluctuations due to the nature of gas as a medium, making robust noise reduction and signal enhancement a critical challenge. This paper aims to address these issues by proposing a more accurate and reliable method for estimating the time delay, thereby improving the overall performance of ultrasonic gas flowmeters.

Methods

The proposed method involves a multi-stage signal processing pipeline:

  • Initial Pre-filtering: A Butterworth band-pass filter is first applied to the raw ultrasonic echo signals to remove significant interferences and background noise.
  • Signal Decomposition and Reconstruction:
    • Empirical Mode Decomposition (EMD): Used as a baseline, EMD decomposes the signal into Intrinsic Mode Functions (IMFs). However, the paper notes its limitations, including endpoint effects and modal aliasing, which can lead to incomplete noise removal.
  • Variational Mode Decomposition (VMD): This is the core novel technique. VMD decomposes the signal into a set of IMFs, each with distinct center frequencies. The paper highlights VMD's superior ability to handle modal mixing and endpoint effects compared to EMD. The Hilbert spectrum is then used to analyze the time-frequency characteristics of these IMFs. The most suitable IMF (Component 4 in this study) is selected for signal reconstruction.
  • Time Delay Estimation: The reconstructed, cleaned echo signals (forward and reverse) are then processed using the Cross-Correlation (CC) algorithm. The time delay ($\Delta t$) is determined by finding the peak of the cross-correlation function between these two signals.

Datasets

The study utilizes experimental data collected from a self-developed ultrasonic gas flowmeter prototype tested on a laboratory-built gas flow standard device. The experiments involved collecting echo signals at various flow rates using air as the medium. The paper does not mention specific publicly available datasets but rather describes the experimental setup and data acquisition process.

Key Insights and Ideas

  • VMD's Superiority over EMD: The paper strongly advocates for VMD over EMD for ultrasonic echo signal processing. VMD's ability to decompose signals into IMFs with distinct frequency bands and its better anti-interference capabilities lead to a cleaner signal with less residual noise and baseline drift.
  • Hilbert Spectrum for IMF Analysis: The use of the Hilbert spectrum provides a valuable time-frequency perspective, allowing for a more informed selection of the most relevant IMF for signal reconstruction.
  • Combined Approach for Accuracy: The synergistic combination of VMD for noise reduction and signal enhancement, followed by Cross-Correlation for precise time delay estimation, is shown to be highly effective.
  • Impact on Measurement Error: The proposed method significantly reduces measurement errors, particularly at lower and higher flow rates, bringing them within stringent calibration requirements.

Key Results

The proposed VMD-Hilbert Spectrum and Cross-Correlation method demonstrated significant improvements in ultrasonic gas flowmeter performance:

  • Reduced Measurement Error: The maximum indication error was reduced to 0.84% within the tested flow range of 60–606 m³/h. This is a substantial improvement compared to results without VMD-Hilbert processing, where errors could reach up to 8.56% at low flow rates and 3.51% at high flow rates.
  • Improved Repeatability: Repeatability was consistently below 0.29% across the tested flow range.
  • Compliance with Standards: The achieved accuracy meets the 1-level accuracy requirements as outlined in the national ultrasonic flowmeters calibration regulation JJG1030-2007.

Limitations and Future Work

  • Flow Rate Limitations: The paper acknowledges that at flow rates exceeding 606 m³/h, signal quality and amplitude decrease. This phenomenon is attributed to potential vibrations from the PVC pipe or increased turbulence intensity.
  • Future Work: The authors suggest further testing and research to address the limitations observed at higher flow rates and to explore the method's applicability in more challenging environments. Further investigation into the impact of pipe materials and flow conditions on signal quality is also implied.

Summary updated

BibTeX
@misc{kong2024time,
  author = {Lingcai Kong and Liang Zhang and Hulin Guo and Ning Zhao and Xinhu Xu},
  title = {{Time Delay Study of Ultrasonic Gas Flowmeters Based on VMD-Hilbert Spectrum and Cross-Correlation}},
  year = {2024},
  doi = {10.3390/s24051462},
}
Figure