Acoustic emission testing is often associated with the detection of metal or composite flaws that are active, or grow, while stressed. In steel for example, fatigue crack growth will emit acoustic emission in the 100 to 500 kHz range. Similarly, acoustic emission bursts are detected in composite material from fiber breaks, matrix cracking and other failure mechanisms. The acoustic emission emitted from fatigue cracks in steel and fiber breaks in composites are discrete bursts of acoustic energy. The acoustic emission intensity and rate at which it is generated may be correlated to damage severity. In comparison, acoustic emission testing for valve leak detection analyses the continuous emission of acoustic emission and largely focuses on acoustic emission amplitude. The objectives of acoustic emission inspection of valves are to detect leaks and estimate leak rate. This introductory article summarizes some of the key components of acoustic emission leak detection in valves.
Figure 1: Valve acoustic emission generated by turbulent flow.
Acoustic Emission Sources in Valves
Acoustic emission is generated in by high pressure, high velocity, turbulent flow in a leaking valve. The magnitude of turbulence is related to the Reynold’s number which is dependent on the density, flow velocity flow path radius, and viscosity [1-3]. Useful acoustic emission is generated in the valve at Reynold’s numbers between 1,000 to 10,000. While unrelated to valve leakage, cavitation is also a source of acoustic emission in valves. Cavitation occurs when the local static pressure does not exceed the saturated vapor pressure. The expansion and collapse of cavitation bubbles leads inner-surface corrosion or pitting of metal surfaces. The corrosion and pitting compromise sealing surfaces and may be a pre-cursor to valve leaks and failure. Figure 1 show a simple setup for acoustic emission valve leak detection . The flow direction is shown on the upstream side of the valve. The valve leak paths along the outside diameter of the ball diaphragm. Turbulence is shown on the downstream side of the valve. Field research has identified the inner surface of the downstream surface and the optimal position for the acoustic emission sensor.
Measuring Valve Leak Acoustic Emission
Steel and composite material failure applications monitor acoustic emission peak amplitude, rise-time, energy and other discrete AE features. Leak detection acoustic emission data acquisition is fundamentally different in that it monitors average acoustic energy (AErms) and average signal level (ASL). AErms is defined as the rectified, time averaged AE signal, measured on a linear scale and reported in volts . The average signal level is defined as the rectified, time averaged AE logarithmic signal, measured on the AE amplitude logarithmic scale and reported in dBae units (where 0 dBae refers to 1 µV at the preamplifier input). Mathematically, AErms is expressed as :
The acoustic emission average signal level in volts (ASLv) is formulated using:
The acoustic emission ASL in dB (ASLdb) is formulated using
Valve Leak Acoustic Emission Analysis
In order to correlate acoustic emission data to leak detection and leak rate qualification, the relationship between the time averaged AErms and average signal level ASLdB acoustic emissions must be understood for specific valve types, valve sizes, and pressure differentials across the valves. To establish this relationship, acoustic emission testing is performed on a range of valve sizes across which a known pressure differential is applied. Acoustic emission data is recorded then analyzed to establish the connection between AE and leak rate. Leak rates are established in milli-liter per second (mL/sec) or comparable units. Figure 2 shows example acoustic emission RMS data. The vertical axis scales the AE RMS data micro-volts (uV). The data analysis has been setup to acquired data from multiple sensors on the valve. Acoustic emission data is color coded by sensor to determine which location is most sensitive to the valve leak. The horizontal axis simply plots time. AE RMS data is typically acquired at 1 milli-second to 1 second sampling intervals. A secondary vertical axis may be used to further correlate the acoustic emission data to pressure differential, flow velocity or any user defined variable.
Figure 3 shows the distribution of acoustic emission during the test duration. The vertical axis scales the acoustic emission average signal level in dB, time axis is in dB, and the left vertical axis records the user defined parametric, or secondary, input. In this scenario, as the pressure differential increases across the valve the amplitude of the acoustic emission increases.
Figure 2: Acoustic emission time averaged signal measurement.
Figure 3: Acoustic emission average signal level measurement.
The output from the analysis of the experimental acoustic emission is a simple equation that considers the following three variables: acoustic emission ASL or RMS, inlet pressure and valve size. Q is the leak rate in the desired units (liter/min, etc.). X, Y, and Z are constants derived from the experimental data.
log(Q) = X log (ASL or RMS) – Y log (Pinlet) + Z Log (Svalve)
1. R. K. Miller and P. McIntire, Nondestructive Testing Handbook: Volume 5 Acoustic Emission Testing, American Society for Nondestructive Testing, Columbus, Ohio, USA, 1987.
2. W. Kaewwaewnoi, A. Prateepasen, and P. Kaewtrakulpong, “Measurement of valve leakage rate using acoustic emission,” in Proceedings of the International Conference on Electrical Engineering/Electronics, Computer, Telecommunications, and Information Technology (ECTI ’05), pp. 597–600, 2005.
3. ASTM E 1316 – 08a Standard Terminology for Nondestructive Examinations.