1 April 2005
Ultrasonic spots leaky valves
There are significant savings involved in identification and monitoring of critical valves.
By Hans Wagner
It is vital to recognize challenges in the oil and gas industry regarding leak identification, quantification, and monitoring of critical areas such as valves, flanges, bends, or joints.
Today, with all the refineries, process plants, and production facilities, it is essential to take action against unwanted leaks and environmental complications.
In some countries, the contents oil and gas companies are subject to strict regulations from their governments. Governmental regulations do not only require the oil and gas companies to report the specific leaks, but also to report quantities of the leak.
Even though the regulations are there, enforcement does not necessarily happen. By using ultrasonic systems, the operators are able to not only comply with governmental regulations, but also keep the level of security very high.
Research shows significant savings are involved in identification and monitoring of critical valves. We'll look at the possibilities with ultrasonic systems for both topside and subsea monitoring of leaks. Further testing is required for subsea monitoring, and projects are in the pipeline.
Losses through passing valves are a general problem for operating sites, and significant cost savings are possible by being able to identify passing valves non-intrusively and thus enable operators to take early action to reduce the loss by repairing the faulty valve.
There are several areas where benefits arise from valve leak detection: reduced losses to flare, reduced losses to atmosphere, reduced losses from process, and maintenance planning.
The monitoring system has these components:
- Ultrasonic intelligent sensor
- Intrinsically safe power supply unit (with power barrier, signal barrier, RS485/RS232 converter)
- Laptop computer
- Leak monitoring software (graphical display of leaks, generates reports)
- Two-pair cable from sensor to PSU—one pair for power, one pair for signal
- Metal clamping bands
- Silicon compound (for best possible contact between sensor tip and pipe/valve)
If all the leaks are noisy
The physical principal used to detect passing valves is well known and is acoustic emission.
All valves passing will emit a frequency range with varying amplitude depending on the volume of the leak. An acoustic transducer can detect this noise or vibration. The physical source of leak noise is the varying pressure field associated with turbulence in the fluid. Turbulence results from flow instability where inertial effects easily dominate viscous drag. For flow in a cylindrical path, the Reynolds number (R) reflects turbulence.
R = rnD / m
r = Fluid density
n = Velocity of the leaking fluid
D = Leak diameter
µ = Fluid viscosity
Turbulence starts when the Reynolds number is between 1000 and 10,000, giving a lower limit for leak detection under ideal conditions.
In practice, the conditions are never ideal, so there are a number of other considerations that come into play. The most important ones are:
- The leak path is not cylindrical but is usual complex for small leaks, resulting in much higher turbulence, hence more noise from real leaks than predicted.
- The product loss may not come from one single leak but from multiple leaks distributed around a valve seat. If these are below the critical point for turbulent flow, then no noise emits.
- If all the leaks are noisy, then the signals will add up to something that is not the same result as one single leak path.
- The noise level at the source is not what the sensor has to detect; the signal has to travel through the valve body to where the sensor mounts on the outside of the tank, and this distance varies for different valve sizes, as does the signal attenuation.
- The final factor is background noise from nearby turbulent flow or from normal plant noise and vibration; this is why measurement position is important.
There are a number of mechanisms that contribute to the 'noise' or vibration signals associated with leakage across valves including turbulent mixing, shock associated noise, cavitations—liquids only, mechanical source—parts of valves that resonate.
A number of reasons can cause the different mechanisms; the most common are seat peening damage, deposit build up between the nozzle and the seat, or simply scoring of the valve plug.
Installation and real time
The best position is as close as possible to the source of signal generation where there is good contact for signal transmission between valve and transducer. One must maximize the contact area between the measured component measured and the transducer.
Where possible, the measurement position should be a smooth, flat, or surface with a large radius. Preparation of the surface at the measured position is not normally necessary, but any loose paint or rust deposit must go so the transducer couples directly to base metal or firmly adhered surface. There can be no insulation or air gaps between the measure point and the signal source.
Other potential signal sources/background noise, which rarely present a problem, but need some consideration when measuring valves thought to be passing include:
- Noisy steam flow, where there are steam tracing lines to/around the valve body
- Nearby atmospheric steam leaks, impinging directly on the transducer or valve body
- Compressed air leaks or bleeds
- Nearby regulating control valves
- Elbows, tees, or the like causing noisy turbulent process flow nearby
To be able to use the gathered information from the sensor as well as possible, the information needs to be of recent quality; the monitoring must happen in real-time. The sensor will monitor every second, minute, day, week, etc. It is of critical importance that the appropriate personnel receive the alarm from a leak.
The signal from the sensor should also be a digital signal, due to the advantages of digital signals over analogue signals. The sensors can give an output of RS485, RS232, Modbus, Profibus, CANopen, Relay option, and 4-20mA. By using the digital option, the operator can have a two-way communication with the sensor, i.e. download new settings or new developments in the sensor software.
Predicative leak equation
When considering the range of areas that can benefit from leak identification and quantification, there are many cost savings involved.
Identification and quantification of leaking valves on process plant can lead to significant savings, reduced losses to the environment, and enhance maintenance and operations. Specific areas and instances where benefits can arise from valve leak identification and quantification include:
- Reduced losses to flare through leaking relief valves (and their bypasses), control, and blowdown valves. Even where a flare gas recovery system is in operation significant savings realize as high value gas, like propane for instance, does not have to downgrade to something less valuable like fuel gas.
- Reduced losses to atmosphere where valves discharge through vent pipes to atmosphere.
- For maintenance planning, the identification of leaking valves prior to shutdown.
- Troubleshooting—tracing sources of loss and cross contamination, identifying losses through pump and compressor recirculation valves, and monitoring deteriorations in plant performance.
The non-intrusive acoustic leak monitor offers a means to quickly and reliably identify and monitor leakage through valves and to quantify the leakage. In order to identify and quantify if a valve is leaking, the valve must be in closed position, and there must be a differential pressure of at least one bar across the valve. The leakage through the valve produces
an acoustic emission signal, which is detectable using a non-intrusive leak monitor. The leak rate is then determined after the specific parameters are input into the predicative equation.
Saudi Aramco, an international petroleum company, has achieved savings of approximately $1.5 million by identifying leaking valves before shutdown and through better maintenance planning by using non-intrusive leak monitors. This sends a clear message that savings are obtainable using this type of equipment actively in maintenance planning and NDT planning.
Another example is PGS, which made a decision on if they should continue producing with one particular valve, or replace it. They used a non-intrusive leak monitor not only to identify the leak, but also to quantify it. The leak through the valve was not substantial, and they decided to keep on using the valve. They found it more economical and within the safety parameters to leave the valve operating. CE
Behind the byline
Hans Wagner is president of ClampOn Inc. The article comes from his ISA EXPO 2004 paper and presentation Innovative techniques to deal with leaking valves.
Ultrasonics: Wave of future
Laser-based ultrasound is a promising technique for remote inspection of demanding in-process applications involving high workpiece translation velocities and high temperatures.
Thanks to their simple design and ability to compensate for wave-front distortions due to dynamic speckle changes, atmospheric turbulence, and vibrations in the factory environment, the new adaptive photo detectors are particularly promising for these applications.
In the next few years, we expect to see more in-process laser ultrasonic systems reach the stage of factory demonstration and installation.
Laser ultrasonics is a noncontact, nondestructive inspection and diagnostic tool with the potential for in situ sensing and process control across a variety of industrial and aerospace manufacturing needs.
Manufacturers that use laser ultrasonics produce parts with closer tolerances, labor and material savings, and higher production yields for materials as diverse as composites, steel, aluminum, semiconductors, and paper.
Laser ultrasonics can determine internal properties (thickness, temperature, defects) and monitor surface processes (thin-film deposition, case hardening, shock peening).
The basic laser ultrasonic technique involves two lasers that, in essence, replace ultrasonic contact transducers.
One laser, typically a high-peak-power pulsed source, generates a pulse of ultrasound in a material upon absorption of the light, while a second continuous wave laser interferometer remotely senses the ultrasonically induced surface displacements on the work piece.
This receiver approach circumvents the need for direct contact, close proximity, or squirter systems. Laser ultrasonics has a number of features that make it very attractive for process control applications:
The key enabling technology for many of the new in-process applications of laser ultrasonics is the class of adaptive interferometers developed to detect the small surface displacements encountered in typical measurement conditions.
Unlike conventional homodyne or heterodyne interferometers, these adaptive laser ultrasonic receivers efficiently process the speckled beams received from rough surfaces and/or multimode fibers.
In addition, they compensate for dynamic wave-front changes resulting from beam scanning, workpiece motion, or atmospheric turbulence.
Receivers for inspection
While laser generation of ultrasound can be more efficient than generation with contact transducers, optical receivers for ultrasound are generally less sensitive than contact transducers for detection of ultrasonic signals.
In recent years, there has been considerable interest in improving the performance of laser ultrasonic receivers. The specific requirements of such receivers are as follows:
Engineers have used laser interferometers for years to detect the small-amplitude surface displacements produced when an ultrasonic wave reaches the detected surface.
Originally, passive homodyne or heterodyne interferometers with coherent detection could not operate effectively with the speckled input beams that result from interrogating a rough surface with a laser probe beam.
Behind the byline
Marvin Klein, Bruno Pouet, Sebastien Breugnot, and Konrad Peithmann contributed this brief. They work for Lasson Technologies, Inc. in Culver City, Calif. Read their entire paper at www.isa.org/isaolp/journals/pdf/intech/20011038.pdf.
PSV leak recorded with clampon leak monitor software
This graphic is a pressure safety valve, which opens up at 5.5 bar. It settles down and opens up when pressure has reached 5.5 bar again. Between 10:10 and 10:11, the valve malfunctions and does not close 100%. The horizontal line indicates a leak of about 1.6 liters per minute (l/min).
From research done on 491 valves, a multiple regression analysis found a relationship that enables one to predict the leak rate by using the measured signal and knowledge of the valve size, valve type, and operating inlet pressure.