Locating Leaks in Pilot Plants and Laboratory Units and Equipment

Locating Leaks

While leak testing remains as much an art as a science, understanding the basic principles involved and the equipment available can frequently make the task simpler and faster. The result will be an improvement in both the productivity and accuracy of research efforts. Some tips that are of special interest to pilot plant and research equipment are discussed below. By no means is the list exhaustive.

Welded joints, even when properly tested hydrostatically, can have pinhole leaks and/or weld defects that can leak gas. These welds should be carefully checked if a problem is suspected. (Code mandated leak testing often allows passing even with leak remaining if they are small enough. And even small leaks can be a major problem in pilot plants and laboratory units.)

Flanged joints may be difficult to check with a soap solution. One way to do so is to cover the open area between the two flanges (where the gasket is held) with tape or a similar material, forming a reasonably gas tight seal. A pinhole is made in the tape and any leakage is directed to this hole, where it can easily be located.

Spring loaded relief devices, the most common type used in pilot plants, may leak at almost any pressure, and will frequently leak intermittently or continuously at 80% or more of set pressure. (the onset of leakage is a function of numerous factors and varies significantly.) The outlets to all relief valves (and rupture disks) should be opened and checked for leakage if difficulty is encountered in sealing a unit. An effective long term solution is to run all vents to an inert oil bubbler made of glass pipe; the presence of bubbles in such a device immediately indicates relief valve leakage.

To detect leakage on a miniature (1.5 cm and smaller) valve, the handle may have to be removed. The clearance between the handle and the packing is so small that the leaking gas may be dispersed too widely to form clearly visible bubbles.

Leakage may occur through a closed valve, particularly if it has a metal to metal seat or small trim. This can usually only be checked by opening a downstream fitting, although a downstream pressure gauge may show a gradual pressure rise if located close to the valve.

Any device that has a diaphragm (such as back pressure regulators, pulsation dampers, pumps and compressors) can develop pinhole leaks through the diaphragm; it should be carefully checked, particularly on the side of the diaphragm away from the process. In addition, the diffusion of a light gas such as hydrogen or helium across a diaphragm can be significant, particularly if the diaphragm is made of an elastomer. To minimize such diffusion, such devices can be loaded with the process gas or helium, to minimize the diffusion effect, or a heavier oil, more resistant to gas absorption, used.

Thermowells, sample lines and similar intrusions should also be closely examined for pinhole leaks. Since such devices are frequently concealed inside vessels and other components, they may be missed.

If all methods fail to find a leak in a piece of equipment, removing it and repressurizing it while it is immersed in a tank of water will almost always locate the leak.

Finally, unless a proper holdback device is used on the main pipe when retightening compression fittings or pipe threads, adjacent fittings may "spring" or loosen, resulting in new leaks because of the stresses placed on them.

Leak Detectors

Leak detectors can be divided into two broad categories. Leak location detectors assist in pinpointing the exact location of the leak to allow repair. Leak rate detectors allow measuring the actual leak rate to determine if the maximum acceptable leak rate has been exceeded. Typically both types are required.

Leak Location Detectors

The most common detector for locating leaks remains soap solution. These are, in their simplest form a mixture of soap and water which is manually applied to each potential leak point. Escaping gas causes bubbles to form which can be visually observed by the operator. Soap solutions are inexpensive, reliable and easy for anyone to use. Typically 90-98% of all leaks can be located by a trained operator. In order to be effective, the soap solution must have a low viscosity so that it will penetrate even small gaps. Commercially available soap solutions are frequently modified to provide better wetting, improved penetration and better adhesion than available in a homemade mixture. While this may increase the cost it still remains relatively inexpensive and it usually significantly more effective. In order to use soap solutions most efficiently it is important to follow the following guidelines.

Apply the solution generously. (This is one time where more is almost always better!) Only areas that are adequately wetted can be tested. Many fittings and seals make it difficult for the soap solution to contact all potential leak points before running off. Hence it is generally required to over spray the fitting to ensure adequate coverage.

Wait for several minutes before deciding no leak is present. Many fittings will develop small dispersed leaks, also known as "fuzz leaks", only after several minutes. These take time to develop a cloud or "fuzz" of bubbles which are visible. While small, the cumulative effect of numerous such leaks is frequently significant.

The use of electronic leak detectors is becoming more and more common particularly when temperatures outside the limits of soap solutions are encountered. Such devices often enable a leak to be found quickly and efficiently; their relatively low cost produces a quick payback period.

Use a soap solution matched to the temperature of the system under test. At temperatures below the freezing point of water (0 C) most soap solutions are ineffective unless modified with alcohol or glycols to lower their freezing point. Similarly above water's boiling point (100 C) most soap solutions will boil off before they can bubble unless modified. At temperatures approaching either limiting value soap solutions will become less effective. Modified solutions, while useful and readily available, are usually less effective due to their higher viscosity and reduced wetting ability.

The most common of these devices is the helium leak detector, which generally senses the presence of helium outside a leaking joint or seal via a change in thermal conductivity. This instrument is very sensitive but, of course, can only be used effectively if the test gas is helium. These detectors can be used at depressed or elevated temperatures and are typically significantly more sensitive to tiny leaks that would tax the ability of soap solution to identify. Typical costs for small portable units range from $500-$1000. Some problems with this type of detector include the following.

The use of electronic leak detectors is becoming more and more common particularly when temperatures outside the limits of soap solutions are encountered. Such devices often enable a leak to be found quickly and efficiently; their relatively low cost produces a quick payback period.

Helium leak detectors are commonly used but require the unit to be pressurized with Helium. They are also frequently triggered (although much less frequently at at much higher levels) by gases or vapors other than helium. Most are actually thermal conductivity detectors which are much more sensitive to helium, with it's relatively high heat capacity, than almost any other gas; they are not, however, totally gas specific. Hence oil or grease on an operator's hand, if exposed to the probe, can cause a false trip. Light oils or similar materials on the piping exterior can also cause problems. Newer models are available which are helium gas specific are available but are significantly more expensive, typically $5,000-$6,000.

These detectors are also very sensitive to temperature differentials and will trip if exposed to moving gas streams or exposed to ambient temperature differentials. While such streams are frequently leaks, testing near control valves with constant air bleeds, in areas of high ventilation, or in exterior areas where wind currents are substantial can produce numerous false trips.

For these reasons, successful use of a helium leak detector generally requires experienced personnel. Each joint should be individually probed, slowly, and if a leak is indicated, the test should be repeated. Unless several repeat tests indicate a reading in the same place, it is likely that a false trip occurred.

Finally, helium leak detectors are frequently not explosion proof, and proper permits may be needed for use in an electrically classified area.

Portable combustible gas detectors are also in extensive use as leak detectors. While it is not advisable to perform initial leak tests with a flammable or combustible gas, it is frequently necessary to do the final testing with hydrogen or other flammable gases or to test the unit while in operation.

Older models of combustible gas detectors were frequently not suitable for use around individual joints, due the instruments' large size or limited sensitivity; newer models are available that detect flammable and combustible gases to very low levels, some down to the high parts per million range. Such units are light, fairly inexpensive and can frequently be obtained in intrinsically safe versions. They are relatively insensitive to soap solution, water, normal greases and dirt unless the sensor head is virtually immersed in such a material. While these detectors can be affected by temperature differentials, they generally can be re-zeroed easily in the field, allowing them to be used in exterior as well as interior areas. They are rugged and easy to operate. Typical models cost from $500-$1,500.

Acoustic leak detectors operate by locating the sound made by a leak. While useful for finding large leaks in inaccessible locations such as pipe racks or utility headers, these devices - at last in my experience -do not operate well in high noise areas and can be adversely affected by vibration in piping and support structures. For this reason, it is difficult for even experienced personnel to interpret their results correctly. In addition, such devices are usually not available for electrically classified areas and are much more expensive, typically $4,000-$8,000, than the devices discussed previously.

Leak Rate Detectors

As mentioned in my previous post, determining whether a test section or entire unit has passed its leak test can frequently be a time consuming task if large volumes or very small acceptable leak rate are encountered. Typical process gauges require long test times - often hours to an entire shift. Thus, alternative methods are valuable to reduce the time spent in leak testing.

The most common approach to this dilemma has been to use a more sensitive gauge that is moved from test section to test section. Use of a 30-60 cm (or even larger) gauge allows a more rapid evaluation to be made. However, this is a cumbersome and frequently expensive solution, as even high sensitivity pressure gauges may require several hours for determining leak rates in larger systems or those with low maximum acceptable leak rates.

A high pressure, low range differential pressure gauge is an ideal alternative for detecting very small pressure drops at high pressures (as low as 0.2 kPa pressure drop and as high as 42,000 kPa operating pressure). These instruments require no power supply, are compatible with any area electrical classification, and are rugged and easy to use. Typical costs range from $500-$1,000. The gauge is attached to the equipment under test and is pressurized with the system; the high pressure side is then isolated from the low pressure side, which is left connected to the unit - any leakage in the unit will cause a differential pressure. Because these gauges have a small range (typically less than 2 kPa), small leak rates can be determined rapidly, frequently in 5-10 min.

Although these differential pressure gauges are useful, their sensitivity is typically due to a large diaphragm movement, which can cause errors in the reference or high pressure side of the unit. It may be necessary to compensate for the large displacement of these diaphragms, as errors well in excess of 100% (depending on the diaphragm's size) can be developed. By attaching a compensating volume, typically 4 liters or larger, the error in such devices can be reduced to less than 5% at pressures up to 21,000 kPa. This compensating volume and the weight of the original unit make the combination rather large and bulky, but it can be mounted on a small cart or roll around table for easier use.

An alternative to the approach described above involves using an electronic or pneumatic differential transducer. Its small diaphragm displacement usually does not require any compensation, thereby eliminating the compensation vessel. However, such devices are frequently more expensive (typically $1,000-$2,000), particularly when very high pressures (over 21,000 kPa) and low ranges (less than 2 kPa) are involved. In addition, a power supply is necessary; this can create problems if used in a remote location or in an electrically classified area.

Thermal mass flowmeters have also been used as leak rate detectors. Unlike the devices discussed previously, mass flowmeters provide leakage rates in actual engineering units, so operating personnel can easily determine rates of widely varying test sections or components. This can be a useful feature if long term (3 - 12 month) tests are being performed. Measurement of leakage rates during turnaround, maintenance or other off condition periods can give the operating personnel an indication of a problem and enable them to take appropriate steps to locate, isolate and repair the problem without losing or seriously affecting the quality of the research results.

Thermal mass flowmeters are expensive (typically $2,000-$3,000 installed) particularly in higher pressure versions. They require both initial calibration and a power supply. They are frequently not available for electrically classified areas; further they require experience to operate successfully, since they are usually very sensitive to small ambient temperature swings and occasionally to process pressure changes. In general, such devices are suitable only where frequent leak testing is routinely required, long term leak free runs are necessary and experienced operating and maintenance personnel are available.

In closing, the best and most reliable way of locating leaks remains a trained and knowledgeable operator equipped with the right tools to do the job right.