In 1990 leaks on the launch pad necessitated grounding of the Space Shuttle fleet until the leak source could be identified, a process costly in monetary and scheduling disruptions. The U.S. Navy continuously monitors for toxic hydrogen sulfide build-ups in wastewater containment areas aboard sea-going vessels. Offshore oil platforms use dynamic- and static-condition leak detection systems on under-sea gas and oil transfer lines. Major airports rely on fuel leak detection systems. Liquid and gas/vapor detection systems are deployed in every conceivable location and industry.

Liquid leak detection can be divided into two areas, detecting the loss of liquids from pipelines and vessels, and detecting the presence of liquids in unwanted areas such as water under a raised computer room floor.

Dynamic and static leak detection systems are being successfully deployed using combinations of standard process flow, pressure, and temperature instrumentation, differential equations to simulate the pipeline, and/or pattern recognition (neural network) technologies.

Generally, dynamic leak detection models provide faster response than static models, but static models are able to detect much smaller leaks than the dynamic models.

Detection of liquids in unwanted areas occurs in a variety of facilities including manufacturers of toxic chemicals. Building-in-a-building designs, negative interior pressure buildings, jacketed piping, monitored trenches and drains are examples of the precautions used to prevent or detect accidental releases. Often detecting liquids in unwanted areas is not considered until after the first bad experience.

A unique liquid detection technology from Raychem (Menlo Park, Calif.) uses flexible, liquid absorbing cables to detect water, conductive liquids, and liquid hydrocarbon fuels and solvents in unwanted areas. The cable contains two sensing wires, a signal wire, and a continuity wire. Using low dc voltage and microamps, a circuit is established and instrumentation monitors the circuit's current flow. When moisture is absorbed into the cable, the circuit shorts and current flow increases generating a spill alert. The instrumentation switches to a "locating" mode and provides the distance from the instrumentation to the spill location.

Hazardous gas/vapor leak detection can be divided into three categories:

• Combustible gases;
• Toxic gases; and
• Oxygen displacing gases.

In each category the required gas monitoring system can be one of two types. For minor risk applications, a simple gas monitor interfaced with a panel that sounds an alert may be adequate. Higher risk applications often require continuous (e.g., 4-20 mA) output sensors; however, before integrating the sensors into the control system consider the risk when the control system is unavailable, such as during a software upgrade.

Pattern recognition (neural network) software provides an opportunity to use sound, such as a hydrophone, and a trained neural network to distinguish between the noises gas bubbles make when leaking from an under-sea gas transport pipeline and other ocean noises such as ship propellers.

Sometimes a fluid leak (gas or liquid) may not present a hazardous risk, but has economic impact. For example, steam or water leakage adversely affects the efficiency of a boiler. The use of acoustic leak detection can locate leaks in boiler tubes, feedwater heaters, pipelines, and valves.

When a fluid moves from high pressure to low pressure, turbulent flow occurs. The turbulence generates an ultrasonic "hissing" sound that increases in intensity near the source. By continually reducing the sensitivity of a portable ultrasonic detector, the user is lead to the source of the leak.

Leak detection does not always involve sensors, software, or instrumentation. Coating materials can be applied at likely leakage points such as flanges and packings. These coatings change color in the presence of toxic chemicals, such as chlorine.

Reducing community and environmental risk from hazardous liquid and gas leaks is becoming more practical and precise with the integration of improved instrumentation and real-time software technologies.