Memorizing only “LNG at the ceiling, LPG at the floor” is not enough

When locating combustible-gas detectors, people are often taught to install them high for natural gas because it is lighter than air and low for LPG because it is heavier. That is a useful starting point for considering a small gas leak into stable air at room temperature. Real leaks at LNG storage, unloading, and vaporization facilities and at LPG filling and storage facilities, however, are more complex than this one sentence suggests. Temperature drops sharply as liquid vaporizes, a pressurized release entrains surrounding air, and fans, louvers, doors, vehicles, and structures alter the flow.

The UK Health and Safety Executive also advises positioning fixed sensors to detect accumulation before gas creates a serious hazard, while considering the process equipment, sensor type, gas properties and dispersion characteristics, and ventilation pattern together. Relative density is one factor in selecting installation height, not a standalone rule that automatically places detectors on a drawing. The correct sequence is to understand the differences between LNG and LPG and then map the expected leak paths.

LNG and LPG differ from the way they are liquefied

LNG is natural gas cooled to about minus 162 degrees Celsius to turn it into a liquid. Its main component is methane, but its actual composition may include ethane, propane, nitrogen, and other constituents, with proportions varying by source and processing conditions. When LNG leaks, it rapidly boils as it exchanges heat with the surrounding ground, air, and water, producing a large volume of vapor. What matters here is not just the behavior of methane at room temperature, but also the temperature of the newly generated cryogenic vapor.

LPG is primarily a mixture of propane- and butane-family hydrocarbons and is stored and transported as a pressurized liquid. If a valve or hose fails, part of the liquid immediately flash-vaporizes, while the remaining liquid may form a cold pool and continue evaporating. The release rate and vapor composition vary with the propane-to-butane ratio and storage temperature and pressure, but propane and butane vapor at room temperature are both heavier than air.

Accordingly, neither “LNG” nor “LPG” is the name of one pure gas. The actual vapor density and detector response are affected by composition, temperature, and the extent of mixing with air. Designers must first confirm a representative composition and its possible variation from current safety data sheets, supply specifications, and the process material balance.

Vapor density changes with temperature and composition

At the same pressure, gas density generally increases with molecular weight and decreases with absolute temperature. Even methane, whose molecular weight makes it lighter than air, can contain more mass in the same volume and be heavier than the surrounding warm air when it is extremely cold. Conversely, as leaked vapor mixes with surrounding air and warms, the density difference decreases. For fuels containing several constituents, average molecular weight also varies with composition.

The NIOSH Pocket Guide to Chemical Hazards gives propane a relative gas density of 1.55 at the same temperature. This means propane vapor remains heavier than air at room temperature. Mixing in butane may increase the average molecular weight further. This number, however, must not be applied as a fixed property of an entire LPG cloud in the field. Vapor cooled by flash vaporization, a cloud mixed with air, and a 2-phase release containing droplets behave differently from the equilibrium properties of pure propane alone.

Relative-density data provide direction, but a real leak cloud continually changes state. A universal height such as “a certain number of centimeters above the floor whenever relative density exceeds 1” is therefore fragile. A sensor too far from the leak source will respond only after dilution, while one too low may fail because of rainwater, wash water, dust, or impact. The location must also account for protective construction and maintenance access.

Cold LNG vapor may initially travel along the ground

Natural gas at room temperature is normally lighter than air and rises. Vapor immediately after an LNG leak is an exception. The US PHMSA explains that vapor produced at LNG’s normal boiling point is about 1.5 times heavier than air at 25 degrees Celsius and has negative buoyancy until it warms sufficiently after release to the atmosphere. This cold, dense cloud can travel along the ground, impoundment, trenches, and low terrain. A visible white cloud is fog formed when the cold gas condenses atmospheric moisture, so it does not precisely coincide with the edge of the flammable range.

As the vapor mixes with ambient air and warms over time, methane-rich vapor may pass through neutral buoyancy and then rise. This must not be interpreted to mean that only high-level sensors are needed because “it will rise after a while.” FERC’s LNG environmental assessment explains that cold LNG vapor is initially heavier than air and becomes lighter as it warms, but actual tests and models show the cloud may already have diluted below the lower explosive limit before lifting off the ground. In other words, the zone where a flammable cloud must be caught early may be at low level near the leak source and along its ground-hugging path.

Other leak scenarios also exist. A small gas leak from a natural-gas line where the gas has already been vaporized and warmed is more likely to move upward. Boil-off gas at the top of an LNG tank, a vaporizer outlet, and pockets below an indoor ceiling require a different arrangement from a low-level release. This is why even a single facility must treat a cryogenic-liquid leak and a room-temperature gas leak as separate scenarios.

LPG has a strong tendency to remain in low areas

Propane and butane vapors remain heavier than air even after warming to ambient temperature. The HSE explains that vaporized LPG moves downward or remains at low levels. Leaked LPG can spread along the floor and collect in pits, sumps, drains, cable trenches, basements, and inside impoundments. The possibility that ignition at a distant low point could flash back along the vapor path to the leak source must also be considered.

That does not mean one detector on the floor is sufficient. A pressurized release initially emerges as a jet in the direction of the valve or flange and entrains surrounding air. Indoor supply fans may carry LPG upward or into another compartment, and outdoor wind can accumulate a low cloud in a recirculation zone behind a structure. The appropriate approach is to detect first near likely leak points—such as filling-hose connections, pump seals, compressors, vaporizers, and pressure-relief discharge points—and then add sensors at expected migration and accumulation points.

Low-level sensors near drains or trenches are important, but flooding or contamination can leave them inoperative precisely when an incident occurs. Confirm the sensor’s permitted temperature and humidity, ingress protection, suitability for hazardous locations, condensate drainage, impact protection, and calibration space. Monitoring a low-level hazard and installing the instrument close to standing water are not the same thing.

Actual placement starts at the leak source

Detector placement begins by dividing equipment into small leak scenarios. Instead of treating an entire storage tank as one point, mark connections, liquid outlets, pump and compressor seals, flanges, sampling points, unloading arms and hoses, filling connectors, vaporizers, regulators, instrument tubing, drains, pressure-relief valves, and vent discharge locations. Include not only normal operation but also startup and shutdown, unloading, purging, maintenance, misoperation, and power-loss states.

Apply the following questions to each scenario.

  1. What is released—liquid, a cold 2-phase mixture, or room-temperature gas—and in what direction?

  2. How do release pressure, hole size, and duration affect air entrainment and travel distance?

  3. Where do ground slope, impoundments, pits, trenches, walls, and equipment trap the cloud?

  4. How do supply and exhaust openings, fan operating modes, open doors, vehicle movement, and seasonal winds alter the flow?

  5. Where must the gas be detected before it reaches workers, ignition sources, control-room air intakes, and evacuation routes?

  6. When one sensor is out for failure, calibration, or maintenance, is the hazardous area left unmonitored?

Indoors, the positions of louvers, doors, supply and exhaust openings, and whether fans are running are important. Outdoors, examine wind direction and speed, terrain, equipment congestion, and building wakes. The HSE notes that smoke tests or computer models can be used where necessary to predict leak behavior and optimal locations. For large LNG or LPG facilities, it is reasonable to perform dispersion analysis and detector-coverage assessment for representative leak scenarios, then supplement them during commissioning with smoke or tracer-gas tests under actual ventilation conditions.

Design multiple lines of defense rather than a single height

In LNG handling areas, first evaluate the expected paths at low levels around pumps, unloading equipment, and piping where cryogenic liquid can leak, as well as inside impoundments. At the same time, assess the potential for high-level accumulation at vaporizer outlets, compressor rooms, and ceiling pockets where warm natural gas is present. Do not assume that a sensor array at one height represents both scenarios. Where necessary, combine low- and high-level sensors, sensors adjacent to leak sources, and monitoring of HVAC intakes.

In LPG areas, prioritize locations near filling and transfer connections and low accumulation points such as pits, trenches, and drains. Rather than repeatedly installing floor-level sensors at uniform intervals throughout a large open area, it is better to increase their density based on actual leak frequency and the paths gas is likely to follow. Also check whether impoundment walls will retain vapor, whether openings connect to an adjacent building, and whether drainage is segregated to a safe location.

Personal portable instruments supplement blind spots in fixed installations, but the two do not replace each other. A design in which a portable alarm sounds only after a worker has entered the hazardous cloud is not early equipment protection. Conversely, even in an area with fixed sensors, use a remote probe before entry into confined spaces, opening drains, opening equipment, or maintenance to measure high, middle, and low levels and expected pockets, and decide whether continuous monitoring is needed during the work.

Sensor type and calibration gas matter as much as placement

Even at the correct location, an alarm will be delayed if the sensor does not respond appropriately to the target mixture. Catalytic-combustion sensors require oxygen and may be affected by catalyst poisons or inhibitors. Infrared sensors can measure hydrocarbons without oxygen, but their target gases and response characteristics vary by product. Check the manufacturer’s performance range for the presence of cryogenic vapor, condensate, high humidity, dust, and silicone-containing substances.

A detector calibrated with methane will not give the same indication for propane or butane, and vice versa. The HSE explains that an instrument should ideally be calibrated with the gas to be measured and that, when another calibration gas is used, the correction relationship provided by the manufacturer should be applied. If one instrument monitors both LNG and LPG, document the least favorable response over the expected composition range, alarm settings, and calibration strategy. A bump test is a functional test confirming that gas reaches the sensor and the alarm operates; it must be distinguished from a full calibration that assures quantitative accuracy.

Detectors do not replace ventilation, leak isolation, impoundment and drainage, explosion protection, or emergency response. Testing must extend to which valves and equipment stop automatically or manually on alarm, whether the ventilation mode directs gas toward evacuation routes, and whether alarms are visible both in the field and in the control room. If equipment layout or ventilation operation changes, review the existing detector positions again.

The conclusion of a site review is a “leak scenario,” not a gas name

The tendency of LPG to move to low areas and of room-temperature natural gas to rise remains valid. Cold LNG vapor, however, may initially flow along the ground, then change buoyancy as it warms and mixes with air. Vapor density itself varies with temperature and composition. Therefore, a rule that always places LNG detectors at the ceiling and LPG detectors at the floor can miss the actual hazard.

A good design verifies the material data, separates liquid, 2-phase, and gas releases, identifies the most likely sources and discharge directions, and places sensors along migration and accumulation paths created by ventilation and terrain. It then validates those assumptions through smoke testing, dispersion modeling, and commissioning tests, and confirms calibration for the target gas and maintenance access. Relative density answers the first question, but the full leak scenario determines a safe detector location.

Sources

Cover photo

Original photo on Pexels