Plan the observation as an interference experiment. Hold the nominated component and camera position steady, change one permitted source of acoustic interference, and classify what happens to the highlighted source. This gives hydrogen leak detection planning a sharper purpose in ventilated work areas: separating a component-attached ultrasonic source from fan, machinery or airflow-related activity.
The experiment answers a location question, not an airflow question. It can show whether an ultrasonic source remains attached to a visible fitting under recorded conditions. Hydrogen identity, gas movement, room safety and ventilation performance require other evidence and site controls.
Pick one candidate and one interferer
Start with a nominated pressurised connection or compact component group. Avoid using the whole room as the test object. A useful candidate might be a regulator connection, valve packing area or flange pair that can be seen from a stable position.
Then identify the most plausible competing ultrasonic source near that view. Examples include an extraction fan, a pneumatic tool, a rotating machine or turbulent airflow at a louvre. Select only one interferer for the first pair. If several conditions change together, the comparison cannot show which change mattered.
Distran ultrasound imaging cameras localise ultrasound and place the source on an optical scene. Naming the dominant interferer creates a focused test of whether the highlighted source remains attached to the candidate component when that interferer changes state.
Lock the geometry before changing state
Mark a repeatable camera position using a floor feature, platform post or measured stand-off. Frame the same visible landmarks around the candidate in both observations. If safe access or the permitted operating arrangement prevents that repeatability, treat the records as separate views rather than a pair.
Before the first capture, note:
- candidate component and visible landmark;
- camera position, direction and obstruction;
- interferer selected for the comparison;
- relevant plant state that will remain fixed;
- ventilation opening states that will remain fixed; and
- file naming or time basis for the two captures.
This short control set prevents a changed zoom, viewpoint or plant condition from masquerading as an effect of the selected interferer. It also keeps the experiment small enough to repeat.
Run an A-B-A sequence when the site permits
A two-state pair is useful, but an A-B-A sequence is stronger when the authorised work plan allows it. Capture state A, change the selected interferer to state B, then restore state A and repeat the view. The return to A tests whether the original spatial pattern also returns.
For a fan comparison, the sequence might be:
- A1: nominated fan running, candidate in the defined process state.
- B: nominated fan stopped, with all other listed conditions unchanged.
- A2: fan running again after the settling period required by the site.
No general rule says a fan should be stopped. The permitted states and settling period come from the site’s arrangements. If only A and B are available, retain the simpler pair and lower the strength of any comparison.
Abandon the sequence if the component process state changes, another strong machine starts, the viewing position moves materially or an access obstruction enters the frame. A shorter valid pair is more useful than three files whose differences cannot be isolated.
Classify the spatial behaviour
Read the captures by source position rather than by a generic increase or decrease. Four patterns lead to different next tests.
Component-locked: the highlight remains centred on the same visible fitting through A1, B and A2. This makes component-focused follow-up the next logical question because the source survived the interferer change.
Interferer-locked: the highlight appears at the fan, louvre or machine when that source is active and reduces or disappears when it is inactive. The next move is to improve separation from that source, not to infer gas movement.
Unstable field: several highlights move around the frame without returning to a repeatable component or interferer location. Try a different view, shorter stand-off or another permitted comparison condition.
No localised source: no stable source appears at the accessible candidate. That result may redirect the work to another view or detection method. It is not a room-wide finding.
The A-B-A shape matters. If a candidate highlight appears in A1, disappears in B and returns in A2, it behaves with the selected state. If it remains at the fitting throughout, it behaves independently of that state. If A2 differs from both earlier captures, look for an uncontrolled change before interpreting the sequence.
Use a second angle to resolve overlap
One optical view may place a fan edge and a connection close together. A second accessible angle can separate those positions geometrically. It should be treated as a second experiment with its own fixed frame, not inserted midway through the first sequence.
Imagine a roof extract fan behind a small hydrogen connection from the first platform position. In that frame, both acoustic origins could overlap. Moving 30 degrees around the connection may place the fan to the left and the fitting to the right. Repeat the permitted state pair from the new position. A source that tracks the fitting in both frames presents a different follow-up question from one that tracks the fan.
This triangulation scenario is especially useful in compact enclosures. It turns a confusing bright patch into two competing spatial hypotheses that the observer can test without making a dispersion claim.
Decide whether acoustic imaging is the next method
Use the classified behaviour to choose the next step. A component-locked source supports a defined component enquiry. An interferer-locked or unstable field supports another observation design. No localised source may justify a different gas-detection method, another accessible face or a decision that the acoustic route is not suitable for the question.
For a defined field observation, on-site gas leak detection and quantification uses a handheld acoustic-imaging device. A good enquiry specifies the hydrogen or process context, candidate connections, permitted A-B-A states, fixed viewing positions and the records required from the experiment.
The final briefing can be concise:
- one candidate component per sequence;
- one selected interferer;
- the states the site permits;
- the fixed geometry and optional second angle;
- any event that would invalidate the comparison; and
- the component-locked, interferer-locked, unstable or absent pattern to be reported.
Let the controlled comparison decide
A bounded hydrogen observation can move from method to field work through the on-site gas leak detection and quantification service at Aquip. Arrange the stated hydrogen task once its permitted conditions are settled. Any conclusion still depends on the indication following the component through the controlled A-B-A sequence while geometry remains fixed.
The field conclusion must still come from the designed comparison. Hold the viewing position, candidate component and process condition steady for A, introduce only the chosen interferer for B, then restore A. Record any geometry change or uncontrolled event that breaks the sequence.
Treat a hydrogen indication as useful only when the source behaviour follows the component while geometry and the chosen interferer remain controlled across A-B-A. If those controls fail, the sequence identifies another observation need rather than a component conclusion.