LDAR coverage before the auditor asks

Most leak detection and repair programs can answer the first question an auditor asks: show me the monitoring. The records are there, by component, by date, with readings, leaks found and repairs made. The harder question comes next. How do you know every component that belongs in the program is in it?

That is a coverage question, and monitoring records cannot answer it, because they only describe components already in the inventory. The answer lives in a comparison between the inventory and the plant. This field note is about making that comparison routine, against a verified P&ID graph, so the answer exists before anyone asks for it.

What an LDAR program has to cover

LDAR programs exist to find and fix fugitive emissions from process equipment: the small, steady leaks from packing, seals and joints that add up across a refinery or a chemical plant. In the United States the obligations come from federal Clean Air Act standards, state and local rules, permits and sometimes consent decrees. Which rules apply to which units, with what leak definitions and monitoring frequencies, varies by jurisdiction and by facility. The structure is broadly similar:

• An inventory of regulated components. Typically valves, connectors, pumps, compressors, pressure relief devices, sampling connection systems and open-ended valves or lines. Each is identified, usually with a physical field tag, and classified by type and by service, such as gas or vapor, light liquid or heavy liquid.
• Periodic monitoring. Often by EPA Method 21, in which a technician moves the probe of a calibrated portable analyzer around each potential leak interface of a component and compares the reading with the leak definition the applicable rule sets. Some programs allow or require optical gas imaging for some equipment, and many add visual inspections of pump seals between instrument readings.
• Repair on a deadline. A first attempt and a final repair within periods the rule sets, with delay of repair allowed only under defined conditions, such as a repair that needs a unit shutdown.
• Special designations. Components that are unsafe or difficult to monitor carry their own schedules and written justification.
• Records and reports. Component identification, monitoring results, leaks, repairs, delays, and periodic reports to the regulator.

Everything downstream depends on the first item. A component that is not in the inventory is never scheduled, never monitored and never repaired, and its absence appears in no monitoring record.

How inventories drift from the plant

An inventory is usually built once, from P&IDs and a field tagging survey, and maintained after that. The plant changes faster than the maintenance.

• Management of change. A bypass around a control valve adds valves, connectors and perhaps a drain. A new transmitter adds a root valve and instrument fittings. If the MOC closes before the LDAR coordinator hears about it, those components are installed and unmonitored.
• Turnarounds. Equipment is opened, spools are replaced, valves are swapped for other types, and temporary connections go in and sometimes stay. Field tags disappear under new insulation or leave with the old valve.
• Process changes. A change in stream composition can bring a line into regulated service or take it out. Nothing physical changes, so nothing prompts a field update.
• Removals. Components demolished in the field linger in the inventory, and monitoring records against equipment that no longer exists are a problem of their own.
• Red-lines that never came back. The drawing the inventory was built from stops matching the plant, and the next review inherits the error. We look at that backlog in management of change starts with a diff.
• Two owners, two lists. Operations and engineering own the drawings. The environmental group, often with an LDAR contractor, owns the inventory. Each list is updated on its own schedule.

None of this points to a careless program. It is what happens when two records of one plant are kept by different people through years of change.

Monitoring records prove you checked the components you knew about. Coverage is about the ones you did not.

Reconcile the inventory against a verified graph

The natural reference for coverage is the P&ID: the document every MOC is supposed to update, and usually the one the inventory was first built from. On paper, a full comparison is impractical. In a verified graph it becomes a query.

Our P&ID platform reads each sheet into a typed graph of equipment, lines, valves and instruments, linked across sheets, and engineers verify it tag by tag before anything runs on it. Line numbers are parsed into size, service and piping class, and every item traces to the sheet and region it came from, as described in reading a P&ID like an engineer.

Start from the applicability decision

Which lines are in regulated service is an applicability determination your environmental team makes from stream data and the rules that apply. It is not something to read off a drawing. The graph takes that determination as an input, often from the highlighted LDAR P&IDs many sites already keep, marks the lines in scope, and carries the boundary forward, so a line added in a later revision surfaces as a scope question instead of falling silently outside it.

Match on location, not only on tag

LDAR tags and P&ID tags are different numbering systems. Many manual valves on a P&ID carry no tag at all; they are identified by the line they sit on and their position along it. An LDAR inventory usually identifies a component by its field tag number and a location description, often with a drawing reference. So reconciliation matches on what both records share: unit, drawing, line, neighboring equipment or instrument, and component type. It compares counts per line segment and type first, then drills into the segments that disagree.

Be honest about connectors

P&IDs show valves, pumps, compressors, relief devices, sample connections, vents and drains well. They do not show every flange or threaded fitting; many joints appear only on isometrics or in the field. So the graph gives a floor for connectors, not a count. Where the inventory holds fewer connectors on a line than the drawing implies, the line goes on the field list. Connector coverage is confirmed in the field.

What the reconciliation finds

Each finding carries its evidence, so a technician knows exactly where to look:

• On the drawing, not in the inventory. A likely untagged component, often from an MOC or a turnaround, with the sheet, region, line and neighboring equipment.
• In the inventory, not on the drawing. A removed component still on the route, or a field change never drafted, with the inventory record and the place on the current sheet where it should appear.
• Type or service mismatches. A component classed one way in the inventory and drawn another, or a line whose service changed.
• Open-ended valves or lines. A vent or drain with no cap, plug, blind flange or second valve shown, which equipment leak rules commonly require.
• Lines in scope with nothing in the inventory. Usually a line added by MOC, or a boundary drawn in the wrong place.
• Changes since the last reconciliation. Components added or removed in each revision delta, tied to the MOC that made them.

Every finding becomes a field check with a location attached. Technicians confirm what is installed and tag what is not, the coordinator updates the inventory, and where the drawing is wrong, an engineer red-lines it.

What a coverage response looks like

When the question comes, from an auditor, an inspector or your own internal audit, a good answer is a package rather than an assurance:

• Scope. The process units and streams in regulated service, the applicability basis, and who determined it.
• Drawing basis. The P&IDs used, with revisions, and when and by whom each was verified.
• Counts. By unit and component type: on the drawings in scope, in the inventory, and matched.
• Every discrepancy, with evidence. The sheet and region, the inventory record if one exists, the field result, and the disposition: added and first monitored on a stated date, confirmed removed, determined not in regulated service with the basis, or drawing corrected.
• The change trail. Each MOC and turnaround since the last reconciliation, and when its components entered or left the program.

The package does not claim the program is perfect. It shows that the program knows its own boundaries, and that the gaps were found by the facility, documented and closed.

What to do about a gap, including whether it has to be reported, is a decision for your environmental team and often your counsel. A reconciliation produces evidence. It does not make that call.

Who stays accountable

Reconciliation changes how gaps are found. It does not change who owns them:

• The environmental team owns applicability: which units, which streams, which rules.
• The LDAR coordinator owns the inventory and decides how each discrepancy is dispositioned.
• Field technicians confirm what is installed and tag what is not.
• Plant engineers verify the P&ID graph and correct drawings that are wrong.
• The people the rules name certify reports.

We do not monitor components, determine applicability or certify anything. Our part is keeping the drawings verified and current, and every comparison traceable to its sheet, with an audit log of every verification and change. More on deployment and security on our enterprise page.

Find the gaps yourself

Coverage is the part of an LDAR program that monitoring cannot show. It drifts quietly through MOCs, turnarounds and process changes, and sooner or later someone finds the drift. A reconciliation against verified drawings, run after each turnaround and kept current with each revision, means the facility finds it first, with the evidence to show what it did about it.

The same verified register supports revamp scoping, as we describe in brownfield before the walkdown. For how we work with operating assets, see energy and chemicals.