Sorting a Backlog by the Outage It Can Actually Be Done In

The plant never stops; only a train comes out

Every framework for planning inspection work assumes a shutdown. Water and wastewater utilities do not have one. Influent arrives at whatever rate the weather and the population dictate, and finished water demand has to be met on the day it occurs. The nearest equivalent to a turnaround is taking one process unit out — a clarifier, an aeration basin, a filter, a digester, a storage reservoir, one side of a duplicated pump station — and asking the remaining units to carry the load.

That single difference reshapes the deficiency backlog. Work is not sorted by priority alone; it is sorted by which unit it sits on and when that unit can be released. Two findings of identical severity on adjacent structures may be separated by four years of scheduling because one sits on a basin that can be isolated any month of the year and the other sits on a reservoir that can only come down in a narrow shoulder season.

So a backlog presented as a single ranked list is actively misleading in this sector. What the maintenance planner needs is the backlog sliced by outage opportunity: everything that can be done during the next filter outage, everything waiting on a reservoir drawdown, everything that requires bypass pumping on a force main. Findings that are not attached to a feasible window are not scheduled work. They are a wish list.

The window is set by permit and by season

The limits on a utility outage are external. An NPDES permit sets effluent quality that must be maintained continuously, and taking a treatment train out reduces the margin available to meet it. Wet weather closes windows entirely at many plants, because peak flow events need every basin in service and the forecast horizon is shorter than the outage. On the potable side, peak summer demand closes reservoir outages for months at a stretch.

This makes window availability a fact to be discovered rather than negotiated, and it changes who owns the schedule. In an industrial turnaround, engineering proposes and operations reacts. In a utility, operations declares what is possible and engineering fits the scope inside it. A deficiency system that models outages as project schedules, freely movable, misrepresents how the decision is actually made.

The practical requirement is that the register knows the outage calendar, and that findings attach to outages rather than to dates. When operations moves a basin outage from October to April because of a wet forecast, everything scoped into it should move with it, and anything that no longer fits should surface immediately as a re-deferral decision rather than being discovered on the morning the crew arrives.

Cure, disinfection and clearance own the critical path

Ask a planner how long it takes to recoat the interior of a steel water storage tank and you will usually get an answer in working days for blasting and application. That answer is the smaller half of the outage. Coatings for immersion service cure on a schedule set by the manufacturer's data sheet at a reference temperature, and the temperature inside an enclosed structure in a cold month is not the reference temperature. Cure that reads as five days on paper can be twice that in practice, and applying immersion service before full cure damages the coating on first fill.

Application itself is weather-bound in a way that surprises people who plan indoor work. Surface preparation and coating are governed by ambient temperature, relative humidity and the margin between substrate temperature and dew point. In a covered reservoir with limited ventilation, dew point control can consume more of the window than application does, and dehumidification equipment is itself a mobilization item that has to be scoped in advance.

Then comes clearance. Dry film thickness verification under SSPC-PA 2 and holiday detection cover the whole coated area, not a sample. Disinfection of a potable water storage facility follows AWWA C652 with a defined method and contact time, and return to service depends on bacteriological results that come back on the laboratory's clock, usually as consecutive sample sets. None of these is labor. All of them are duration, and a plan built on crew hours contains none of them.

Access dependency is the field most backlogs are missing

Look at a typical utility deficiency record and you will find asset, description, severity, estimated cost and a status. What you will not find is how the work is physically reached. Yet access is what determines whether two findings can share an outage, what the mobilization costs, and how long the window has to be.

The dependencies worth making into structured fields are few and highly repetitive. Does this require dewatering, and of what volume. Does it require permit-required confined space entry. Does it require bypass pumping or temporary treatment. Does it require scaffolding or rope access. Does it need materials certified to NSF/ANSI/CAN 61 because the surface contacts potable water. Does it require disinfection and clearance before return to service. Six yes-or-no fields, answered when the finding is raised by the person who saw the asset.

The payoff is immediate and arithmetic. With those fields populated, the planner can ask the backlog which findings share a dewatering event, which ones justify a single confined space setup, and which ones will extend the window regardless of how small the work is. Without them, the same questions require reading every record's free text, which is why in practice nobody asks them and outages get scoped from memory.

PACP grades, and what a grade does not tell you

Collection systems already solved part of the normalization problem that plagues other sectors. NASSCO's Pipeline Assessment Certification Program gives CCTV operators a defined defect vocabulary and a one-to-five severity scale, separating structural defects from operations and maintenance conditions and rolling individual defects into segment ratings. Two contractors coding to PACP produce comparable output, which is more than most inspection disciplines can claim.

The limitation is that a PACP grade describes the pipe, not the consequence of the pipe failing. A grade 5 structural defect in a short run under an open field and a grade 4 under a hospital access road or crossing a creek are not the same problem, and no coding standard can tell them apart because consequence is not visible in the video. Utilities that rank rehabilitation purely by grade end up funding the wrong segments and can defend the decision, which is worse than not being able to.

A deficiency register earns its keep here by carrying the second dimension: criticality of the segment, based on diameter, depth, surface use, proximity to receiving water, redundancy and customer impact. That value is assigned once, at the asset, not repeatedly at each finding. Combining a coded defect grade with a stored criticality gives a rehabilitation ranking that survives a council presentation, and it is reproducible year to year as new CCTV comes in.

Deferred findings and an aging clock that counts outages

Everything in a utility backlog is deferred at some point, and most deferrals are correct. The reservoir cannot come down until spring, the basin is needed through the wet season, the funding is in next year's program. Deferral is not a failure of discipline; it is the normal operation of a system with fixed windows and an annual budget.

What fails is the tracking of it. A finding deferred into the next outage is typically written into a note, and the note is read by whoever plans that outage, if they know to look. When staff change, when a consultant does the planning, or when the outage is delayed twice, the note stops being read. Findings from the previous internal inspection of a reservoir routinely resurface in the next one, five years later, as new findings, because nothing carried them forward.

The fix is to make deferral a first-class action that requires a target — the next outage of a named type on a named asset — and to age findings in windows rather than days. A finding that has missed two qualifying outages is escalated automatically and visibly, regardless of its original severity, because missing two windows means something in the planning process is repeatedly discarding it. That single metric surfaces the systemic problem that a calendar-based aging report never will.

Fiscal years, project numbers and emergency declarations

A private operator can authorize a repair the day it is scoped. A public utility usually cannot. Capital work moves through a capital improvement program adopted annually by a council, board or commission, with a budget line and a project number, and the approval calendar has nothing to do with when an inspector finds a problem. Operating funds cover small work; anything of size waits for the cycle or requires an emergency declaration that carries its own scrutiny.

This means a utility deficiency register has a financial dimension that industrial registers do not. A finding needs to know which funding path it is on — operating budget, an existing CIP project, a future CIP request, or an emergency route — because that determines when it can be executed at least as strongly as its severity does. A backlog that shows only technical priority will repeatedly propose work that has no money attached, and the credibility cost of that lands on the engineering group.

It also creates a real analytical need at budget time. The engineering manager preparing next year's CIP request needs to show the backlog grouped by proposed funding year, with the consequence of each deferral stated in terms a non-technical board can weigh. A register that can produce that grouping directly turns budget preparation from a month of spreadsheet work into a query, and it makes the request defensible because every line traces to a dated inspection finding with evidence behind it.

Evaluating a system before the next window opens

Test it against a real outage you have already run, preferably one that overran. Load the findings that were in scope, the access dependencies each one carried, and the actual durations. Ask the system to reproduce the window. If it plans the outage on labor hours and returns a duration close to the original optimistic estimate, it will make the same error on the next one.

Then test the dependencies and the deferrals. Ask where dewatering, confined space entry, bypass pumping, cure and clearance live, and confirm they are structured fields the planner can filter and aggregate on. Ask what happens when operations moves an outage by six months. Ask what a finding deferred at the last reservoir inspection looks like today, and whether the planner of the next reservoir outage will see it without being told to look.

Finally, test the money. Ask whether a finding can carry a funding year and a project reference, and whether the backlog can be grouped that way for a budget submission. Atlantis builds deficiency and recommendation tracking as a module of an Odoo-based NDT ERP, alongside inspection reporting, digital twin and 3D laser scanning services for structures where access is the constraint. To walk a real outage through the model before your next window opens, contact info@atlantisndt.com.

Request a consultation