Scheduling inspections when the data comes back from someone else

The calendar is set by hydraulics, not by maintenance preference

In a refinery a unit outage is a planning decision. In a water utility it is a capacity calculation. An elevated tank cannot be drained for interior inspection unless the remaining storage in that pressure zone still covers peak-hour demand plus the fire-flow duration the utility's rating assumes. That arithmetic changes with the season, which is why tank washouts cluster in spring and autumn and why a schedule that treats a tank inspection as a movable half-day task will collide with the one narrow period the hydraulic picture actually allows.

On the wastewater side the constraint inverts. A clarifier, digester or wet well can only come down when a parallel train carries the flow, and the flow is weather. A crew mobilised into a basin during a wet-weather event is a crew standing on the bank being paid for standby. The scheduling module has to hold a seasonal availability rule on the asset itself, not as a comment inside a work order, so that a due date landing in the wrong month is flagged when the annual plan is built rather than the week the crew is booked.

Both are dependencies rather than preferences, and they are why a generic maintenance calendar under-serves a utility. Due dates come from the asset management plan and from state primacy requirements. Access windows come from the hydraulic model and the rainfall record. Crews come from a term contract with a fixed annual quantity and a fixed unit rate. A system that cannot represent all three at once will always be reconciled by hand in a spreadsheet that one person maintains and nobody else can audit.

You are not buying an inspection, you are buying a deliverable

Almost nothing in a utility's inspection programme is performed by utility staff. Tank interiors go to a specialist contractor or a potable dive team. Sewer condition goes to a CCTV crew. Transmission main assessment goes to an electromagnetic inspection firm. Cathodic protection goes to a corrosion consultancy. Coating condition goes to a certified coating inspector who may work for any of them. Each of those firms has a house report format that was designed to satisfy their own quality system, not to feed yours.

The consequence is predictable and expensive. Two surveys of the same tank floor, five years apart, cannot be subtracted from each other because the second contractor used a different grid, a different reporting unit and a different convention for what counts as the governing reading. The utility ends up with a shelf of technically competent reports and no trend. Capital renewal decisions are then made on the most recent report alone, which is exactly the decision the inspection programme was supposed to improve.

The point of leverage is the work order, and it exists before the crew arrives. If the scheduling record carries the required grid identifiers, the required units, the required minimum and average convention, the technician qualification level under SNT-TC-1A, the calibration reference and the required file format, then the deliverable is specified. If it carries a scope line and a date, the deliverable is whatever the contractor already had a template for. Very little of this is negotiated after the fact; almost all of it is settled by what the work order said.

What 'trusted' has to mean when nobody from the utility was standing there

Trust in third-party data is not a feeling, it is a set of attached artefacts. For a thickness survey it is the technician's certification level and method, the equipment serial number, the calibration block and the verification times before and after the survey. For a coating assessment it is the gauge type and its calibration on a known standard, plus a stated separation of coating thickness from remaining substrate. For a cathodic protection survey it is the reference electrode type and the explicit statement of which acceptance criterion was applied. Without those, the number is an opinion with a decimal point.

The scheduling module should require those artefacts at closure, not at audit. A work order that cannot be closed until the calibration record, the certification record and the raw data file are attached produces a different behaviour in the field than a work order that closes when the invoice is approved. It also gives the utility a defensible answer during a state inspection, a consent decree review or an insurance survey, because the evidence sits with the asset rather than in a folder named after a contractor who no longer holds the contract.

There is a second-order benefit that engineering teams notice within one cycle. When the raw data is attached rather than summarised, a later question can be answered without remobilising. Someone asks whether the low reading on a shell course was a genuine pit or a couplant artefact; if the A-scan or the raw dataset is on the record, an experienced reviewer can settle it in an afternoon. If only the PDF summary survives, the only way to answer is to drain the tank again.

Three grading scales, one asset register

A utility's condition data arrives on incompatible scales by design. NASSCO PACP assigns structural and operations grades from one to five against defined defect codes for a pipe segment. A tank report expresses condition in engineering units of remaining thickness against a design minimum. A cathodic protection survey expresses it as a potential in millivolts against a criterion. A coating survey expresses it as percentage breakdown and adhesion rating. None of these can be averaged with any of the others, and any system that tries has quietly destroyed the evidence.

The correct treatment is to store each in its native form against the asset and derive a governed condition index on top for planning purposes, with the derivation visible. A renewal programme needs one comparable number to rank candidates. An engineer defending a decision needs the underlying grade, the raw reading and the date. Both requirements are satisfiable at once, but only if the schedule and the asset register were built to keep the native record rather than to compress it on import.

This is also where asset naming does most of its damage. The CCTV contractor identifies a segment by upstream and downstream manhole. The GIS identifies it by a feature ID. The financial system identifies it by a capital asset number. Reconciling those three after the fact is a data project. Issuing the work order with the utility's own identifier as the mandatory key, and treating the contractor's identifier as an alias, removes the reconciliation entirely and costs nothing at the point of scheduling.

The return-to-service tail that turns a six-hour inspection into a twelve-day outage

The single most common scheduling error in potable water work is planning the inspection instead of the outage. Consider a routine steel tank washout. The tank has to be drained and the water routed or wasted under the discharge permit. The interior has to be ventilated and cleared for entry. The inspection itself may take six hours. If any coating repair is authorised, surface preparation, application and cure follow, and cure time is dictated by ambient temperature and humidity, not by the schedule. Then the facility must be disinfected under AWWA C652 and refilled, and bacteriological clearance samples must be taken and returned before the tank goes back into service.

Add those together and the inspection is a small fraction of the outage. Two days to drain and ventilate, one day to inspect, several days for coating work and cure, a day to disinfect and refill, and then the clearance sampling period governed by the state primacy rule. Twelve days is an ordinary figure. And there is a branch that most plans ignore: if a clearance sample fails, the last several days repeat. A schedule that shows a one-day task against a tank in July is not optimistic, it is wrong in a way that will be discovered by the distribution operator at the worst possible moment.

Modelling the tail properly changes decisions upstream. Once the true outage duration is on the calendar, the case for combining the interior inspection with the cathodic protection anode replacement, the mixing system service and the coating touch-up in the same entry becomes obvious, because the expensive part of the work is the outage rather than the labour. The scheduling module earns its place by making that bundling visible before the annual plan is committed.

Permit-required confined space is a scheduling constraint before it is a safety one

Digesters, wet wells, drained tanks, valve vaults and dry pits are permit-required confined spaces. Under 29 CFR 1910.146 an entry needs an entry supervisor, an attendant who does not enter, continuous or periodic atmospheric monitoring, and a rescue capability that can actually reach the entrant in time. On the wastewater side hydrogen sulphide and oxygen deficiency are the governing hazards; in a tank being recoated, solvent vapour is added to that list.

The scheduling consequence is that the inspection crew is not the crew. Booking two technicians against a digester entry and calling it resourced is how a window gets stood down at eight in the morning. The module has to model the entry team as a unit — supervisor, attendant, entrants, monitoring equipment, rescue standby, and the ventilation blower that has to run before anyone descends — and treat the whole unit as the resource whose availability constrains the date.

There is a knock-on effect on contractor data quality that most utilities discover late. The entry permit, the atmospheric readings and the gas-free certificate are the records a regulator or an insurer will ask for, and they are almost always held by the contractor rather than the utility. Requiring them as closure attachments on the same record as the inspection findings costs the contractor nothing at the time and saves the utility a document hunt years later, when the crew that performed the work has long since moved on.

How to evaluate the module against a low-bid, term-contract procurement model

Public procurement shapes everything about a utility's inspection programme. Work is awarded on term contracts or task-order agreements with fixed unit rates, often to the lowest responsive bidder, and rebid on a cycle measured in years. Any evaluation of a scheduling module that assumes a long-term relationship with one inspection vendor is evaluating the wrong thing. The realistic assumption is that the firm producing next cycle's data is not the firm that produced the baseline.

So the questions to ask a vendor are concrete. Can a deliverable specification be attached to a work order as a requirement rather than a note? Can closure be blocked until the required attachments exist? Can the utility's asset identifier be mandatory while a contractor alias is stored alongside it? Can seasonal availability windows sit on the asset and generate a conflict at plan time? Can a term contract's annual quantity be tracked against consumed task orders, so the programme knows in September whether it has spent its footage? Can the entry crew be modelled as a resource unit rather than a headcount?

Ask for the failure demonstration too, not the happy path. Show me a due date that lands outside the allowed window. Show me a contractor upload with the wrong units. Show me a work order somebody tried to close without a calibration record. A module that handles those three cleanly will hold a utility's programme together across a change of contractor. One that only demonstrates a clean cycle is being demonstrated in conditions the utility will never actually operate in.

What Atlantis configures for a water and wastewater programme

Our inspection management platform is configured around the utility's asset register rather than a generic equipment list, so tanks, mains, basins, digesters and pressure equipment each carry their own interval logic, allowed-window rules and required deliverable set. Work orders issue with the specification bound in: identifiers, units, method, technician qualification level, required attachments and the file format the contractor must return. Closure is gated on those attachments, so the utility's evidence is complete at the moment the invoice is approved rather than years later.

On the data side we map contractor formats on ingest instead of asking the contractor to change how they work. PACP-coded exports, thickness datasets, potential surveys and coating reports land in their native structure against the correct asset, with the derived condition index calculated on top and its derivation visible. Historic PDF-only cycles can be back-loaded as evidence so the trend line starts before the platform did, which matters when the first renewal decision arrives twelve months in.

The platform is built on Odoo, so the inspection programme sits alongside procurement, term-contract quantity tracking, contractor qualification records and the asset register rather than in a silo beside them. It is affordable, accessible and fully customisable to the way a specific utility runs its plan. If you want to see it against your own asset list and one real contractor deliverable, request a demonstration or a scoped consultation at info@atlantisndt.com.

Request a consultation