Asset Integrity Management Software — Built Around the Inspection Data
Most AIM platforms started as historians, EAM suites or IoT platforms and had integrity bolted on. This one started with the UT reading.
Asset integrity management software exists to answer three questions: what condition is this equipment in, how fast is it changing, and where exactly is the damage. Answering them honestly requires measured inspection data at the component level — thickness readings tied to specific CMLs, indications tied to specific welds, damage mechanisms tied to specific process conditions. Platforms built outward from process historians or maintenance systems tend to model everything except that. Atlantis starts there.
The three questions the software has to answer
- Condition — what does measured wall thickness, indication history and damage-mechanism susceptibility say about this component today, at CML resolution rather than equipment-tag resolution.
- Trend — what corrosion rate does the measured thickness time series show per CML, rather than a generic default corrosion rate.
- Location — where on the physical asset the damage sits, and when the next inspection is due under API 510, API 570 or API 653.
How the data gets in and stays trustworthy
Inspection data arrives from UT thickness surveys, PAUT and TOFD scans, radiography, MFL and inline inspection runs, corrosion mapping, and visual inspection reports — in whatever format the instrument produced. Each reading is bound to a CML or TML with a persistent identity, so a 2019 reading and a 2026 reading on the same location are genuinely comparable and corrosion rates are computed from a real time series rather than from an assumed default.
Damage mechanisms are assigned per API RP 571 against the actual process service — sulfidation, naphthenic acid corrosion, HTHA per API 941, wet H2S cracking, chloride stress-corrosion cracking, corrosion under insulation, MIC, erosion-corrosion — so the inspection plan targets the mechanism that is actually credible for that circuit instead of applying a uniform scan coverage everywhere.
Every record carries provenance: which procedure revision was in force, which technician performed the work and what their certification state was at that moment, and whether the instrument was in calibration. That is the bundle a regulator, insurer or client audit asks for, and assembling it after the fact is the single most expensive routine task integrity teams face.
Thickness trends and damage mapping on measured data
An integrity decision is only as good as its condition input. When a corrosion rate is a default rather than a measurement, the inspection plan is a re-labelled calendar. Measured thickness trends per CML show which circuits are genuinely degrading and which have been proven stable for a decade.
Thickness grids, corrosion maps and indications are rendered on the 3D model against the component they belong to, colour-coded by condition. The inspection records behind each reading are retained, so the basis for any decision can be reconstructed years later.
Honest comparison with the platforms you are also evaluating
| Platform | Origin | Choose it when | What integrity teams find missing |
|---|---|---|---|
| AVEVA PI System / OSIsoft | Process historian | Time-series process data is the primary problem and you already run PI across the estate. | Inspection data at CML resolution, damage mapping and inspection-evidence provenance are not native. |
| Bentley iTwin | Engineering / BIM | Capital projects and as-built engineering models dominate the use case. | Damage mapping and thickness trending from measured NDT data. |
| Cognite Data Fusion | Industrial data platform | You have a large data-engineering team and want to contextualise many source systems. | Out-of-the-box integrity workflow — it is a platform to build on, not an integrity application. |
| IBM Maximo APM / Hexagon EAM | Maintenance & EAM | Work management and reliability across a large owned estate is the priority. | Thickness trending per CML, 3D damage mapping and inspection-technique provenance. |
| GE Vernova APM (Predix/Meridium lineage) | OEM asset performance | Rotating equipment and power-generation assets, especially GE-manufactured fleets. | Fixed-equipment corrosion management driven by inspection data rather than sensor telemetry. |
| Atlantis Digital Twin | NDT and inspection | Fixed equipment — vessels, piping, tanks, exchangers, structures — where the integrity case rests on inspection data. | Not the right tool for OEM rotating-machine telemetry models or enterprise-wide maintenance work management; integrate rather than replace. |
Deployment and data ownership
A first unit typically goes live in ten to fourteen weeks: geometry capture or import, CML register reconciliation, historical thickness import, damage-mechanism assignment, then dashboard and report configuration. Geometry can come from LiDAR or photogrammetry capture, from drone survey, or from existing BIM, CAD and isometrics — there is no requirement to re-scan a plant that already has good as-builts.
Data ownership is explicit. Full REST API, documented schema and bulk export are available throughout, and the export includes the inspection history and assessment records, not just a rendering. Integrity data has a multi-decade life; it should never be hostage to a platform decision made in one budget cycle.
Frequently Asked Questions
What is asset integrity management software?
It is the system that holds the technical case for continued safe operation of fixed equipment: condition data from inspection, the damage mechanisms credible for each circuit, and the resulting inspection plan and due dates under API 510, API 570 and API 653. Done properly it is the evidence base an inspector, regulator, insurer or client audit examines.
How is an asset integrity digital twin different from a 3D model?
A 3D model is geometry. A digital twin binds live condition data to that geometry: each CML carries its thickness history, each weld its indication history, each circuit its governing damage mechanisms. The distinction that matters in practice is that a twin changes when new inspection data arrives — corrosion rates recompute and the condition map updates — whereas a model does not.
How does it integrate with SAP, Maximo or our historian?
Through documented connectors and a full REST API. Typical patterns: inspection findings raise notifications and work orders in SAP PM, Oracle eAM, IBM Maximo or ServiceNow; process conditions are read from AVEVA PI or OSIsoft historians to inform damage-mechanism susceptibility; and engineering geometry is imported from Bentley iTwin, AVEVA or plain CAD and isometrics. The intent is to sit alongside those systems rather than displace them.
What data do we need before starting?
At minimum: an equipment and circuit list, a CML register (even an imperfect one), historical thickness readings in any tabular format, process service conditions per circuit, and current inspection procedures. Geometry is helpful but not blocking — many first deployments start from isometrics and P&IDs and add scan-derived geometry later. Reconciling the CML register is usually the longest single task and is worth starting before the platform decision is made.
How long until the first unit is live?
Ten to fourteen weeks for a first process unit is typical: two to three weeks of geometry capture or import, three to four weeks reconciling the CML register and importing historical thickness data, two to three weeks assigning damage mechanisms per API RP 571, then dashboard configuration and integrity-team training. Subsequent units are substantially faster because the data model and conventions are already established.
Is Atlantis affordable compared with the enterprise APM suites?
The platform is positioned as affordable, accessible and fully customizable, and the commercial model is deliberately simpler than enterprise APM licensing — which typically combines platform fees, per-asset or per-tag fees and a substantial systems-integrator engagement. Scope, asset count and integration depth drive the terms, so request a tailored quote. What we will commit to publicly is the data-ownership position: full export, documented schema, no lock-in.
See it running on your own workflow
Your job types and your reporting formats, walked through by an ASNT NDT Level III. Affordable, accessible, fully customizable — request a demo and a tailored quote.
Related: Atlantis Digital Twin platform · Digital twin vendor comparison · vs AVEVA PI System · vs GE Predix / Vernova APM · vs Bentley iTwin · Digital twin ROI calculator
OSHA's Process Safety Management standard requires a documented mechanical integrity program, and the audit trail an inspector actually checks is the closed loop from inspection finding to fitness-for-service or RBI disposition to verified deficiency correction. Asset integrity software supports that element by keeping those three records — finding, disposition, correction — linked to the same equipment record instead of scattered across separate systems.
29 CFR 1910.119(j) requires inspection and testing procedures that follow recognized and generally accepted good engineering practice, and it requires that deficiencies outside acceptable limits get corrected before further use, or in a safe and timely manner otherwise. In an audit, that translates into a specific evidence chain an inspector will pull: the inspection record that found the deficiency, the engineering evaluation — often an API 579 fitness-for-service assessment — that determined whether the equipment could keep running and under what conditions, and the closure record proving the deficiency was actually corrected, not just logged. When those three pieces live in three different systems, or worse, in email threads and spreadsheets, reconstructing that chain for an audit becomes a manual research project instead of a query.
Source: OSHA 29 CFR 1910.119 (Process Safety Management)
Mechanical integrity under OSHA PSM: closing the loop from finding to correction
The mechanical integrity element of a PSM program is judged less on whether inspections happened and more on whether findings that fell outside acceptable limits were tracked to actual resolution. A wall-thickness reading below the retirement limit, a crack indication exceeding acceptance criteria — these need a documented disposition path: a fitness-for-service evaluation determining whether the component can run as-is, needs a rerate, or requires repair, followed by a verified correction if repair or replacement was the outcome. Software support for this means the deficiency record doesn't close itself when the inspection report is filed; it stays open until the disposition and, where applicable, the corrective action are both documented against it.
The connection to RBI matters here too: a deficiency that gets corrected should also update the risk-based inspection interval for that circuit going forward, because the corrective action changes the actual condition the next interval calculation should be based on. Treating the PSM correction record and the RBI risk model as separate, unconnected systems means the interval schedule keeps running on outdated assumptions even after the underlying problem was fixed.
This is a narrower slice of what a full mechanical integrity program covers — inspector qualification, equipment-specific inspection frequency, and documentation retention are all part of 1910.119(j) as well. The dedicated treatment of mechanical integrity program requirements covers that fuller scope; the point here is specifically how the finding-to-disposition-to-correction chain has to work as a connected record, not a set of independent filings, for the audit trail to hold up.