Chemicals: damage mechanisms across many small units

A refinery integrity engineer can name the damage mechanisms in a crude unit from memory. A chemical plant integrity engineer cannot, because the site may run forty small units, three of which changed product last year. The reactor that saw chlorides in one campaign sees caustic in the next. The stainless piping that was fine at the old temperature is now in the chloride stress corrosion cracking window. The program is not one large problem repeated; it is many small problems that differ. This paper is about what chemical plant integrity software has to do to keep up with that.

Why is the damage mechanism the organizing fact?

In a thinning-dominated service, the inspection plan is thickness monitoring and the record is a CML with readings. In a chemical plant, thinning is one mechanism among many, and for most of the others a thickness reading tells you nothing. Chloride stress corrosion cracking of austenitic stainless steel, caustic cracking of carbon steel, hydrogen-related damage, intergranular attack after sensitization, corrosion under insulation on cold and cyclic lines, erosion-corrosion at slurry bends, and thermal fatigue at mixing points each need their own inspection method, their own locations, and their own interval logic.

API 571 is the reference that connects each mechanism to susceptible materials, operating conditions, appearance, and inspection method. Using it well means the record must say, for each piece of equipment, which mechanisms are credible and why. That is what turns a list of equipment into an inspection plan.

In intelliSPEC damage mechanisms are first-class records, with thinning and cracking as the base families, attached to the equipment and component. That means "every item with a cracking mechanism in this unit" is a query, and the NDE scope for a shutdown can be built from it. The intelliINTEGRITY module is where the mechanism, rate, and risk records live.

How do you manage dozens of small corrosion loops?

A corrosion loop groups equipment and piping that share a process environment and are expected to degrade the same way. In a refinery, loops are large and few. In a chemical plant they are small and many, and the boundaries move when a unit changes campaign.

The record has to handle:

• Many loops per site, each with its own mechanism list and its own rate history, without forcing the team to duplicate configuration for each.
• Reassignment of equipment between loops when the process changes, with the history following the equipment.
• Loop-level review, where a rate spike at one CML prompts a look at every CML in the loop.
• Small equipment counts per loop, where a single unexpected reading is a large fraction of the evidence.

The asset hierarchy in intelliSPEC is a graph of nodes with roles rather than a fixed tree, so a loop can be represented as a grouping over components that may sit under different physical assets. CML points carry the full thickness basis (nominal, tMin and method, design pressure, material, allowable stress, joint efficiency, corrosion allowance) and three corrosion rates, and the corrosion summary reports short-term, long-term, and maximum rates with the maximum governing. Remaining life is (t minus tMin) divided by the governing rate, which is the arithmetic an engineer would do by hand.

How does API 580 RBI work with many mechanisms?

API 580 describes risk-based inspection as the product of a probability of failure and a consequence of failure, evaluated per item, used to rank equipment and focus inspection where risk is highest. API 581 supplies a quantitative methodology; many chemical sites use a semi-quantitative matrix because the data to run a full quantitative model on forty small units does not exist.

The mechanics that matter for the record:

• Probability is driven by the active damage mechanisms and how well the inspection history characterizes them. An item with a credible cracking mechanism and no recent crack detection carries higher probability than one with a good UT history and thinning only.
• Consequence is driven by inventory, toxicity, flammability, and location. A small reactor of a toxic intermediate can outrank a large tank of a benign product.
• The result is a rank, and the rank is only as good as the mechanism records behind it.
• The assessment must be revisited when the process changes, because the mechanisms change.

intelliSPEC implements API 580/581 semi-quantitative RBI scoring as probability by consequence with a rank from 1 to 5. The scoring is a calculator with a visible basis, and the inspection interval calculation caps at half the remaining life regardless of the rank. The point of the tool is not to replace the RBI team's judgment; it is to keep the scores next to the mechanism records that justify them.

What does the record need for cracking and other non-thinning mechanisms?

A thickness reading is a number. A cracking examination is a set of facts:

• The material and the environment that make the mechanism credible.
• The locations examined, ideally by weld or by zone, with drawings marked.
• The NDE method, technique, and sensitivity (a PT examination and a PAUT scan are not equivalent evidence).
• Indications found, with position, length, and depth where measured.
• The disposition: acceptable, monitor, repair, replace, and the engineering basis.
• The next examination interval and why.

The asset graph in intelliSPEC includes weld as a node role, so a crack examination can be recorded on the weld, not just the vessel. intelliNDT holds the NDE results. Drawing annotation supports marking the examined zones and automatic CML point detection on drawings reduces the setup work for new equipment. For corrosion under insulation, which is often the largest single inspection spend on a chemical site, the CUI wizard family covers visual, suspect area, insulation removal, NDE UT, advanced NDE, repairs, recoat, inspection test plan, and frequency assessment as connected steps on the same component.

How does the inspection program follow changing chemistry?

This is the question that separates chemical plants from every other industry. A campaign change can put a new mechanism in play on equipment that was stable for a decade. The program has to react in weeks, not in a vendor's release cycle.

What the software has to allow the integrity team to do without code:

• Add a mechanism to the credible list for a unit and have the inspection plan reflect it.
• Add or change fields on the inspection form when a new mechanism needs different data recorded.
• Change reference lists (mechanism types, NDE methods, disposition codes) as the program matures.
• Keep the old records readable after the forms change.

intelliSPEC is metadata-driven: the product is configured by entity definitions and workspace and form definitions, and tenants change forms and workflows through the form designer rather than through code. The second half is the one that protects the history: completed records pin the field configuration, locale, and reference options they were executed with, so an inspection recorded on last year's form renders exactly as executed after this year's revision. Visual inspection records for these programs run through intelliINSPECT.

What does a small-unit turnaround look like in the record?

Chemical plants shut units more often and for shorter periods than refineries, and a site may have several units down in the same month. The inspection scope per unit is small; the coordination problem across units is not.

The record should let the team:

• Build each unit's scope from due intervals, open findings, and mechanism-driven examinations.
• Track findings by equipment during the shutdown, with photos bound to the component.
• Record repairs and post-repair NDE against the finding.
• See progress by unit and by day while the shutdown is running.

intelliPROJECTS models the shutdown as an event with activities, crews, daily progress, delay events, gate reviews, and cost entries, and the findings land on equipment rather than on the event. Because the field app works fully offline with no time limit on the local store, inspection inside a reactor with no signal is captured the same way as inspection in the open. Field-level conflict records show what happened when two inspectors touched the same item.

How intelliSPEC handles it

The facts a chemical plant integrity team should weigh:

• Damage mechanisms are first-class records attached to equipment and components, with thinning and cracking as the base families.
• The asset hierarchy is a graph with roles down to the weld, so loops and welds are real objects.
• CML points carry 64 fields including the full tMin basis; the tMin engine routes to ASME VIII Div 1 for vessels and ASME B31.3 for piping and returns the design code and formula with the number.
• Corrosion summary reports short-term, long-term, and maximum rates, with the maximum governing, and remaining life computed from them.
• API 580/581 semi-quantitative RBI scoring ranks items 1 to 5; intervals are capped at half the remaining life.
• The full CUI wizard family connects visual, suspect area, insulation removal, NDE, repairs, recoat, ITP, and frequency assessment.
• Tenants change forms, mechanism lists, and workflows through the form designer, and completed records pin their configuration.
• Everything lands in a hash-chained audit trail per tenant.

The chemical solutions page describes the module mix for a multi-unit site. The evaluation that matters is to bring one unit's mechanism list and last shutdown package and see whether the next shutdown's NDE scope can be built from the record.