API 570 common mistakes
API 570 · Understand the traps and how to avoid them.
On a scope question, treating every pipe on the unit as API 570 piping - for example counting a cross-country crude line, boiler external piping, or instrument/control tubing as in-scope.
The exam tests the scope article's inclusion/exclusion logic, but candidates answer from the everyday meaning of the word 'piping' and never check the exclusions.
What to do instead
Before committing to an answer, ask 'is this item covered by another code or explicitly excluded (pipelines under B31.4/B31.8, boiler external piping, instrument and control tubing, pressure vessels, furnaces, etc.)?' Rehearse the exclusion list until the check is automatic.
Reading a short damage description and pattern-matching to the first familiar mechanism - labelling flow-accelerated/erosion-corrosion as plain general corrosion, or missing corrosion under insulation because 'the pipe looks painted and dry'.
Mechanism names share vocabulary and the exam gives only a line or two of process data, so candidates anchor on a keyword instead of the driving conditions.
What to do instead
For each scenario, extract three facts before choosing: fluid and velocity, temperature/insulation status, and chemistry or contaminants. Then localize: flow-driven thinning shows up at elbows, tees, reducers and downstream of control valves; CUI shows up where the susceptibility window and insulation/water-trapping conditions coincide.
Laying out thickness measurement locations only at the circuit's generic corrosion allowances and leaving injection points, mixing points, and deadlegs out of the plan.
Candidates think of a piping circuit as one uniform corrosion environment and treat these features as fittings rather than as locally severe, separately managed locations.
What to do instead
Independently list every injection point, mixing point (where two streams join), and deadleg in the circuit; give each its own TMLs and a shorter, separately justified inspection interval than the circuit as a whole.
Classifying work as a 'repair' just because welding was involved, or as an 'alteration' only when a new nozzle is added.
The distinction turns on the definition (change of pressure/temperature rating or service fluid, or a physical change in a component), not on the type of workmanship, and candidates classify by activity instead.
What to do instead
Run a two-part test every time: (1) Does it change the pressure rating, temperature rating, or service fluid? (2) Does it physically change a component beyond restoring it to its original condition? Either 'yes' makes it an alteration requiring engineering approval and possible re-rating; restoring original condition is a repair.
Computing remaining life with a rushed formula - subtracting from nominal instead of measured thickness, using only the long-term corrosion rate, or mixing mils per year with millimetres per year.
The arithmetic looks trivial, so candidates do it in their heads under time pressure and skip the short-term versus long-term comparison.
What to do instead
Write the equation before doing arithmetic: remaining life = (measured thickness − minimum required thickness) ÷ corrosion rate. Convert every reading into one unit system, calculate both a short-term rate (from the two most recent readings) and a long-term rate, apply the governing (higher) rate, and round the resulting interval down, not up.
Treating re-rating as a paperwork exercise - recalculating a maximum allowable working pressure at one temperature without re-deriving allowable stress, and assuming re-rating only ever means increasing pressure.
Candidates memorize 're-rating needs an engineer' but not that the calculation is temperature-dependent and bidirectional.
What to do instead
For any re-rate, state that it can be an increase or a decrease; re-derive allowable stress at the new design temperature, verify every component in the circuit (pipe, flanges, valves, fittings, supports), require engineer approval, and then update records and markings accordingly.
Setting an inspection interval from a half-remembered number, or extending an interval because the last inspection found nothing of note.
Candidates memorize class intervals as isolated facts and forget the remaining-life constraint that can shorten them and the justification requirement that governs any extension.
What to do instead
Apply the rules in order: the class-based maximum from the interval table, then the cap tied to estimated remaining life (whichever is shorter governs). A clean inspection history alone is not a basis for exceeding the table maximum; any extension needs documented engineering justification.
Assuming an external visual inspection satisfies corrosion-under-insulation coverage, or selecting an NDE method that does not fit the geometry, material, or access available at the location.
Candidates treat 'inspection' as a single activity and default to the cheapest or most familiar method rather than matching method to mechanism.
What to do instead
Select the method from the damage mechanism and the access: CUI requires insulation removal at representative and suspect locations, not just a walk-by; internal or under-insulation loss needs volumetric or profile techniques, and austenitic or difficult geometries may rule out a method that works on carbon steel.
Writing an inspection summary that states findings but omits required elements - measurement locations and readings, the evaluation of remaining life or fitness, and the authorized inspector's approval.
Candidates narrate what they saw and stop, because in practice the paperwork is filled in by others.
What to do instead
Check the finished answer against a four-part test: what was inspected, what was found (with TMLs and readings), what it means (rate, remaining life, suitability for continued service), and who approved it. If any part is missing, the report is incomplete.
Closing out a repair or alteration without revising the isometrics/drawings and circuit records, or discarding old thickness readings once a report is issued.
Candidates see documentation as the end of the job rather than as an input to the next corrosion-rate calculation.
What to do instead
Treat every repair or alteration as triggering a records update: as-built drawings and isometrics, circuit and TML records, and re-rating documentation, all retained with the equipment file. Keep historical readings - without them, no future short-term or long-term corrosion rate can be calculated.