API 936 common mistakes
API 936 · Understand the traps and how to avoid them.
Answers material-selection questions by citing API 936 as if it were a property specification - quoting density, strength, or service-temperature limits from 936 for a candidate castable.
The standard is an installation quality-control document covering sampling, mixing, curing, testing, and inspection; it deliberately does not set or tabulate design properties, so candidates memorise the test names and then assume the acceptance values come from the same document.
What to do instead
State the scope explicitly before answering: property values come from the owner/engineering specification and the manufacturer's data sheet; API 936 tells you what to inspect, sample, test, and record, and what documentation to require. When a stem asks 'what property must this material have', the correct source is the project specification, not 936.
Selects a material class by generic strength or popularity - a conventional high-cement castable for an insulating or thermally cycled service, a dense low-cement castable where abrasion is not the mechanism, or a phosphate-bonded material in a wet-steam environment - without matching the class to the dominant degradation mechanism.
Candidates learn families of monolithics (conventional, low-cement, ultra-low-cement, phosphate-bonded, insulating, gunning mixes) as a list rather than as trade-offs, so they pick the 'best' material instead of the one whose chemistry, density, and bond suit the damage mode.
What to do instead
Build a decision table keyed to the failure mechanism: thermal shock cycling favours low-cement, low-modulus, or fibre-toughened materials; abrasion favours hard, dense aggregate; insulating duty requires low-density material with matching strength; chemical attack requires checking aggregate and bond chemistry against the specific slag, alkali, or acid. Always check maximum service temperature against actual hot-face temperature plus margin.
Reviews a lining specification only for material grade and thickness, and never checks expansion-joint layout, anchor type/spacing/coverage, layer interfaces, or how the lining terminates at nozzles and shell transitions.
Design content feels like the engineer's job, so candidates skim it, and exam items that hinge on a missing expansion joint or on anchor spacing at a transition are answered from general refractory 'common sense' rather than from the specified detail.
What to do instead
Use a fixed design-review checklist for every design question: material and thickness, anchor alloy/type/spacing/coverage and weld quality, expansion joints (location, width, depth, and that they are not filled with castable), construction and cold joints, interface between different materials, and termination details at penetrations. Name the specific item the scenario violates instead of answering in generalities.
Treats expansion joints and construction joints as interchangeable, or accepts a design in which joints are omitted at shell openings, anchor transitions, or long runs, or are filled with the same castable during installation.
Both are gaps in the lining, so they get merged mentally, but one accommodates thermal growth and the other only reflects placement logistics; filling an expansion joint removes the single feature preventing buckling and spalling.
What to do instead
Define both before answering: expansion joints are designed gaps that absorb thermal expansion and must remain open (with the specified compressible filler or sacrificial material removed at dryout); construction joints are placement boundaries between lifts or panels. In scenarios, check that joint spacing/width matches the design and that placement did not fill or mortar them.
Allows or tolerates 'just a little more water' at the mixer to improve pumpability, trowelability, or to speed placement, without a measured consistency check against the approved water range.
Placement crews judge workability by eye and the mixer operator wants to keep the line moving; the strength, density, and shrinkage penalty is invisible on site, and candidates underestimate how sensitive monolithic properties are to water beyond the qualified range.
What to do instead
Fix the water at the qualified addition for the batch size, measure it (calibrated container or flow meter), and verify consistency with the method the manufacturer specifies rather than by feel. If workability is off, investigate mixing time, temperature, and aggregate moisture before touching water; record actual water added per batch on the installation record.
Approves placement over anchors that are rusty, oil-contaminated, or missing the specified coating/plastic cap, or over a shell that has not been cleaned and checked, because anchor and surface preparation are treated as the contractor's private business.
Preparation happens before the inspector's attention is drawn to the lining, and the work is hidden within minutes of placement; candidates also forget that uncoated anchors restrain the lining locally and that the specified anchor coating or cap is an expansion-relief feature, not cosmetic.
What to do instead
Make surface and anchor preparation an explicit hold point: verify cleanliness, anchor pattern and spacing against the drawing, weld integrity by visual and specified testing, anchor coating/cap type and full coverage, and that no anchor is bent to fit. Record the release before any material is placed, then re-check on a sample basis during placement.
Signs off an installation without defining or witnessing the inspection hold points - for example, letting the lining be closed over on a shell whose preparation, anchor welds, or first lift were never released.
Candidates read 'inspection' as testing specimens in a laboratory and forget the sequential, before-cover nature of lining inspection; once the next lift or shell section is placed, the earlier work cannot be verified without destructive removal.
What to do instead
List the hold points in order: material receipt and storage, shell and anchor preparation, mixing/water verification, placement of each lift and joint installation, curing, dryout, and final acceptance. For each, state who releases it, what evidence is recorded, and what happens if the hold point is missed - typically documented investigation and possible removal.
Handles sampling and specimen preparation loosely: samples taken at the wrong point in the batch or the wrong frequency, specimens cast and cured differently from the lining they represent, or a single specimen relied on for a test.
Sampling feels administrative, and the standard's requirements about when in the discharge to sample, how many specimens per sample, and how they are cured alongside or under the same conditions are exactly the details candidates skim.
What to do instead
Work from the sequence, not memory: sample during discharge at the specified interval, take the specified number of samples per lot or batch, cast the required specimen set for each test (density, cold crushing strength, modulus of rupture, and other specified tests), and cure specimens under the same conditions as the installed lining. Identify which specimen supports which acceptance criterion before testing.
Compares test results to acceptance criteria casually - accepting a result below the specified minimum because 'it's close', or condemning a lot on one anomalous value without checking sample identity, testing error, or the specified retest/rejection path.
The candidate knows a number was measured but not whose value it is (manufacturer's published typical versus project-specified minimum) or what the document requires when a result falls outside the limit.
What to do instead
Always answer in three steps: identify the acceptance limit and its source (project specification or manufacturer's qualified value), compare the properly identified result, then follow the specified disposition - investigate, retest from the retained sample or a new sample, or reject and remove. Never invent an intermediate 'engineering judgement' acceptance when the document defines one.
Collapses curing and dryout into one step - assuming the heat-up schedule performs the cure, or starting dryout before the specified cure period has elapsed, or using a fast heat-up without regard to steam escape and spalling risk.
Both involve time and temperature and both happen after placement, so they get merged; candidates also assume that a slow ramp is 'safe' regardless of thickness, moisture content, venting, or the material's permeability.
What to do instead
Separate them explicitly: curing holds water in the material to complete the hydraulic bond for the specified time and temperature; dryout removes free and chemically bound water through a controlled heating schedule that must respect lining thickness, vent paths, and the material's spalling tendency. Check both against the approved procedure, and require documented temperature records and hold periods at the specified plateaus.
Describes repairs as 'cut out and patch' without removing all deteriorated material to sound substrate, without roughening and cleaning the interface, without a compatible repair material and its own cure, or by welding/anchoring over hot or contaminated surfaces.
Repair seems like the practical, low-technical part of the exam, so answers stay at the level of intent; the specific steps that determine whether the patch survives the next thermal cycle are the ones omitted.
What to do instead
Answer with the full repair sequence: define the extent of removal (to sound, uncontaminated material, typically verified by sounding or inspection), prepare the surface and any added anchors, select a repair material compatible with the existing lining and its service, place, cure, and dry out the patch under the same discipline as new work, then record the repair and its acceptance. Note when the required action is replacement rather than repair.