How these are written. Each case below is drawn from real MTIS project work and has been anonymised before publication. No client, operator, site, plant, project, asset or tag identifier appears anywhere on this page, and no project photographs are published. Nothing here is embellished: no outcome figures, cost savings, percentages or customer statements are claimed, because those were not what the evidence established. Where an investigation stopped short of root cause, the case says so.
Case lesson 1 — What is actually blocking a water-injection system?
A water-injection train on an oil and gas production facility was accumulating solids on its filters. The question was not academic: if the solids are formation sand, the answer lies in completions and filtration; if they are scale, in water chemistry; if they are corrosion product, the plant is consuming itself and the injection system has a materials and corrosion-control problem. The three answers lead to three different budgets.
- Evidence
- Eight used filter membranes taken from points across the injection train — feeder, manifold, pump outlet, storage tank and individual injection points. Visually there was very little gross debris; the only naked-eye clue was that the centres of the membranes were more strongly coloured than their edges.
- Testing
- Sub-samples were cut under clean conditions — powder-free nitrile gloves, sterilised scissors and tweezers — and mounted on aluminium stubs. Examination was by environmental scanning electron microscopy, deliberately chosen so the deposits did not have to be sputter-coated: coating a deposit alters exactly the surface you are trying to analyse. Composition was by EDX, area scans on the deposit layer and point scans on individual particles standing proud of it. Carbon, oxygen and fluorine were excluded from quantification because the filter membrane itself is rich in all three and masks everything else.
- Mechanism
- The membranes were not loaded with sand. They were covered by a continuous layer-like deposit that was predominantly iron corrosion product — iron oxide and/or iron carbonate, which EDX cannot separate — carrying significant iron sulphide. Sodium and chlorine were present throughout, consistent with dried brine. Discrete alumino-silicate particles were also present, consistent with a minor fraction of formation fines. The iron sulphide is the significant finding: sulphide corrosion product indicates sour service.
- Root cause
- Not established, and deliberately not asserted. No operating history, water chemistry or materials data was supplied with the samples, so the report stated that limitation instead of filling the gap with a plausible story.
- Action
- The solids were reclassified as internally generated corrosion product rather than carried-over sand, which moves the follow-up work to corrosion control and souring monitoring, and the recommendation was to combine the characterisation with the water-chemistry and operating data that the laboratory did not hold.
The lesson. Characterisation identifies the mechanism; it cannot on its own identify the root cause. A laboratory given samples and nothing else can tell you truthfully what the deposit is made of, and should tell you just as truthfully what it cannot conclude. Characterisation for non-specialists →
Case lesson 2 — An unknown solid, and the preparation step that decided the result
A quantity of fine solids recovered from an oil and gas production stream arrived for identification with no history attached. The client wanted a compound identification, not an element list: knowing there is calcium present does not tell an engineer whether the deposit is carbonate scale, and the difference determines whether the fix is a scale-control programme or something else entirely.
- Evidence
- A single solids sample, visibly contaminated with hydrocarbon, supplied with no operating or process background.
- Testing
- The sample was dried slowly at 90 °C for 48 hours to drive off the bulk of the hydrocarbon before analysis — not housekeeping, but a controlling step, because organic material and water suppress the signal-to-noise ratio in X-ray diffraction badly enough to make phase identification unreliable. Composition was then measured by XRF, and the phases present were identified by XRD.
- Mechanism
- XRD gave strong evidence for calcium carbonate and for sodium chloride. The material behaves as carbonate scale plus dried brine salt rather than as an iron corrosion product or a predominantly siliceous sand.
- Root cause
- Outside the agreed scope. The engagement was a characterisation, not an investigation, and the report was written accordingly — the reported limitation was explicit: with no background information supplied, no literature search could be run to corroborate or challenge the interpretation.
- Action
- The finding directs the next step towards water chemistry and scale control, and away from a sand-management or metallurgical corrosion investigation that the evidence does not support.
The lesson. Elemental techniques such as XRF and EDX tell you which elements are present; they are poor at telling you which compounds they form, and their accuracy for light elements is limited. Phase identification needs a phase-sensitive technique. And sample preparation is not a formality — here it determined whether the technique could produce an answer at all. How we choose the test →
Case lesson 3 — Micro-moulded polymer gears failing on form accuracy
Not every problem is a broken component. A micro-moulding programme producing polymer microgears was losing yield to poor form accuracy, and the decision at stake was where to spend: re-cut the tool, change the polymer, or move the process window. Parts of this size are too small for the usual diagnosis — cavity temperature and pressure are the standard indicators in injection moulding, and a micro-cavity has no room for the instrumentation.
- Evidence
- Moulded gears produced across a designed range of process settings, together with part weight and gear tip diameter measurements, and process data from a monitoring system built for the purpose because standard in-cavity measurement was not practical at this scale.
- Testing
- Detailed examination of the moulded gear teeth by low-voltage field-emission scanning electron microscopy combined with elemental analysis, run against a statistically designed parameter set: metering size, melt temperature, hold pressure and hold time, and mould temperature.
- Mechanism
- Incomplete filling. Micro features sit at or near the end of the flow path and are the last part of the cavity to fill, so they are the first to suffer when packing is marginal. Gears moulded at low hold pressure and time with a low metering size showed clear short-fill defects at the tips. Part weight correlated strongly with gear tip diameter — low weight, small tips.
- Root cause
- The process window, not the tool and not the material. The parameter set in use was below the threshold needed to pack the micro features to form.
- Action
- Moulding moved to the optimised parameter set, which produced gears of higher form accuracy. Part weight — cheap, fast and non-destructive — was available as a practical proxy indicator for fill quality.
The lesson. Neither approach would have settled this alone. Process monitoring without characterisation shows that settings changed but not what changed in the part; characterisation without process data shows a defect but not which knob controls it. Coupling the two, with designed experiments, reaches answers a single technique cannot.
What these cases have in common
Each one follows the same spine: evidence → testing → mechanism → root cause → action, and each one is explicit about the point at which the available evidence ran out. That last part matters most. An investigation that names a root cause it cannot support is more dangerous than one that says plainly what further information is needed, because the plant then acts on it.
How MTIS investigates failures → · Root-cause analysis in materials failures → · What to preserve after a failure →
Further anonymised case lessons are added as candidates clear technical and confidentiality review. Insights carries fifteen mechanism-level technical notes in the meantime.