UT Procedure Authority for Pasadena Refineries, Terminals and Yards

What a UT Level III signs, and what nobody else can sign for them

Three responsibilities get mixed up on almost every unit along the channel. The API 510, 570 or 653 inspector authorizes and signs the in-service inspection of the vessel, circuit or tank. The UT technician is certified under the employer's written practice, to SNT-TC-1A or CP-189, to operate the instrument and report what the screen shows. The ASNT Level III in UT sits behind both: approving the procedure the technician works to, approving the practice under which that technician was certified, and answering for both when a client asks who decided the technique was adequate.

The Level III's signature therefore lands on procedures, technique sheets, examination material, certification records and technical reviews. It does not belong on production reports as a routine approval, and it does not belong on a mechanical integrity document. Where those signatures blur, the audit question that follows is always the same and always awkward: which credential were you signing under.

The practical value of the role is that it forces decisions to be made before mobilization. Reference block, couplant, surface condition, gain and reference level, coverage, resolution, temperature correction, reporting threshold, acceptance criteria. Every one of those is a Level III decision, and every one of them is otherwise made informally on a scaffold by whoever is holding the probe.

The calibration block is a component decision, not a shelf item

Amplitude-based ultrasonic examination is a comparison. The indication is judged against a response from a known reflector in a block, so the block has to represent the part. ASME Section V Article 4 is explicit about the dimensions of that representation: material, thickness range, surface curvature where it matters, and heat treatment condition. Change any of them and the attenuation changes, the beam behaves differently, and the reference curve you built no longer describes the component you are examining.

In shops and service companies around Pasadena the failure is almost always economic rather than technical. One well-made carbon steel block gets used for everything because buying and maintaining blocks for stainless, duplex, clad and heavy-wall work is expensive and the difference is invisible on the screen. It is not invisible in a client's technical review, and it is not invisible in the sizing.

The related trap is transfer correction. The block is machined and clean; the component is scaled, painted or rough, and often curved. The difference in coupling and surface attenuation between the two is real and measurable, and a procedure that does not require it to be measured and applied is transferring an unknown error into every amplitude decision made under it.

DAC and DGS answer different questions

A distance amplitude correction curve is empirical. It is built from actual responses off side-drilled holes or notches at several sound paths in a block that matches the component, and it accounts for beam spread and attenuation as observed rather than as modeled. It is the default basis where a referencing code specifies a basic calibration block, and it is defensible precisely because it was measured on representative material.

A distance gain size diagram is theoretical. It relates amplitude to an equivalent reflector size for a specific probe using a modeled attenuation, calibrated from a single reference reflector. Used correctly it is fast and repeatable. Used carelessly it fails in exactly the places refinery work lives: coarse-grained austenitic weld metal, nickel-alloy filler, castings, and sound paths near the near-field limit where the amplitude relationship is not well behaved.

The Level III's job is to state which basis applies, under which code, for which material, with which corrections recorded, and to say plainly where the basis may not be used. A procedure that offers both without conditions is not a procedure. It is a menu.

The corrosion-rate arithmetic that manufactures and hides metal loss

This is the trap that costs Ship Channel operators the most and gets discussed the least. Thickness measurement has finite repeatability. Between two surveys performed by different technicians, on slightly different spots, with different couplant and surface preparation, a spread of a few thousandths of an inch is normal even when no metal has been lost. Now compute a corrosion rate by subtracting one reading from the other and dividing by the interval.

Over a twelve-month interval, a four-thousandths measurement difference becomes an apparent rate of four mils per year. On a wall with a hundred and fifty mils above minimum, that fabricates a remaining life of under forty years where none was lost, or in the other direction, cancels a real loss that happened to be measured optimistically the first time. Short intervals amplify the effect; the shorter the interval, the more the rate is measurement noise rather than corrosion.

The controls are unglamorous and effective. Fix and mark the measurement locations so the same spot is read each time. Require the same surface preparation. Report to a stated resolution rather than to whatever the instrument displays. Compute both long-term and short-term rates and require the divergence to be explained rather than averaged. And have the procedure say that a rate derived from two readings inside the repeatability band is not a rate at all. The Level III writes those rules; the integrity engineer inherits them.

Thin wall, doubled echoes and readings that are exactly twice wrong

Dual-element thickness probes are the workhorse of refinery and terminal survey work, and they carry a specific failure mode on badly thinned wall. As remaining thickness falls, the instrument can lock onto a second or third backwall echo instead of the first, and report a value that is a clean multiple of the true wall. The number looks plausible, the display looks stable, and the reading is the most dangerous kind of wrong: reassuring.

The counter is procedural rather than clever. Require A-scan verification rather than digital-only readout wherever remaining wall approaches the thin end of the probe's range. Specify a probe and delay line appropriate to the expected thickness rather than one probe for the whole job. Require any reading that is suspiciously close to double an adjacent reading to be re-taken and confirmed on the waveform. State a minimum reportable thickness for the technique.

There is a mirror-image error on the thick side: mode-converted or off-axis signals accepted as backwall on curved or heavily corroded internal surfaces, giving readings that are thin and trigger unnecessary intervention. Both errors originate in the same omission, which is a procedure that does not require the technician to look at the waveform.

Coverage is the rule that decides whether a survey means anything

Two technicians can take entirely valid readings on the same circuit and reach opposite conclusions about its condition, purely because the procedure never said where to read. Coverage is the least glamorous clause in a UT procedure and the one that most determines whether the data supports a decision. It has to state how measurement locations are selected, how many readings constitute a location, whether a grid or a point convention applies, and how locations are marked so the next survey reads the same steel.

The selection rule matters more on refinery piping than anywhere else, because corrosion in these systems is not uniform. Injection and mix points, dead legs, downstream of control valves, the bottom of horizontal runs, elbows on the extrados, and soil-to-air interfaces on buried-to-aboveground transitions all corrode at rates unrelated to the straight-run average. API 570 and API 574 are direct about this. A procedure that says take readings at four points on the circumference at each location, without saying which locations, has delegated the entire integrity question to whoever could reach the pipe.

The Level III owns the coverage rule and owns its limits. That includes stating what the survey does not cover, which is the honest half that usually goes missing. A thickness survey performed only at accessible elevations, or only outside insulation windows, is a valid survey of a subset, and the report should say so. Owners can plan around a stated gap. They cannot plan around one they were never told about.

Where Pasadena's ultrasonic hours actually are

The local demand profile is unusual only in its density. Refinery and chemical fixed equipment along the channel, including the refinery now under Chevron ownership at Pasadena and the large refining and chemical complex at Deer Park next door, generates continuous corrosion monitoring and turnaround weld examination. That is the largest single block of UT hours in the area, and almost all of it is thickness and angle beam weld work rather than exotic technique.

The terminal sector is the second block. The bulk liquid storage running from the turning basin down toward Bayport is inspected under API 653, and every floor scanning campaign generates ultrasonic prove-up of indications. Shell course thickness, repair weld examination and settlement-related work follow. Prove-up is where data disputes concentrate, because the scanning contractor and the prove-up technician frequently work to unrelated procedures.

The third block is marine and fabrication. Barge and vessel hull gauging for the fleets working the channel is thickness work with its own reporting conventions and its own operator qualification expectations from class societies. Spool and skid shops serving the plants examine B31.3 welds continuously. All three blocks buy UT from small and mid-size employers, and all three audit those employers before award.

What a UT certificate does not authorize

A UT Level II certificate authorizes the holder to perform and interpret ultrasonic examination within the scope the written practice grants, to the procedures the Level III approved, on the materials and configurations the practice and the procedure cover. It is not a license to select the technique, to deviate from the procedure, or to decide acceptance where the code assigns that decision elsewhere.

It is not an API inspector credential. It does not authorize signing an in-service inspection report, setting an inspection interval, or accepting a remaining life calculation. It is not a welding inspector credential and does not carry authority over weld acceptance under the code of construction beyond the ultrasonic acceptance criteria the procedure cites. It does not automatically extend to phased array, TOFD, encoded corrosion mapping, immersion testing, or examination of materials the practice never addressed.

The most useful thing a Level III does with all of that is put the limits on the certificate itself. A certificate that reads UT Level II, thickness measurement only, carbon and low-alloy steel is a stronger document than one that reads UT Level II and leaves the rest to be argued about later. It also protects the technician, which is a point rarely made and usually appreciated.

Procedure qualification, demonstration and the client specification layer

Beyond the code minimum sits the client specification layer, and on the Ship Channel it is where most real requirements live. Owners impose their own demonstration expectations, their own reporting formats, their own coverage rules and, increasingly, their own requirements for encoded data and archived waveforms. A procedure that satisfies ASME Section V and fails the owner's specification is still a failed procedure as far as the award is concerned.

Where a code route requires demonstration rather than assertion, that route has to be planned. Ultrasonic examination used in place of radiography under Section VIII Division 1 carries demonstration and documentation obligations that pull in Section V requirements and, in practice, a fair amount of preparation with representative flawed specimens. Deciding to take that route the week before a fabrication starts is not a plan.

Atlantis develops and approves UT procedures and technique sheets, builds and maintains the examination material behind UT certification, reviews ultrasonic data independently where a second opinion is needed, and represents the technical position when a client or third-party audit reaches those documents. To scope a review, contact info@atlantisndt.com with your current UT procedure set, your certified UT population by level and technique, and the client specifications you are being held to. Consultations and quotes are arranged on request.

Request a consultation