UT Level III Authority for Deer Park's Ship Channel Plants and Terminals

What UT Level III authority actually means

The phrase "Level III" gets used loosely along the Ship Channel, usually to mean the person who signs things. Inside a functioning ultrasonic programme it means something narrower and considerably more useful. The ultrasonic Level III is the technical authority for one method inside one employer's written practice. They establish the UT training outlines, write or approve the general, specific and practical examinations, grade or accept the practical performance, approve every UT procedure and technique sheet the employer issues, and fix the reference level and calibration scheme those procedures depend on.

That is a document authority rather than a field-presence authority, and the distinction matters when you are buying it. A Level III does not need to be standing on the scaffold to be accountable for what happens there. They are accountable for whether the instruction in the technician's hand can detect the flaw the circuit is actually growing, and for whether the certificate in that technician's file was earned against examinations that reflect this employer's work rather than a generic question bank.

The second half of the role is defensive, and it is the half most owners underestimate. When a client's inspection group or a third-party auditor challenges a technical decision — why sixty degrees and not forty-five, why this basic calibration block, why a notch rather than a side-drilled hole, why the coverage in this scan plan is sufficient given the nozzle geometry — somebody has to answer in code language, on the spot, without conceding ground that was never lost. Nobody below Level III is positioned to do that.

What a UT Level III certificate does not cover

A Level III certificate is method-specific. Ultrasonics is one method; radiography, magnetic particle, penetrant, eddy current and visual are separate methods with separate examinations and separate certificates. A UT Level III cannot approve a radiographic procedure, cannot write the RT sections of a written practice, and cannot sign an RT certification file. The habit of engaging one outside Level III and applying that signature across a four-method scope is common on the Ship Channel and it is the fastest way to have an entire certification population challenged.

Within ultrasonics, the certificate also does not automatically extend to every technique. Phased array, time-of-flight diffraction, corrosion mapping, immersion and the advanced techniques used for high-temperature hydrogen attack are techniques of the UT method that carry their own training, examination and demonstration expectations under SNT-TC-1A and CP-189. A procedure for encoded phased array approved by someone whose own file shows only conventional pulse-echo is a finding waiting to be written. It is a fair and specific question to ask any Level III before engaging them: which techniques does your own certification file evidence, and on what specimens.

Finally, the role has a boundary at the process unit. Atlantis supplies NDT technical authority. It is not a process safety management auditor and does not act as the API inspector of record. The API 510, 570 or 653 inspector authorises and signs the in-service inspection; the NDT technician performs examinations under certification issued through the employer's written practice; the Level III approves the procedures those technicians work to and the practice itself. Conflating those three roles is the most consistent source of confusion in this market, and getting the boundary right on paper protects everyone in the chain.

What the Deer Park industrial base does to a UT scope

Deer Park sits at the eastern end of the Houston Ship Channel and its industrial base is unusually concentrated. The Highway 225 corridor carries a refinery and a large petrochemical complex, the Shell Chemical site, the Intercontinental Terminals tank farm, and a dense cluster of bulk liquid terminals with their own barge docks, manifolds and transfer lines. Around them sit the fabrication yards, machine shops and inspection contractors of Pasadena and La Porte that service the whole corridor. The 2019 terminal fire put storage mechanical integrity here under a spotlight that has not moved off it.

That mix produces a very particular ultrasonic workload. Fired heater and cracking furnace tubes need wall measurement and creep-related assessment on surfaces that are oxidised and hot. Hydrogen-service circuits need consideration of high-temperature hydrogen attack, which a routine thickness survey will not find. Insulated lines carry corrosion under insulation that only becomes visible when someone strips a section and shoots it. Aboveground tanks need shell course thickness, critical-zone work and floor screening. Dock and jetty steel needs structural weld examination in a marine splash zone. Very little of that is generic plate work.

Then there is the labour pattern. Turnaround season pulls transient crews through every gate on the corridor, and the technician on your scaffold in March may have worked for three other contractors in the preceding year, each with its own written practice and its own view of what a Level II practical looks like. Certificates travel; the qualification behind them does not always. A Level III engaged for Deer Park work spends real time on incoming certification review, because the alternative is discovering the gap when a client auditor pulls the file.

Calibration: DAC, DGS and the reference level nobody defends

A distance amplitude correction curve is built by recording the response from a series of side-drilled holes at increasing metal path in a basic calibration block of the same material, product form, heat treatment and curvature as the component. It compensates for attenuation and beam spread with distance, and it is only meaningful if the block genuinely represents the part. A DAC drawn on a flat plate block and applied to a six-inch schedule 80 elbow is not compensating for anything real.

Distance gain size sizing is a different animal and carries a different trap. DGS diagrams are specific to a probe's beam characteristics — effective element size, near field, frequency. Swapping in a nominally identical probe from another manufacturer, or continuing to use a probe whose wedge has worn, silently invalidates the setup. Procedures that permit DGS should name the probe by type and require verification of the reference before each use, and the Level III should have decided which sizing basis applies to which examination rather than leaving it to the technician on the day.

Transfer correction is where audits land most often. Calibration happens on a machined block; examination happens on a coated, scaled, weathered or curved component with different attenuation and different coupling. Transfer correction measures the difference and shifts the reference accordingly. It disappears from turnaround work because it costs time and its absence produces no visible symptom. The symptom is invisible by definition: a sensitivity shortfall that keeps real indications below reference so they are never reported. A procedure that requires transfer correction and reports that never record one is a documented practice not being followed, which is a worse finding than never having required it.

The temperature arithmetic on hot Deer Park piping

Ultrasonic thickness is a computation, not a measurement: the instrument multiplies a calibrated velocity by half the measured time of flight. Sound velocity in carbon steel falls as temperature rises, by roughly one percent per hundred degrees Fahrenheit. If the instrument is calibrated on a cold block and the reading is taken on a line at six hundred degrees, the transit time is longer while the velocity constant is unchanged, so the computed wall comes out around six percent thicker than the metal actually is. The error runs in the unsafe direction, which is why it deserves a paragraph in every hot-service procedure.

The consequence compounds in the corrosion rate. API 570 remaining life is driven by the difference between two thickness measurements over the interval between them. If the previous reading was taken cold and corrected, and this one is taken hot and uncorrected, the arithmetic can produce a corrosion rate near zero or even negative, and the resulting remaining life is nonsense that nonetheless gets entered into the data management system and defended for years. Inspection managers usually discover this when a line springs a leak that the file said had two decades left.

There is a physical dimension too. Standard couplants boil off; standard wedges soften and change velocity; contact dwell time heats the delay line and drifts the reading during the measurement itself. A hot-service UT procedure has to specify the couplant by temperature rating, the probe and delay line construction, the maximum contact time per reading, the cool-down between readings, and the correction factor or the requirement to apply a hot-calibrated velocity. Writing "apply temperature correction as required" and leaving it there is not a procedure; it is an aspiration.

Austenitic welds, cladding and hydrogen service

Coarse columnar grain structure in austenitic stainless and nickel alloy weld metal scatters, attenuates and redirects a shear wave. A conventional forty-five or sixty degree contact examination developed on ferritic pipe will produce noise, skewed beam paths and false depth. Dissimilar metal welds are worse because the velocity changes across the fusion line. The correct answer is usually a low-frequency dual-matrix transmit-receive longitudinal arrangement with a technique demonstrated on a mock-up of the same weld configuration — and the wrong answer is a procedure that says nothing about it and is quietly used anyway.

Clad and weld overlay surfaces raise a different problem: disbond detection at the clad-to-base interface needs a technique aimed at that interface, not a thickness survey that reports total wall and misses a separated layer. Where overlay has been applied as a corrosion barrier in a Deer Park unit, the inspection question is generally interface integrity and base metal condition beneath, and both need to be named as separate objectives with separate techniques.

High-temperature hydrogen attack in hydrogen-service equipment is the case where a normal thickness programme is actively misleading. HTHA is internal decarburisation and fissuring; the wall thickness may be intact until very late. Detection relies on advanced ultrasonic approaches — backscatter, velocity ratio, full matrix capture and total focusing method imaging — applied by technicians demonstrated on representative damaged specimens, and interpreted against API 941 operating limits. A Level III's most valuable contribution here is often to state plainly that the existing procedure does not cover HTHA, rather than to let a thickness survey stand in for an assessment it cannot make.

Terminals, tank floors and the API 653 arithmetic

The terminal population along the Ship Channel makes tank work a permanent part of the Deer Park scope. A typical out-of-service inspection combines magnetic flux leakage floor screening with ultrasonic prove-up of every indication, shell course thickness on a defined grid, critical zone examination at the shell-to-annular junction, and weld examination where repairs are planned. The UT procedure has to define prove-up: what constitutes confirmation of an MFL indication, what probe and technique are used on a pitted underside, and how a reading on a rough surface is qualified.

The recurring arithmetic trap is the initial thickness used in the corrosion rate. API 653 remaining life depends on the difference between an original and a current thickness, and it is routine to see the nominal plate thickness from the tank drawing used as the original figure. Plate is supplied to a mill tolerance and is frequently thinner than nominal from day one; using nominal manufactures corrosion that never happened, or, when the plate ran thick, hides corrosion that did. Where a genuine baseline exists it must be used, and where it does not the assumption has to be stated in the report rather than buried in a spreadsheet column.

The second trap is averaging. A shell course thickness reported as a single averaged number across a grid can sit comfortably above minimum while an individual location is already below it. Localised pitting and underside floor corrosion do not obey averages. The procedure should specify how readings are recorded, what the minimum recorded value triggers, and that the governing figure for a locally thinned area is evaluated on its own terms rather than smoothed into a course average that reassures nobody who understands it.

Records, traceability and the process safety audit

Deer Park is wall-to-wall covered process. OSHA's process safety management standard requires mechanical integrity inspection and testing on equipment in covered service, performed to recognised and generally accepted good engineering practice, with records retained. In practice the API codes supply the RAGAGEP and the NDT programme supplies the data. That means an examination record is not a technical artefact only; it is a compliance artefact, and it is read as one.

The chain an auditor follows is short and unforgiving. The report cites a procedure and revision; the procedure carries an approval signature; the signature belongs to a Level III whose certification in that method must be current and evidenced; the technician who performed the work must hold a certificate issued under a written practice whose UT content that same Level III approved; and the technician's file must show the experience, examinations and vision records the practice requires. A break anywhere in that chain devalues everything downstream of it.

The failure mode is almost never dramatic. It is a procedure revised in the office while field reports keep citing the superseded revision. It is a technique sheet updated without a corresponding revision to the parent procedure. It is a certificate signed by a person who is not the certifying agent the practice names. It is a Level III whose own certification lapsed eight months ago and whose signature has continued to appear. Remediating these is unglamorous, entirely documentary, and far cheaper before an audit than after one.

How a Deer Park engagement is structured

Atlantis works a mobilised model: document authority carried out wherever the work is best done, combined with scheduled site attendance for practical examinations, technique demonstrations, mock-up work and audit representation. Houston is the operating base, which puts Deer Park inside a short drive rather than a flight, and that proximity matters most during turnaround season when a practical examination or a technique demonstration cannot wait a week.

A typical sequence starts with a gap review: the existing written practice's ultrasonic content, the live procedure set, a sample of certification files, and a sample of recent reports read against the procedures they cite. That produces a findings list ranked by what a client auditor would raise first. From there the work is usually some combination of procedure rewriting and qualification, examination bank development against your actual codes and procedures, practical specimen preparation, certification file remediation, and standing Level III of record coverage thereafter.

Audit representation is a distinct deliverable and worth naming separately. When a client's technical authority challenges a technique, a coverage decision or an acceptance call, having the person who approved the procedure answer the challenge directly changes the outcome. For a consultation on your ultrasonic programme, or a scoped quote covering procedure development, examination and certification support, contact info@atlantisndt.com.

Request a consultation