Ultrasonic authority for the Calcasieu corridor: procedures, calibration and examinations
A UT Level III approves the ultrasonic procedures your technicians work to, fixes the calibration and sizing rules inside them, prepares and administers the practical examination, and rules on indications the procedure did not anticipate. Around Lake Charles that spans carbon steel pressure work, cryogenic nickel steel, austenitic and duplex welds and tank floors, under four different referencing codes.
Ultrasonics is the method most often bought as a commodity and the method where a bad procedure does the most quiet damage. Radiography either produces a readable film or it does not. Ultrasonics produces a number every time, whether or not the calibration was valid, whether or not the beam reached the volume of interest, and whether or not the material was capable of returning a usable signal at all. That number then travels into a corrosion rate, a remaining life calculation and an inspection interval, and nobody revisits it. Level III authority in this method is mostly about deciding, in writing and in advance, which numbers you are allowed to believe. That means reference reflectors, transfer correction, sizing rules, scan plans and coverage, and an honest list of the materials on your site where conventional angle beam should never have been specified.
Source: Written against ASME BPVC Section V Articles 4 and 5; ASME Section VIII Division 1; ASNT SNT-TC-1A (2020) and ANSI/ASNT CP-189; API 510, 570 and 653 with API RP 577 for examination practice; API RP 571 for the damage mechanisms; AWS D1.1 for structural work; and ASTM E797 for manual pulse-echo thickness measurement.
| Material or component | Where it is found locally | Why standard angle beam underperforms | What the procedure must specify |
|---|---|---|---|
| 9% nickel steel LNG tank and cryogenic welds | LNG export trains and storage on the ship channel | Nickel-alloy weld metal is coarse, attenuative and anisotropic, and the base metal is ferromagnetic, so amplitude drops and the beam skews | Low frequency, transmit-receive longitudinal probes, a reference block cut from the same weld procedure, and a documented transfer loss |
| Austenitic stainless and nickel alloy welds | Chlor-alkali, cold boxes, caustic and acid service piping | Columnar grain structure splits and steers the beam; a shear wave technique calibrated on carbon steel is simply invalid | Dual matrix or TRL longitudinal wave, velocity established on the actual weld, and geometry-versus-flaw discrimination rules |
| Centrifugally cast HP-modified furnace tubes | Olefins cracking heaters and reformer furnaces | As-cast coarse grain scatters the beam so severely that useful pulse-echo inspection is not available | Do not specify ultrasonics for wall loss or creep; route to eddy current, profilometry or radiographic techniques and say so in the procedure |
| Duplex and super duplex | Seawater and brine service, some vinyls units | Two-phase microstructure attenuates and skews, and sigma phase embrittlement changes the acoustic response | Frequency and probe selection proven on production welds, plus separate acceptance treatment for the flaw types actually expected |
| Hot carbon steel piping in service | Refinery hot lines, crude and vacuum transfer, steam | Sound velocity falls as temperature rises, so uncorrected readings over-report remaining wall | High temperature couplant and delay line, a stated temperature compensation rule, and a maximum surface temperature for the equipment used |
| Tank floor plate and shell courses | Bulk terminals, refinery tankage, dock storage | Magnetic flux leakage screening loses sensitivity under thick coatings and heavy plate, and spot readings miss isolated pitting | Screening technique with its stated limits, a prove-up rule for indications, and a defined prove-up ultrasonic technique with resolution suited to pitting |
What ultrasonic authority means, line by line
In the ultrasonic method the Level III is not a senior technician with a bigger badge. The role is defined by the documents that carry a signature. Procedures and their revisions, because a referencing code will not accept an examination performed to an unapproved procedure. Calibration and reference standard decisions, because the reference reflector determines what counts as recordable. Practical examination specimens, because that is how the employer establishes that a technician can actually find what the procedure requires them to find.
The fourth item is the one people forget: interpretation of the indications a procedure did not anticipate. Every plant has them. A cluster of mid-wall reflectors in a plate that turns out to be lamination rather than cracking. A geometric response from a counterbore that a new technician calls as root lack of fusion. A cluster of small amplitude indications in a weld that a scan plan never intended to insonify. Those calls are made under time pressure, in front of a client, and they need someone whose name is on the procedure.
Locally that authority has to be exercised across an unusually wide material range. A single week can cover refinery hot piping, a chlor-alkali nickel alloy line, a fabrication shop carbon steel nozzle weld and a cryogenic weld on the ship channel. The competent answer differs in each case, and in one of them the answer is that ultrasonics is not the right method.
What UT certification covers, and four things it does not
Certification in the ultrasonic method establishes that a person has the training, experience and demonstrated skill to set up equipment, calibrate to a reference, scan to a procedure and evaluate indications against stated acceptance criteria. It is granted by an employer under that employer written practice, and it is specific to the method and to that employer. It is a real qualification and it is worth something.
It does not, by itself, cover encoded phased array or time of flight diffraction unless the practice says so and the examinations proved it. It does not make the holder an API 510, 570 or 653 inspector; those are separate individual certifications that authorise in-service inspection decisions. It does not qualify anybody to perform a fitness for service assessment under API 579, which is engineering work built on top of the measurements rather than an extension of them.
And it does not confer the authority to deviate. A certified Level II who decides on the day that a different angle would work better, or that a corroded surface is close enough, has stepped outside the approved procedure. Whether that call was technically sound or not, the examination record now says something the procedure does not support. The Level III exists so that deviations become approved revisions instead of silent ones.
DAC, DGS and the transfer correction nobody wants to do
A distance amplitude correction curve is built empirically from a reference reflector, usually side-drilled holes, in a block that matches the component in material, thickness range and curvature. It is honest about attenuation because the attenuation is baked into the measurements. Its weakness is that the block has to genuinely match, and blocks migrate: the block cut for a shop job in one alloy ends up in the field van and gets used on something else entirely.
A distance gain size curve is analytical. It relates echo amplitude to an equivalent flat bottom hole diameter using the beam characteristics of one specific probe. It is efficient, it removes the need to carry a block for every thickness, and it becomes worthless the moment the material stops behaving predictably. On coarse-grained castings, austenitic weld metal or heavily attenuating nickel alloys the underlying assumptions collapse. It is also worth remembering that an equivalent flat bottom hole size is not a flaw size and should never be reported as one.
Both approaches depend on transfer correction, which is where field practice usually fails. Sensitivity is established on a clean machined block and then applied to a rough, painted or corroded component that absorbs and scatters more sound. Unless the loss is measured and added back, the examination is running below its intended sensitivity and reporting clean. The procedure should require the measurement, state the method, and require the value to appear on the report.
Angle beam arithmetic that decides whether the volume was covered
The geometry is simple and constantly got wrong. For a refracted angle in a plate of thickness T, the sound path to the first leg is T divided by the cosine of the angle, and the surface distance to the first full skip is twice T multiplied by the tangent of the angle. On twenty five millimetre plate a sixty degree probe needs about eighty seven millimetres of clear scanning surface to reach the far toe on a full skip. If a support, a clamp or a nozzle reinforcement sits inside that distance, the far side of the weld was never insonified regardless of what the report says.
Refracted angle also moves. It is set by the velocity ratio between the wedge and the material, and both are temperature dependent. A wedge left on a warm line softens and its velocity changes, so the nominal sixty degrees becomes something else, and the calculated flaw depth follows it. Procedures for hot service should state a maximum surface temperature, require angle verification at working temperature, and specify the couplant that will survive it.
Coverage is the deliverable, not the angle. A scan plan that shows the beam sweeping the root, the fusion faces and the cap for the stated thickness range is worth more than a list of probe angles, and it is the artefact an auditor can actually check. When a procedure covers a wide thickness range, the plan needs to hold at the extremes of that range, not just in the middle where the example was drawn.
Thickness readings, pitting and the coverage illusion
Manual pulse-echo thickness measurement is treated as the simplest thing in ultrasonics and it produces more bad data than any other technique. Doubling on thin wall, mode conversion in corroded backwall, coupling through a rust scab, reading through a coating the gauge was never set up to subtract, and probe placement drift between surveys all produce plausible numbers that are wrong in a consistent direction.
The statistical trap is worse. Take a twenty foot spool with twelve spot readings. The probe face covers a fraction of a square inch each time, so the survey physically interrogates a few square inches of a surface measured in thousands. Localised pitting, under-deposit attack and corrosion under insulation do not distribute themselves conveniently. Reporting that spool as having a minimum wall based on twelve spots is a statement about twelve spots, and the procedure should say so.
The fix is not more spots, it is the right technique for the damage mechanism. Where API RP 571 says the expected morphology is localised, the procedure should call for scanning or corrosion mapping over a defined area rather than a point grid, and should locate that area where the mechanism actually attacks: the extrados of a bend downstream of a control valve, the low point of a dead leg, the crevice under a support shoe, or the band under damaged insulation jacketing.
Tanks, docks and hulls on the Calcasieu Ship Channel
Storage tankage in this corridor is examined under API 653, and the ultrasonic component of that work carries its own decisions. Floor scanning by magnetic flux leakage is a screening technique whose sensitivity depends on plate thickness, coating thickness and scanning speed. It does not measure anything; it flags. The procedure has to state the prove-up threshold, the prove-up technique, and how top side and underside loss are distinguished, because a repair decision that treats them as equivalent will be wrong.
Shell course readings, critical zone measurements and the settlement survey interact. A shell thickness that satisfies the minimum for its course can still be inadequate once out of plane settlement and edge settlement are considered, and the ultrasonic data feeds directly into that assessment. Anyone specifying the ultrasonic scope should know which calculation their numbers are about to enter.
Marine work adds another layer. Hull and barge gauging along the ship channel and at the port is performed to class society and regulatory requirements, with their own expectations about who may take the readings, how the survey grid is laid out and how the data is reported. It is ultrasonic thickness measurement, but the acceptance framework is not API, and applying refinery habits to a hull survey produces a report the surveyor will reject.
The practical examination is where a UT programme is proved
Written examinations are easy to administer and easy to pass. The practical is the part that separates a real certification programme from a paperwork exercise, and in ultrasonics it is unusually informative because the candidate has to produce a calibration, a scan, and an evaluation that can all be checked independently.
A defensible practical uses specimens representative of the work: the product forms, thickness ranges and weld configurations the candidate will actually meet, with documented discontinuities whose type, size and location are recorded and controlled. It scores calibration setup as well as detection, because a candidate who finds the flaw with the wrong sensitivity has not demonstrated the skill. It records what was missed as well as what was found, and it is retained as evidence.
This is also the point at which a written practice reveals whether it was written for this site. If the specimens are all carbon steel plate butt welds but the crews spend half their time on austenitic piping and cryogenic welds, the examination proves the wrong thing. Aligning the practical set with the site material inventory is one of the highest value changes available to most programmes.
Bringing the method under a single authority
A typical engagement starts by reading the ultrasonic procedures against the codes the work actually invokes, then checking each one against what the crews do in the field. The gaps are usually consistent: thickness ranges that exceed what the calibration blocks support, no transfer correction requirement, no scan plan, sizing rules that were never stated, and no treatment at all of the materials on site where conventional angle beam does not work.
From there the work is procedure revision, calibration standard rationalisation, a practical examination set matched to the site, and the ongoing approvals and interpretations that keep the programme current. Where the right answer is that ultrasonics is the wrong method, that is written into the procedure rather than discovered on a scaffold at the end of a shift.
If you want your ultrasonic procedures, calibration standards and technician examinations reviewed by a Level III who will put a name on the outcome, request a consultation at info@atlantisndt.com. The first deliverable is a written assessment of the current procedure set against the codes you actually work to.
What exactly does the Level III approve in an ultrasonic procedure?
The referencing code and acceptance criteria, the material and thickness range, surface condition and preparation, probe frequency, size and angle, wedge and couplant, the instrument and its performance checks, the calibration block and reference reflector, the calibration and recalibration interval, the scan plan and coverage, the method of discriminating geometry from flaws, the sizing technique, and the reporting and record requirements. Change any of the sensitive ones and the procedure requires requalification.
Does a UT Level III certification cover phased array and TOFD?
Not automatically, and this is where written practices most often fall behind their own field crews. Conventional ultrasonic certification demonstrates competence with pulse-echo instruments and manual scanning. Encoded phased array and time of flight diffraction bring focal laws, wedge delay and velocity calibration, encoder setup and imaging interpretation. Most credible practices treat them as separate qualification or as a technique endorsement with its own training, examination and practical.
When is DGS the wrong reference and DAC the right one?
DGS is analytical: the curve belongs to a specific probe and assumes a predictable beam and known attenuation, so it fails on coarse-grained, anisotropic or heavily attenuating material and on badly corroded surfaces. DAC is empirical, built on side-drilled holes in a block matching the material, thickness and curvature. Where the referencing code names one, that settles it. Where it does not, the material decides.
How much does temperature distort a hot thickness reading?
Enough to matter. Sound velocity in carbon steel falls as temperature rises, and the customary field correction is to reduce an uncompensated reading by roughly one per cent for every hundred Fahrenheit degrees above the calibration temperature. On a six hundred degree line that is around five per cent of over-reported wall. A pipe genuinely at minimum thickness can read comfortably above it, and the interval gets extended.
Why does transfer correction get skipped, and what does skipping it cost?
Because it takes two extra probes and ten extra minutes on a scaffold. The calibration block is clean, fine grained and machined; the component is rough, painted, corroded or cast. Sensitivity established on the block therefore over-states what reaches the component, sometimes by several decibels. The consequence is under-recording: real reflectors sit below the recording level, the report reads clean, and nobody knows the sensitivity was never achieved.
Does the Level III act as our API inspector or process safety auditor?
No. The API 510, 570 or 653 inspector authorises and signs the in-service inspection, and process safety auditing is a separate function. This engagement supplies NDT technical authority in the ultrasonic method: procedure development and approval, personnel qualification within your written practice, independent review of examination data, and technical representation when a client or regulator challenges a result. The two roles complement each other and should never be merged.