ESI Composites 6 — Autoclave, Quality Assurance & Certification

Before you start

How this course works

The capstone. Everything to this point has been about making a part. This course is about proving one — autoclave processing where the properties are highest and the documentation is the deliverable, then inspection, testing, traceability and the industry certifications that turn an ESI technician into a recognised one. Fourteen chapters, 80% on the final — the same pass mark ACMA uses on its Certified Composites Technician exams.

Chapter 1

🔧The Autoclave — What Pressure Buys

"Why not just use the oven?" Because the spec says a void content the oven cannot reach.

Easy Composites reduce it to one sentence: "The difference between these two pieces of equipment is pressure; an autoclave is a pressure chamber... whereas a curing oven only has normal atmospheric pressure inside."

Composites 4 Chapter 1 established the ceiling on a vacuum bag: one atmosphere, and no more, because you are only ever removing air so the outside can press in. An autoclave breaks that ceiling by pressurising the whole vessel, so the pressure on the laminate is atmospheric plus whatever the vessel is run at.

• Collapses voids. Gas trapped in a laminate is compressible. More pressure means smaller voids, and small enough voids are effectively gone. This is the main reason autoclaves exist.
• Raises fiber volume. Higher consolidation pressure squeezes the plies closer together.
• Improves ply-to-ply intimacy, which is what interlaminar strength depends on.
• Suppresses volatiles. Anything that would boil off or expand at cure temperature is held down.

Easy Composites are blunt: it is "rare to find an autoclave for less than around £50k (60k euro, $70k)," larger vessels considerably more, and they are "heavy, expensive to run, requiring regular inspection and servicing to ensure their safety."

Gurit note where the gap opens up: for thick laminates — above about 3 mm — it becomes difficult to remove entrapped air between plies and around details, and traditional prepregs need multiple warm debulking stages that significantly increase manufacturing time. That is exactly the territory where applied pressure earns its cost.

VBO — vacuum bag only, the out-of-autoclave route of Composites 5 — is a serious process, not a compromise, and the industry treats it that way. Abaris run comparative training in which panels made by different repair methods are made to ASTM D5687M and compared by short beam shear interlaminar shear strength to ASTM D2344M, explicitly comparing vacuum bag only against autoclave results. The point of that exercise is that the comparison is measured rather than assumed.

Chapter 2

📊The Autoclave Cure Cycle

Pressure goes on before the resin has flowed. The part is locked in with the air still in it.

An autoclave cycle controls four things at once against time: temperature, pressure, vacuum and the resulting part state. Getting the relationship between them right is the whole skill.

Think about what the resin is doing. As temperature rises, viscosity falls — the resin gets thinner and can flow and let trapped air escape. Then it starts to gel, and Gurit define that point precisely: the gel point is where "the resin is no longer liquid and has lost the ability to flow."

• Pressure applied too early, while the resin is still stiff and the air has not had a chance to move, can trap air rather than expel it — and can squeeze resin out before it has done its job.
• Pressure applied too late, after the resin has begun to gel, cannot consolidate anything — the resin has lost the ability to flow and the voids are locked in.
• The window is between those two, and it is material specific. That is why the cycle specifies it rather than leaving it to judgement.

Composites 5 Chapter 11 applies in full and matters more here, because an autoclave load is usually bigger and more varied:

• On the tool and near the part, positioned to represent what the laminate is experiencing.
• Include the lagging location — the thickest section, the heaviest tool, the most shadowed corner. The hold does not start until that one is at temperature.
• Include a leading location too where over-temperature is a risk on a thin section.
• Pad the leads with breather where they could puncture the bag — CKN note this use specifically.
• Verify and calibrate. An uncalibrated thermocouple produces a trace that proves nothing.

Worth knowing that the industry is moving past time-and-temperature recipes. Abaris teach Material State Management — described as moving from temperature-based cure cycles to material-state control, using embedded and microwire sensors, encapsulated-specimen rheometry (ASTM D7750-12) and CMH-17 glass transition methods, together with shelf-life and out-time management. The principle: control the cure by what the resin is actually doing, not by what the clock says it should be doing.

Chapter 3

⚠️Autoclave Safety and Loading

It is a pressure vessel at temperature with people around it. Treat it like one.

An autoclave is a pressure vessel operating at temperature. Easy Composites note they are "heavy, expensive to run, requiring regular inspection and servicing to ensure their safety." That is not a maintenance note — it is a statement about the hazard class of the equipment.

• Certified and inspected. Pressure vessels are subject to inspection and certification requirements, and the certification record is a condition of operation, not paperwork.
• Door interlocks are safety devices. They exist to make it impossible to open a pressurised vessel. They are never bypassed, taped, defeated or worked around, for any reason.
• Relief devices are tested. Pressure relief is what stands between a fault and a vessel failure.
• Only authorised, trained operators run it. This is not a machine you learn by watching.
• Nothing enters a pressurised or hot vessel. Ever.

• Hot surfaces. Tools, racks and the vessel itself. Burns are the most common autoclave injury and they happen at unloading, not during the run.
• Let the load cool before handling — which is also what the part requires (Composites 5 Chapter 12). The safety requirement and the quality requirement agree.
• Volatiles. Curing resin releases volatiles. Vessel venting and shop ventilation must handle them, and the area is not somewhere to stand and breathe at the end of a cycle.
• Nitrogen-inerted vessels, where used, introduce an asphyxiation hazard in and around the vessel. That is a confined space and oxygen-monitoring question, and it is governed — Chapter 296-809 WAC for confined spaces, and WAC 296-304 for shipyard operations.

Composites 1 Chapter 9 applies here without modification. The chapters that bear most directly on autoclave work:

Citations from current Washington Administrative Code. Confirm current text before relying on any citation operationally, and work to the equipment manufacturer operating procedures and the ESI written program.

Every part in the load must see the cycle the specification requires. That is a loading problem as much as a control problem.

• Space for circulation around every tool. A tool packed against a wall or another tool is in dead air and will lag.
• Group compatible cycles only. Everything in the load must want the same schedule — same as an envelope bag in Composites 5 Chapter 8, and with more at stake.
• Instrument the extremes, not the average. Thermocouples on the heaviest, most shadowed tool and on the thinnest fastest-heating part.
• Route lines so nothing is trapped, pinched or abraded when the door closes or the load shifts.
• Nothing sharp against a bag. Pad tool corners, rack edges and thermocouple leads with breather.
• Confirm vacuum on every bag after loading, in its final position, before the door closes.

Chapter 4

📏Fibre Properties — The Numbers Behind the Choices

"Carbon is stronger than glass." Stronger how, and per what?

Composites 1 Chapter 2 taught you to recognise reinforcements. This chapter attaches numbers to them, because at this level "stronger" is not an argument.

These are independent. A material can be very strong and quite flexible, or very stiff and relatively weak. Most marine structural problems — a flexing panel, a soft deck, a hull that oil-cans — are stiffness problems, not strength problems. Choosing a high-strength fibre to fix a stiffness complaint solves nothing.

Fibre property figures quoted from the Gurit Guide to Composites, Table 1. Confirm against the supplier data sheet for the specific product before design use.

Read that table carefully, because it makes the point better than any explanation:

• HS and HM have the same tensile strength — 3500 MPa — but HM is far stiffer.
• IM is the strongest at 5300 MPa, and it sits between the other two on stiffness.
• All three have the same density. So the choice is genuinely between strength and stiffness, with no weight penalty either way.

Specific modulus is stiffness divided by density — stiffness per unit weight. It is the number that actually matters when weight is a design constraint, and it is why carbon is used in marine and aerospace work at all. A material can be beaten on absolute stiffness by steel and still win decisively on specific modulus, because it is a fraction of the weight.

Gurit list the variables that determine the properties of a finished laminate: the resin properties, the fibre properties, the fibre volume fraction, and the fibre orientation.

Note that two of those four — fibre volume fraction and orientation — are entirely under the control of the shop floor. That is the whole reason this ladder exists. You cannot change what carbon fibre is, but you decide how much of it ends up in the part and which way it points.

Chapter 5

⚖️Process Comparison — Where Everything Sits

Six courses in. Now put the whole map on one page.

This is the map of the whole ladder, and it is the chapter to come back to when somebody asks how a part should be made.

Gurit identify where conventional prepreg starts to struggle: for thick laminates — above about 3 mm — "it becomes difficult to remove entrapped air between plies and around details," and traditional prepregs need multiple warm debulking stages that significantly increase manufacturing times.

That is a real cost, and it is what several newer approaches exist to attack.

Gurit describe a resin film sandwiched between two dry fibre layers — vacuum extracts the air from the laminate, monolithic or sandwich, before heat softens the resin and it impregnates. Air out first, resin in second. The claimed advantages are worth knowing because they explain what the industry is optimising for:

• Fibre volumes accurately achieved with low void contents for very thick laminates — up to 100 mm.
• High resin mechanical properties, from the solid initial polymer state and an elevated-temperature cure.
• Enables lower-cost heavyweight materials, e.g. 1600 gsm, and fast deposition rates, reducing component cost.
• Very robust and repeatable, with resin content accurately controlled.

Gurit also note that out-of-autoclave prepregs are low-temperature curing, at 60 to 120 °C — a wider and lower range than the general 100 to 150 °C figure.

Process comparison content quoted from the Gurit Guide to Composites and Easy Composites guidance. Confirm against current supplier documentation before specifying a process.

1. What is specified? A property, void content or standard named in the specification decides it immediately.
2. How many? One is a hand layup. Twenty is a tool and a repeatable process.
3. How thick? Above roughly 3 mm, conventional prepreg starts paying a debulking penalty.
4. How big, and what is the flow distance? The constraint that kills large infusions.
5. What tooling exists and what can it survive?
6. What is the cost of failure? That sets how much process control you buy.
7. Can we prove it? If the answer has to be demonstrated, you need the documentation chain of this course.

Chapter 6

📜Traceability and the Part Record

A part is questioned two years later. Nobody can say which batch of prepreg went into it.

Traceability means being able to answer, for any part, at any point in the future: what material went into it, who made it, how, on what equipment, and what was measured. It exists because composite quality is invisible after the fact — you cannot look at a finished part and see the cure cycle, the bond preparation or the ply schedule.

A nonconformance is anything that did not go according to the specification. A vacuum excursion, a thermocouple that never reached dwell, a ply out of sequence, a material past out-life, a missed debulk. The handling is fixed and it is not negotiable:

1. Stop. Do not continue processing a part that is already nonconforming without a decision.
2. Record it exactly as it happened, with times, values and conditions. Not a summary. Not a softened version.
3. Raise it to whoever is authorised to decide. It is not your call and it is not your risk to carry.
4. Disposition is recorded — use as is, rework, repair, scrap — with who authorised it and on what basis.
5. The record travels with the part. Permanently.

On certified or documented work, the customer is buying two things: a physical part, and evidence that it is what it is supposed to be. Without the second, the first has no demonstrable value — because nobody can prove anything about it after the fact.

• Record as you go, not afterwards. Reconstructed records are wrong records, and the numbers you need are exactly the ones you will not remember.
• Record actuals, not intentions. What the trace shows, not what was programmed. What was used, not what was planned.
• Batch numbers at the point of use, off the container in your hand.
• Photograph. Layups, stacks, defects, findings. A photograph settles arguments a description cannot.
• Sign what you did and only what you did.

Chapter 7

👁️Visual Inspection and Tap Testing

The part passes a glance. Under raking light it has porosity across the whole face.

Gurit list visual inspection first among NDT methods for a reason: it is free, immediate, and it finds a remarkable amount. Their stated detection list for visual inspection is worth learning:

• Raking light. A light low and nearly parallel to the surface, sighted along. It exposes porosity, waviness, print-through and surface defects that overhead light hides completely. Composites 2 Chapter 5.
• A defined route. Same sequence every time, covering every area, so nothing is missed because attention drifted.
• Corners and radii first. That is where bridging voids are — nine out of ten prepreg defects, Composites 5 Chapter 5.
• Edges and trim lines. Ply drop-offs, dry edges, delamination exposed by cutting.
• Compare against a reference part where one exists. "Different from the good one" is a far more reliable judgement than "looks wrong."
• Photograph anything you find, with something in shot for scale and location.

Gurit describe it as tapping with a coin and listening for a change in pitch or volume, and rate it cheap, simple and quick — with one stated disadvantage: it relies heavily on the operator.

• Sound laminate rings sharp and bright. Delaminated, voided or wet laminate goes dull and dead. Composites 2 Chapter 1.
• Work a grid, not a wander. Mark the grid if the part is large.
• Calibrate your ear on known-good material on the same part before you start, and re-calibrate as you move between thicknesses.
• Thickness changes the sound. A thick section sounds different from a thin one and from a cored one. That is not a defect.
• Mark the boundary of anything dull as you find it, and photograph it.

Chapter 8

📡Instrumented NDT — Methods and Their Limits

"Thermography found nothing." The defect was four times deeper than its diameter.

Every NDT method has a blind spot. Choosing one without knowing its blind spot is how a part gets cleared with a defect still in it. Gurit publish a comparison, and the disadvantages column is the important one.

Method comparison quoted from the Gurit Guide to Composites. Select and qualify NDT methods against the applicable specification for the work.

Two methods are explicitly ruled out on uncured material: ultrasonics is "unsuitable on uncured material," and for Lamb wave sensing "waves will not propagate in uncured material." Both rely on mechanical wave transmission through a solid, and uncured resin does not behave like one.

The practical consequence: you cannot NDT your way out of a process problem in progress. Inspection of a laminate happens after cure. Which brings you back to the theme running through this whole ladder — the quality is built in during setup and layup, and inspection confirms it rather than creating it.

Chapter 9

🔬Test Coupons, Standards and What Gets Measured

"It is strong." Compared to what, measured how, to which standard?

Everything in this ladder has described making laminates. This chapter is about the numbers that describe one, because on documented work "good" is not a description — a measured value against a named standard is.

This is genuinely useful on the shop floor. Measure the thickness of a cured laminate, count the plies, and you can work backwards toward the fiber volume you actually achieved — a real check on whether the consolidation was what you intended.

Gurit cite the following for laminate and sandwich testing. You are not expected to run these — you are expected to know they exist, that they are named in specifications, and that "tested" without a standard is meaningless:

From the Abaris course material, three more that appear constantly in bonded and repair work:

• ASTM D3762 — the wedge test for adhesive bond durability
• ASTM D2344M — short beam shear, for interlaminar shear strength
• ASTM D5687M — panel fabrication for test specimens
• ASTM D7750-12 — encapsulated-specimen rheometry, used in material state management

Standard references quoted from the Gurit Guide to Composites and Abaris course descriptions. Standards are revised; confirm the current edition and the applicable specification before use.

A travelling coupon is a test panel made from the same material, in the same cycle, in the same vessel, as the production part. It goes through everything the part goes through, and then it can be destructively tested — because the part cannot.

• Same material and batch as the part.
• Same layup and cure cycle, in the same run.
• Identified and traceable to the part and the run.
• Retained for the period the specification requires.

Gurit note that for a polyester/glass woven roving laminate, micro-cracking typically occurs at about 0.2% strain, with ultimate failure not occurring until 2.0% strain. That gap is the thing to understand: a laminate can begin micro-cracking at a tenth of the strain that finally breaks it — and micro-cracks are where water gets in. Long before a laminate fails, it can stop being watertight.

Chapter 10

💧Micro-cracking, Water and Long-Term Durability

The laminate has not failed. It has been leaking through micro-cracks for two seasons.

Structural failure is not how most marine laminates actually die. They die slowly, from water, and the entry point is micro-cracking.

That is a factor of ten. The laminate begins to crack internally at a tenth of the strain that finally breaks it. Everything between those two numbers is a laminate that is structurally intact and no longer sealed.

This is why a boat can flex within its design limits for years and still develop wet core, blisters and delamination. Nothing failed. It just cracked, quietly, at a strain nobody would call excessive.

1. Micro-cracks open a path through the resin between fibres.
2. Water enters and, in a cored structure, travels along the core and the bond lines — often a long way from where it got in.
3. The bond line degrades. Water at a core-to-skin interface is exactly where a sandwich panel loses the shear path that made it a structure (Composites 1 Chapter 6).
4. Freeze-thaw widens everything in a climate that does that.
5. Osmotic blistering develops where water reaches water-soluble material inside a polyester laminate.
6. Wood core rots. Foam does not, but the bond still fails.

Chapter 11

🔗Adhesive Bonding — Surface Preparation Science

The bond passed every visual check. The wedge test showed it failing at the interface.

Composites 2 Chapter 7 established that surface preparation is the bond. This chapter is the engineering version, because on documented work a bond has to be demonstrated, and there is a whole discipline devoted to it. Abaris devote a five-day course to adhesive bonding of composites and metals built around exactly this.

Surface preparation methods listed from Abaris course content. Select and qualify a preparation method against the applicable specification and adhesive system.

Notice the common thread: every one of them either removes contamination, increases mechanical key, or raises surface energy so the adhesive wets out. Usually more than one. That is the entire theory of surface preparation.

Composites 3 Chapter 6 taught the water break test for release systems, and it is the same physics read in the opposite direction:

One caveat carried over: the water break test is recommended to be run on a sample coupon of the substrate, not on the actual part surface.

ASTM D3762, the wedge test, is the standard method for assessing adhesive bond durability rather than just initial strength. A wedge is driven into a bonded joint to open a crack, and the specimen is then exposed to a hot, wet environment. The crack growth over time, and where the crack runs, is the result.

Abaris also list lap shear and DCB (double cantilever beam) coupons alongside it, and short beam shear to ASTM D2344M for interlaminar shear strength.

When a bonded joint breaks, where it broke tells you more than the load did:

Chapter 12

🎓ACMA Certified Composites Technician — The Programs

"Which certification should I do?" It depends entirely on what you actually do.

The Certified Composites Technician program is the American Composites Manufacturers Association industry certification for composites manufacturing. It is the recognised external credential for the work in this ladder.

ACMA also offer Basic Composites Training, CCT Onsite Training and Recertification.

ACMA describe it as Enroll → Learn (self-paced) → Test (online, year-round) → Get Certified, with a digital certificate issued immediately on passing.

An enrollment includes online learning modules, study guides and reference materials, practice activities, the Candidate Handbook, and two exam attempts. A third attempt requires purchasing a new enrollment. Digital study manuals are included in the enrollment fee for the CCT Open Molding and CCT Vacuum Infusion Process initial exams only.

• 18 or older.
• Relevant work experience or educational background in composites or a related field.
• Adherence to ACMA enrollment procedures.

Note what is not required: a formal qualification or a fixed number of years. The program is built for working technicians.

The Open Molding study guide and exam are also available in Spanish.

As an example of the depth expected, the Open Molding body of knowledge is weighted:

CCT program details quoted from ACMA published materials and the CCT Candidate Handbook. Program structure, content and fees are revised — confirm current details with ACMA before enrolling.

Chapter 13

📝Sitting the CCT Exam and Staying Certified

"Unanswered counts as incorrect." Read that again before you plan your two hours.

The mechanics matter, because several of them catch people out.

• Work the ACMA study guide for your track. The enrollment includes it, and the exam is built on it.
• Use the module weightings. If Laminating Application Technology is 30% of the Open Molding exam, that is roughly 30 questions. Spend your study time proportionally.
• Do the practice activities in the enrollment. They calibrate you to the question style.
• Use these ESI courses. Each chapter check and each final exam here is deliberately built the same way — multiple choice, with an explanation on every answer, at an 80% pass mark.
• Learn the safety content properly. It is a full module and it is the part working technicians most often skim.
• Rehearse the clock. Sit a timed set. Seventy-two seconds per question feels different from reading at your own pace.

Exam mechanics, fees and program details quoted from ACMA published materials and the CCT Candidate Handbook, and from Abaris and IYRS published course information. All are revised periodically — confirm current details directly with the provider.

Chapter 14

📋Building the ESI Quality System

Everyone knows how to do it right. Nobody can point at where it is written down.

A quality system is not a binder. In practical terms it is the answer to four questions, for every job the yard does:

Everything else is administration around those four.

Production Tech 1 taught that you cannot tell whether a change is an improvement without a stable baseline. A quality system is that baseline, formalised:

1. Standard work is the current best-known way — the baseline.
2. Records measure what actually happened against it — the data.
3. Nonconformances and defects point at where the standard is failing — the signal.
4. Corrective action and change control update the standard — the Act step of PDCA.
5. And the new standard becomes the next baseline.

Chapter 15

🔍Audits and Being Audited

The auditor asks one question: show me. Everything depends on the answer.

An audit checks one thing: can you demonstrate that what you say you do is what you actually do? It is not a test of whether the parts are good. It is a test of whether the system produces good parts reliably and demonstrably.

The discipline is established enough that Abaris run a dedicated five-day course, Fundamentals of Advanced Composites for Auditors — a whole training program aimed at people whose job is to assess composites operations.

1. Reads what you say you do. Procedures, specifications, work instructions.
2. Watches what you do. On the floor, in real time.
3. Traces a part. Picks one and follows it end to end — material batch, tool, layup, cure trace, inspection, test, disposition, and the people who signed each stage with their qualification records.
4. Looks for gaps between the three.

• Do not clean up for an audit. An operation that looks different on audit day is telling the auditor exactly what normal looks like.
• Check that documents are current and that the version on the floor is the version in the system. Superseded documents at the workstation are a classic finding.
• Check calibration is in date — thermocouples, gauges, ovens, scales.
• Check training records match the people doing the work.
• Close out open corrective actions, or be able to show they are being worked.
• Trace a part yourself first. Pick one at random and follow it end to end. Whatever you cannot answer, the auditor will ask.

A finding is somebody outside the operation, with fresh eyes and a structured method, telling you where your system does not work. That is exactly what a Kaizen system is trying to generate internally, and it arrives already written up.

• Treat a finding as a signal, not an accusation. Processes, not people.
• Find the root cause rather than closing the instance. A superseded document at one workstation is usually a document control problem, not a workstation problem.
• Change the standard, then train it, then verify it stuck. That is the Act step.
• Look for the same failure elsewhere. If one process had it, check the others.

Chapter 16

♻Closing the Ladder — Kaizen, Mastery and What Comes Next

Six courses, eighty-four chapters. The point was never the certificates.

Six courses, eighty-four chapters, six certificates. The material has run from what a composite is to how you prove one. This chapter is about what to do with it.

That single fact drives everything this ladder has taught:

• Why quality has to be built in rather than inspected in.
• Why setup is where the inspection really happens — the last point anything can be corrected.
• Why standard work matters more here than in trades where you can see the result.
• Why the record is the deliverable, not the paperwork around it.
• Why never covering a defect is not a moral flourish but a structural requirement of the whole system.

Notice that the top of that table is not about working faster. It is about deciding better and writing it down. Production Tech 1 said improvement is part of the job rather than extra work on top of it. This is what that looks like in a composites shop.

Every one of those is a number a technician can generate. None of them requires permission.

Lean stands on two pillars: continuous improvement and respect for people. It is worth being concrete about what the second one means in this yard, because it is the one that gets treated as a slogan.

• Flags are aimed at processes, not people. Every defect table in this ladder describes a mechanism, not a culprit.
• Safety is never traded for speed. A faster method that is less safe is not an improvement, does not get flagged as one, and does not become a standard.
• The person doing the work is the one who sees the improvement. That is why the ** marker exists and why there are no bad flags.
• Nobody is asked to carry a risk that is not theirs. Nonconformances go up. Safety concerns go up immediately.
• Training people toward recognised external certification makes them more valuable everywhere, not just here. ESI does it anyway.

Final certification

🎓Final exam

Forty-six questions drawn from all sixteen chapters — the same 80% pass mark ACMA uses on its CCT exams. Put your name on the certificate and your completion is recorded to ESI training automatically.

ESSENTIAL SHIPYARD UNIVERSITY

Certificate of Completion

This certifies that

has successfully completed ESI Composites 6 — Autoclave processing, inspection, quality assurance & certification — and passed the final examination to the ESI standard.

Jason Knott

President

Essential Shipyard Industries

★CERTIFIED

Date Issued

Essential Shipyard University