How a Chinese expert helped a European aerospace foundry cut inclusion defects from 12% to 2.3%. Filtration, degassing, refractory control, and traceability.
In early 2024, I received an email from a quality director at an aerospace foundry in France. His company produces aluminium alloy castings for commercial aircraft and defence programmes. The foundry had been NADCAP accredited for years. Its pyrometry was compliant. Its traceability was solid. But its inclusion rejection rate had climbed to 12%.
The customer had issued a warning. If the rejection rate did not drop below 3% within four months, the foundry would lose the contract.
The quality director told me: “We have changed the ceramic filters. We have increased degassing time. We have replaced the refractory in the launder. Nothing works. The inclusions are still there.”
I spent two weeks on site. We did not replace the furnace. We fixed the melt treatment system. Six months later, the inclusion rejection rate was 2.3%. The contract was retained. The foundry is now producing flight-critical castings for two additional aerospace programmes.
This is how we did it.
The Starting Point: 12% Inclusion Rejection
The foundry operates a 3-tonne gas-fired reverberatory furnace feeding a holding furnace and a casting station. They melt 7050 and 6061 alloys for structural aerospace components. Production is around 2.5 tonnes per day, two shifts, five days per week.
The customer specification was strict:
| Parameter | Aerospace Specification | Foundry Result | Gap |
|---|---|---|---|
| Inclusion count (≥40 μm) | ≤1,000 particles/kg | 3,600 particles/kg | 3.6x too high |
| Hydrogen content | ≤0.06 ml/100g | 0.14 ml/100g | 2.3x too high |
| X-ray rejection rate | <3% | 12% | 4x too high |
The inclusions were visible on X-ray and ultrasonic inspection. They appeared as small, irregular dark spots. Some were alumina films. Some were refractory particles. Some were spinel clusters.
The foundry had tried three different filter suppliers. All failed. The problem was not the filter. It was the system around it.
Step 1: Diagnosing the Root Cause
I started with a full process audit. I used PoDFA — the Porous Disc Filtration Apparatus — to identify inclusion types and quantities. The results were revealing.
| Inclusion Type | Share | Likely Source |
|---|---|---|
| Alumina films | 42% | Turbulence during transfer |
| Refractory particles | 28% | Launder and furnace lining |
| Spinel clusters | 18% | Oxidation in holding furnace |
| Titanium boride clusters | 12% | Grain refiner addition |
The alumina films were the biggest problem. They form when molten aluminium is exposed to air during transfer. The transfer distance was 9 metres. The metal was poured in an open launder. Every pour created new oxide films.
The refractory particles came from the launder lining. It was cracked and spalling. Small pieces of refractory were breaking off and entering the melt.
The spinel clusters formed in the holding furnace. The furnace atmosphere was not controlled. Oxygen was entering through a worn door seal.
The titanium boride clusters came from the grain refiner. It was being added to the holding furnace instead of the melting furnace. The particles did not have time to disperse properly.
The root cause was clear: the melt treatment system was not designed for aerospace work. The furnace was fine. The transfer, filtration, and holding were not.
Step 2: Fix the Transfer System
The first priority was to reduce turbulence during transfer. Alumina films form when the metal surface is disturbed. A smooth, covered transfer minimises that.
We made four changes:
Shortened the transfer distance from 9 metres to 3 metres. We repositioned the holding furnace next to the melting furnace.
Replaced the open launder with a covered, heated launder. The cover reduced air exposure. The heating kept the metal at temperature.
Added a launder baffle to smooth the flow. Simulation-guided baffle design has been shown to improve cleanliness in aerospace-grade structural castings.
Preheated the launder before each pour. A cold launder causes the metal to freeze on the walls, creating turbulence.
Cost: €14,000 for the new launder and installation.
The result: alumina film inclusions dropped by 60% in the first PoDFA test after the change.
Step 3: Install Two-Stage Ceramic Foam Filtration
The foundry was using a single ceramic foam filter. It was not enough. For aerospace work, two-stage filtration is standard.
We installed:
| Stage | Filter Size | Purpose |
|---|---|---|
| Coarse | 30 ppi | Remove large inclusions >100 μm |
| Fine | 60 ppi | Remove smaller inclusions 40–100 μm |
The filters were preheated before use to prevent thermal shock. The metal flow rate was controlled to maintain laminar flow through the filters. Turbulence would re-entrain inclusions.
We also changed the filter housing. The original housing had dead zones where metal could stagnate. The new housing had a tapered inlet and outlet to maintain smooth flow.
Cost: €8,500 for the filtration system upgrade.
The result: refractory particle inclusions dropped by 70%. Titanium boride clusters dropped by 50%.
Step 4: Control the Holding Furnace Atmosphere
The holding furnace was open to air. Oxygen and moisture were being absorbed continuously. That created spinel clusters and increased hydrogen pickup.
We sealed the furnace with a high-temperature gasket and installed a nitrogen blanketing system. The oxygen content inside the furnace was maintained below 0.5%. The moisture content was maintained below 50 ppm.
We also replaced the worn door seal. The original seal was flat and leaking. The new seal was a high-temperature ceramic fibre type.
Cost: €4,200 for the sealing and nitrogen system.
The result: spinel cluster inclusions dropped by 80%. Hydrogen content dropped from 0.14 to 0.08 ml/100g.
Step 5: Upgrade Degassing
The foundry was using a manual degassing lance. The operator inserted the lance for 4–5 minutes, then moved to the next ladle. There was no measurement. No control. No consistency.
We installed a rotary degasser with argon. The parameters:
| Parameter | Setting |
|---|---|
| Rotor speed | 650 rpm |
| Argon flow rate | 20 L/min |
| Degassing time | 10 minutes per 500 kg batch |
| Gas purity | 99.999% argon |
We also installed a portable hydrogen analyser. The operator measures hydrogen before every pour. If the reading is above 0.06 ml/100g, the melt goes back for additional degassing. If it is below, the melt is released to casting.
Cost: €12,000 for the degasser and analyser.
The result: hydrogen content dropped to 0.05 ml/100g within two weeks. The key was not just the equipment. It was the measurement. You cannot control what you do not measure.
Step 6: Fix the Grain Refiner Addition
The titanium boride clusters came from adding grain refiner to the holding furnace. The particles did not have time to disperse. They clustered together and were captured as inclusions.
We moved the grain refiner addition to the melting furnace. The longer residence time allowed the particles to disperse fully. We also reduced the addition rate by 15%. The grain refinement was still effective. The clusters disappeared.
Cost: €0. Just a procedure change.
Step 7: Improve Process Control and Traceability
Aerospace requires full traceability. Every casting must be linked to its melt history. The foundry had a system, but it was not capturing all the data.
We implemented a simple but effective traceability system:
| Data Point | Recording Method |
|---|---|
| Charge material source | Barcode scanning at receiving |
| Charge weight | Load cell on charging system |
| Melting temperature | PLC data logger |
| Degassing parameters | PLC data logger |
| Hydrogen measurement | Manual entry with time stamp |
| Filtration parameters | PLC data logger |
| Casting temperature | PLC data logger |
| Inclusion test results | PoDFA and K-Mold records |
All data was stored in a central database. Each casting was marked with a melt number. The melt number linked to all process data. This data became part of the PPAP documentation that the aerospace customer’s auditors reviewed.
Cost: €6,000 for the data logging system.
Step 8: Train the Team
The best equipment in the world fails without trained people. The foundry had competent operators. But they had never been trained on aerospace inclusion control.
We spent three days training:
- Why inclusions matter and how they form
- How to operate the rotary degasser
- How to handle ceramic foam filters
- How to maintain the nitrogen blanket
- How to record data accurately
- What aerospace auditors look for
We also wrote new standard operating procedures. One page per critical process. No 50-page manuals that nobody reads.
Cost: €4,500 for training and procedure development.
The Results After Six Months
| Metric | Before | After | Aerospace Spec |
|---|---|---|---|
| Inclusion count (≥40 μm) | 3,600 particles/kg | 680 particles/kg | ≤1,000 |
| Hydrogen content | 0.14 ml/100g | 0.05 ml/100g | ≤0.06 |
| X-ray rejection rate | 12.0% | 2.3% | <3% |
| Ultrasonic rejection rate | 8.5% | 1.8% | <2% |
| Aerospace audit | Warning | Passed | — |
The customer cancelled the contract warning. They increased their order volume by 25%. The foundry is now producing flight-critical castings for two additional aerospace programmes.
What Made the Difference
Three things drove the improvement:
1. Transfer and filtration. Reducing turbulence and installing two-stage ceramic foam filtration removed the alumina films and refractory particles. This was the single biggest win.
2. Atmosphere control. Sealing the holding furnace and adding nitrogen blanketing eliminated spinel clusters and reduced hydrogen pickup.
3. Measurement and training. Installing a hydrogen analyser and training operators on aerospace procedures made the improvements sustainable.
I have written about how to achieve aerospace-grade aluminium melt cleanliness. The principles apply here.
What You Can Apply
If you are fighting inclusion problems in aerospace castings, ask yourself these questions:
Can you measure hydrogen before every pour? If not, you cannot control it.
Do you have two-stage ceramic foam filtration? If not, inclusions will exceed aerospace limits.
Is your furnace atmosphere controlled? If not, spinel clusters will form.
Is your transfer covered and heated? If not, alumina films will form.
Where do you add grain refiner? If it is in the holding furnace, you will get clusters.
Do you have full traceability? If not, you will fail the audit.
These are not theoretical. They are the minimum requirements for aerospace work. I have seen foundries in China, Southeast Asia, and Europe fail because they ignored one of these.
When to Call an Expert
You can measure these things yourself. But aerospace qualification is not a DIY project. It requires expertise in metallurgy, furnace design, and quality systems.
The French foundry had a competent team. They had kept the furnaces running for years. But they had never been trained on aerospace inclusion control. They did not know what they did not know.
My fee for the two-week audit and commissioning support was €22,000. The value of the retained contract was over €4 million per year. That is a good return.
If you are fighting inclusion problems, and your rejection rate is above 3%, call someone. Or use the checklists on this blog. Either way, do not wait.
Final Thoughts
Solving aerospace inclusion problems is not about buying a new furnace. It is about controlling the melt treatment system. Transfer, filtration, atmosphere, degassing, grain refiner addition, and traceability. Each one must be right. If one fails, the others cannot compensate.
In my 25 years, I have seen foundries transform their aerospace quality by fixing the system, not the furnace. The furnace did not change. The people did.
If you run an aerospace foundry, start by measuring your inclusion count and hydrogen content. If either is out of specification, fix it. You will see the difference in your rejection rate.
Internal links: How to Achieve Aerospace-Grade Aluminium Melt Cleanliness, The State of China’s Aerospace Aluminium Melting Industry – Expert Analysis, Case Study – Achieving NADCAP Certification for a Sichuan Aerospace Melting Plant, Case Study – How a Chinese Expert Achieved Aerospace Melt Purity for a US Supplier, How to Reduce Gas Porosity in Die Casting Aluminium Melts
