Hardness and impact test for centrifugal impeller of air compressor

Last fall, a chemical plant’s maintenance crew pulled a third-stage impeller from a process air compressor after a sudden spike in vibration. The blades and shroud looked fine — no rub marks, no impact dents. Out of habit, someone shot a few hardness readings across the hub with a portable Leeb tester. The numbers came back almost 20% below the original mill certificate. That stopped everyone. They sectioned a sacrificial blade and sent it for Charpy V-notch testing. The result? Impact energy had dropped below 12 J at operating temperature, against a required minimum of 27 J. The impeller had been running 24,000 hours and was frighteningly close to a brittle burst.

For anyone accountable for a centrifugal impeller in an air compressor — whether you are a quality inspector signing off an incoming part, a procurement manager writing a purchase specification, or a maintenance lead deciding whether a wheel goes back into the machine — hardness and impact testing are not just paperwork exercises. They can catch what dimensional checks, chemistry reports and even dye penetrant will miss.

Hardness testing: reading what the surface can’t tell you

Walk into a quality lab that handles compressor impellers and you will find a bench Rockwell or Brinell tester sitting as close to the inspection workflow as the CMM. The test itself is simple: press an indenter into a prepared spot, measure the dent, convert to a scale. But getting numbers you can trust on a centrifugal impeller takes far more discipline than a lab manual suggests.

Location is the first discipline. A centrifugal impeller isn’t a uniform block of metal. The hub is thick, the blades are thin, and the shroud curves. On a shrouded wheel, the back shroud near the bore experiences the highest triaxial stress, so that is a mandatory test zone. Blade leading edges can be under 2 mm thick and often require micro-Vickers (HV) because a Rockwell indenter would wreck the profile. Over the years I’ve seen smart QA teams mark exactly the same three blades, hub face and back shroud on every incoming wheel and draw a simple hardness map. That map later becomes a field baseline — one of the most useful documents a maintenance crew can have.

Surface preparation separates a useful reading from a guess. A portable Leeb tester, the workhorse of field inspections, is sensitive to roughness, curvature and part mass. On a lightweight aluminum impeller for a low-pressure stage, readings can drift if the part isn’t solidly backed. Put the device on a sand-cast surface without grinding a small flat, and the scatter might exceed 10-15 HBW — enough to accept a bad wheel or reject a good one. In incoming QC, I always insist on a spot dressed with at least 400-grit paper. The extra few minutes per area is cheap insurance.

And then there is the “higher must be better” trap. Centrifugal impellers in air compressors often use precipitation-hardening grades like 17-4PH or 15-5PH. Condition H900 can push hardness above 40 HRC, but impact toughness falls off a cliff — especially if that wheel will see a cold start or a surge event. H1150 or H1150M, sitting in the 28–35 HRC band, usually strikes a safer balance. When talking to heat treatment shops, I repeat one request: give me the hardness range and the corresponding Charpy values from the same coupon. One without the other is an incomplete picture.

Impact testing: the cheap test that gets skipped until something breaks

The Charpy impact test (ASTM E23, ISO 148) swings a hammer and breaks a small notched bar. It costs very little, yet impact requirements frequently get left out of procurement specifications because some legacy standards for centrifugal impellers — especially on integrally geared compressors — still hang everything on tensile strength and hardness. That oversight ignores reality. An impeller absorbs start-up torque spikes, surge transients, and the occasional slug of liquid. If the material lacks impact toughness, a tiny casting flaw or a weld underbead crack can propagate in dozens of cycles, not thousands.

For welded impellers, the heat-affected zone is the weakest story nobody wants to investigate. A good welding procedure qualification will pull Charpy specimens directly from the HAZ and demand values that match the base metal minimum. Without that check, a wheel can pass all surface NDE and still carry HAZ areas with impact energy under 10 J. I recall a case where a 13Cr-4Ni fabricated impeller burst during a shop overspeed test at 115% of rated speed. Dye penetrant had shown no flaws. The autopsy found Charpy values of 7–9 J in the HAZ — the material had gone brittle. After that incident, the end user made HAZ impact testing a contractual hold point on every new build. A cheap test could have prevented a dangerous failure.

For maintenance crews, destructive impact testing on a part that needs to go back into service isn’t possible. That is where the incoming baseline report earns its salary. Within a given alloy family, hardness and impact toughness tend to move together: if hardness creeps down over years of service temperature exposure, impact properties often degrade too. When an impeller is finally scrapped, cutting a Charpy sample from it gives a post-mortem data point that helps set retirement limits for the next one.

Incoming QA: a workflow that rejects trouble before it enters the rotor bay

If your job is to accept or reject impellers at the warehouse door, a systematic approach catches problems that paper certificates will never show.

• Match the heat number stamped on the hub with the mill certificate. No stamp, no acceptance — serial traceability is the only thread connecting test results to the physical part.

• Take hardness at the agreed grid points. Record the indenter type and scale. Convert only when necessary, and state the conversion standard (e.g., ASTM E140). Blending Rockwell C and Brinell without transparency invites trouble.

• For welded or cast impellers, demand separate test coupons that went through the same post-weld heat treatment or casting cycle. Test those coupons for both hardness and impact energy. Do not accept generic heat lot data from the steel mill as a substitute.

• On precipitation-hardened alloys, ask for the aging curve — hardness against aging time at temperature. This reveals whether the supplier used a proper cycle or rushed the furnace to meet a delivery date.

• Budget for periodic destructive testing. Set aside one impeller from a batch, or a spare forging, slice it, map hardness through the thickness, and run Charpy specimens from the core and surface. The variation often uncovers residual stress patterns invisible to surface methods.

Procurement: writing a purchase spec that acts like a filter

Vague PO lines like “Material: 17-4PH, hardness 35 HRC max” leave too much room for interpretation. You might receive a part that hits the hardness number but has no verified impact toughness — or worse, one re-heat-treated multiple times to rescue a bad batch, with a scrambled microstructure nobody notices until it spins apart.

A tighter spec that high-quality manufacturers already work to looks more like this:

• Material grade and heat treatment condition, for example “15-5PH H1025”.

• Hardness range, not just a ceiling. “28–34 HRC, tested per ASTM E18 on a machined flat at the back shroud.”

• Minimum Charpy V-notch impact energy at the lowest expected start-up temperature, with specimen location stated: “27 J average min. at −10°C, transverse specimens from a hub prolongation.”

• For welded wheels: “HAZ and weld metal impact tests required; values shall reach at least 90% of base metal minimum.”

• Witness rights: “Purchaser may witness hardness and impact testing at the foundry. Test reports for mechanical properties to be submitted for approval before shipment.”

• Incoming retesting: “Purchaser may perform random incoming hardness and impact verification on 2% of impellers per batch. Any failure triggers 100% testing at supplier cost.”

This reads as strict, but it filters out shops that substitute paperwork for process control. Good foundries and forging houses rarely push back — their records already meet these levels.

Maintenance teams: field hardness as a health indicator, not a scorecard

When the compressor is open for turnaround, you have hours, not days. The Charpy pendulum won’t be sitting on the mezzanine, but a well-maintained portable hardness tester will. Use it on the spots that tell the biggest story:

• Hub bore area: softening here often signals overtemperature from a thrust bearing failure or a lube oil event.

• Blade tips near the shroud: a rub can locally heat the material and alter hardness in a narrow band. A row of tip readings reveals the pattern.

• Welded joints on fabricated impellers: check the hardness gradient from base metal into the weld. A drop over 15–20% points to post-weld heat treatment degradation.

Record the numbers and overlay them on the baseline map from the original QA file. Plot the trends across several overhauls. A gradual hardness loss beyond 8–10% on the hub often means microstructural aging that saps both fatigue strength and impact toughness. At that point, grit-blasting and recoating are not a fix — scrap the impeller or send it for a detailed lab evaluation.

One maintenance lead I know uses a straightforward rule: if the average portable hardness on the back shroud of a stainless impeller slips below 250 HBW from an original 290 HBW, he kicks off the procurement process for a spare. That habit has saved two unplanned outages.

Repair and re-rate: why you cannot ignore the tests when the boundaries move

Impeller repair — especially weld buildup on eroded blades — rewrites the local hardness and impact story. A repair vendor may promise to restore the profile, but unless they repeat the full post-weld heat treatment cycle, you end up with a softened HAZ carrying dangerously low toughness. Hardness mapping after repair, plus impact specimens from a mock-up coupon welded with the identical parameters, should be non-negotiable deliverables.

If an air compressor is being re-rated for higher flow or pressure, the impeller stresses climb. Impact testing becomes critical because the original material may have had just enough toughness for the original operating window. Re-running impact values at the new, possibly colder start-up condition can flag a risk that hardness alone will never reveal.

A final word to the people who keep the wheels turning

Hardness and impact tests for a centrifugal impeller of an air compressor are not abstract lab exercises. They are the difference between a wheel that weathers twenty years of starts, surges and stops, and one that writes a very expensive failure report. When you hold the indenter, write the spec, or decide whether a used impeller goes back onto the pinion shaft, you are reading the material’s actual condition. The chemistry might be perfect, the dimensions spot on, but if the hardness map drifts or the Charpy bars snap with only a handful of joules, the impeller is talking to you. Listen to it.