Copper-nickel protects itself with a film, and a film can be washed off. Every copper alloy used in seawater has a velocity above which the water removes the oxide layer faster than the metal can rebuild it, and once that happens the tube corrodes at bare-metal rates rather than protected rates. The attack is concentrated where turbulence is highest, which on a tube bundle is the first few inches inside the inlet, and the result is a distinctive crescent or horseshoe groove rather than general thinning.
The important consequence is that this is a design fault, not a materials fault. Nobody chooses a tube velocity directly; it falls out of the seawater flow divided by the free flow area of one pass. A four-pass unit sees four times the velocity of a single-pass unit of the same frame at the same flow, so pass count is the lever, and selecting the smallest frame that meets the thermal duty often means adding passes and quietly buying an erosion problem.
Two other things push velocity effects harder than the number alone suggests. Suspended sand and silt turn water into an abrasive, so a boat working a shallow silty river erodes tubes at velocities that would be safe offshore. And entrained air, from a leaking seacock, a marginal strainer seal or a partly drawn-down sea chest, causes impingement damage far in excess of what the liquid velocity would predict.
A bundle worn at the tube entries and sound everywhere else is not old, it is fast, and replacing it like for like buys the same failure again.
| EVIDENCE | MECHANISM | WHERE IT APPEARS | WHAT TO CHANGE |
| Crescent groove inside tube entry | Film stripped by entry turbulence | Inlet pass only, first few inches | Pass count or frame size |
| Polished bright bore, thinning wall | General erosion corrosion | Whole inlet pass length | Reduce velocity or change alloy |
| Pitted craters near the entry | Air impingement | Random tubes at the top | Cure suction air leaks |
| Sandblasted matt finish | Abrasive suspended solids | All wetted tube surfaces | Strainer and intake position |
| Wear behind plugged tubes | Flow redistribution | Adjacent tubes in the same pass | Replace the bundle |
Tube-side velocity is the volumetric seawater flow divided by the total cross-sectional area of the tubes carrying it at any instant. Split a bundle into four passes and only a quarter of the tubes carry the flow at a time, so velocity quadruples. That is why the same frame can be perfectly safe as a two-pass unit and marginal as a four-pass one.
It also means the pump matters as much as the exchanger. A raw water pump generously sized for cooling margin, or an engine-driven pump running at full speed while the vessel works at reduced load, delivers more flow than the selection assumed. We would rather size against the real flow and check the pressure drop than trust a nominal figure.
Erosion corrosion, galvanic attack and simple fouling leave different marks, and a few minutes with a torch at the tubesheet usually settles which one you have. Erosion is localised at the inlet, sharply defined, and confined to the pass taking the incoming water. Galvanic attack concentrates at the tube-to-tubesheet joints and is often accompanied by wasted or missing anodes.
Sand abrasion is different again, producing a uniformly matt or sandblasted bore over a longer length. Air impingement makes discrete craters, often in the upper tubes where entrained air collects. Photographing the inlet face and a couple of extracted tubes gives us enough to advise, and it is the first thing we ask for.
The lowest cost fix is usually fewer passes in a physically larger frame, which drops velocity, drops pressure drop, and often costs less to run because the pump has an easier job. Where space forbids a bigger frame, moving from 90/10 to 70/30 copper-nickel buys real velocity headroom for a modest premium on the tube cost.
Titanium removes the velocity limit entirely and is the right call where the flow genuinely cannot be changed, for example an engine-driven pump with no bypass on a fixed installation. That is a legitimate reason to specify it, and it is set out alongside the other cases on the titanium comparison page.
Each grade has an established erosion threshold, with 70/30 tolerating more than 90/10, but the useful answer is a calculation rather than a single figure. We take your seawater flow and the free area of the pass arrangement, work out the actual velocity, and select the frame and pass count that keeps it comfortably below the limit.
Erosion is localised and directional. It appears as a crescent or horseshoe groove in the first few inches of the tubes in the inlet pass, with the rest of the bundle looking sound. Corrosion is more general or concentrated at the tube-to-tubesheet joints, and usually comes with evidence of wasted anodes.
Not if it is the same frame and pass arrangement on the same pump. A two-year life with inlet grooving is a velocity result and it will repeat. Send us the flow figures and photographs of the inlet face and we will select something with fewer passes, a larger frame, or a different tube alloy.
Yes. Below a certain tube velocity sediment settles in the bores, growth establishes, and deposits create the shielded low-oxygen pockets that drive under-deposit corrosion. A grossly oversized unit can foul and pit faster than a correctly sized one, which is why the selection aims for a velocity band rather than a maximum.
Considerably. Entrained air from a leaking seacock, a poor strainer seal or a sea chest drawing down produces impingement attack that appears as discrete craters, often in the upper tubes, and it can do damage well beyond what the liquid velocity would cause. Cure suction side air leaks before replacing a bundle.
Balancing velocity against what your pumps can deliver
+ Learn More
Titanium units where velocity cannot be reduced
+ Learn More
Silty shallow water and the erosion it causes
+ Learn More