Cracked Cast Iron Pump Housing: Identify, Prepare, Decide
A wastewater plant delivers a cast iron volute pump housing with a 6 in crack running from a threaded 1 in NPT gauge port across the machined flange face into the casing wall. The maintenance supervisor wants it welded and back in service Monday; a replacement is six weeks out. The housing has handled sewage and oil-bearing water for 20 years. The tag reads 'cast iron,' but the shop foreman says it might be cast steel because 'they look the same,' and proposes E7018 at 120 A with no preheat, followed by a quick water quench to check for cracks. On hand are E6011, E7018, E309L-16, ENi-CI (nickel) in 3/32 and 1/8 in, an O/A torch, a 400 °F temperature stick, a fire blanket and a grinder. The flange face is a sealing surface and the port thread must remain usable.
1. How would you identify the base metal before committing to a procedure, and why does the identification change everything?
Read the worked answer
Grinding spark test and chip test, done on an inconspicuous edge, will separate grey cast iron from cast steel before any welding.
Spark test: grey iron gives short, dull red sparks that stay close to the wheel with fine forking and no distinct bursts; carbon/cast steel gives longer, brighter straw-to-white streaks with feathery bursts.
Chip test with a cold chisel: grey iron fractures into short brittle flakes with a black graphitic smear; steel produces a continuous ductile chip that curls.
Fracture appearance and the casting/part number on the housing are supporting evidence. Magnetism is not a discriminator - both are magnetic.
Why it matters: for cast steel, a normal low-hydrogen procedure (E7018, preheat per section thickness) applies. For grey cast iron, E7018 with no preheat is close to a guaranteed failure: carbon from the graphite flakes dissolves into the weld metal and the partially melted zone, forming hard white iron and martensite that crack as the weld cools, especially in a thick, restrained housing. Grey iron requires a nickel or nickel-iron electrode, 400–600 °F preheat, short stringer beads, peening and slow cooling, or brazing. The quench the foreman proposed would crack either material.
2. Given 20 years of sewage and oil service, what preparation does the joint need before any electrode is struck?
Read the worked answer
The casting is porous and has absorbed oil and water well beyond the visible crack. Preparation is a sequence, not a single step.
- Degrease with solvent and remove all paint, scale and sludge from both faces and the surrounding 2 in.
- Sweat the oil out with the torch, playing heat over the area until oil stops bleeding from the porosity - around 400–600 °F - and wipe while hot. Welding over trapped oil produces porosity and carbon pick-up at the fusion line.
- Gouge or grind a U-groove through the full crack section to sound metal, and grind the groove faces bright. The graphitized skin left by machining or fracture will not fuse and must be removed mechanically.
- Drill 1/8 in stop holes at each crack end so the crack does not run ahead of the weld.
- Preheat the whole repair zone to 400–600 °F, verified with the temperature stick, and hold it during welding; use ENi-CI in small stringers, peen each bead, and cool slowly under the fire blanket.
Cleaning and groove preparation are what make a cast iron repair hold; electrode choice alone will not.
3. Weld it or replace it? Make and justify a recommendation to the plant supervisor.
Read the worked answer
Recommend against welding as the primary repair, and put the reasoning in writing.
The crack crosses a machined sealing face and a threaded 1 in NPT port. Welding a grey iron housing will distort that face, and no weld restores a thread; the port would have to be built up and re-tapped, and the flange remachined after welding - with a real chance the remachined face still will not seal because the iron around the weld is now a hard, crack-prone partially melted zone. A sewage pump sees vibration, thermal cycling and pressure; a cast iron repair in that location is a leak waiting to happen, and a leak at a gauge port on a wastewater line is both a maintenance and an environmental problem.
Options to put to the supervisor, in order: obtain a replacement housing or a used/spare unit; use a mechanical repair (metal stitching/pinning or a clamped/epoxy patch) if the crack is not through-wall critical and the plant accepts it; or weld only as a documented temporary repair, with ENi-CI, preheat, slow cooling, post-weld remachining of the flange face, re-tapping the port and a hydro test before return to service. The customer owns the risk decision; the shop must not make it silently.