Two components can carry the same Cr26 analysis and still differ by a factor of two in service life. The chemistry is usually not the reason.

Conversations about slurry pumps tend to start with metallurgy: chromium content, hardness in HRC, whether the iron is eutectic or hypereutectic. Those numbers matter. What they leave out is the part of the story that actually separates one supplier from another. How the carbides were formed, whether feeding reached the thick sections, how much residual stress is locked into an impeller before it ever sees slurry, and whether anyone can trace a finished liner back to the heat it came from.

This note walks through the route in the order it happens on the floor. Some of the practices described below are documented in public sources and widely used across the industry; none of them requires exotic equipment, but each one only works if somebody owns it and records it.

Figure 1　Slurry pump wet-end casting route (parameters shown are typical working ranges; refer to the process specification and test reports for any given part)

1. Start with the charge

High chromium white iron is usually melted in a coreless induction furnace. Induction melting has one trait worth stating plainly: it does not remove phosphorus or sulphur. Whatever goes into the crucible stays there. Foundries that run clean, low-residual charge materials get a clean melt; those that hope to fix contamination later generally cannot. This is why serious control starts with purchasing rather than with flux additions at the furnace.

Charging order matters too. Ferromolybdenum and part of the high-carbon ferrochrome go down first, then scrap steel packed firmly at the bottom and loosely on top so it collapses into the melt as it softens. Chromium oxidation losses run roughly 5 to 15%, so the bulk of the ferrochrome goes in late, once the scrap is fully molten, and usually preheated. Charging it all at the start means losing chromium before the bath is even formed.

Temperature follows a fairly consistent pattern in practice: melting held around 1450 to 1480 °C, power reduced once the charge is down to limit gas pick-up and oxidation, then a short superheat to 1480 to 1520 °C held for three to five minutes so composition evens out and inclusions can float, followed by power off and slag removal. Too little superheat leaves the bath inhomogeneous; too much accelerates the loss of carbon and chromium and costs more than it returns.

A spectrometer reading before tap is the gate. Out of specification means no tap. Simple, and it is where most variation is stopped.

2. Deoxidation and inoculation

Once everything is molten, ferromanganese and ferrosilicon go in as a first deoxidation step, followed by aluminium for the final kill. Aluminium is typically metered at about 0.08 to 0.1%, which works out near one kilogram per tonne, and the heat is tapped soon afterwards so re-oxidation has less time to work.

The step that shapes the carbides is inoculation in the ladle. The ladle is preheated to 400 to 600 °C, the modifier is placed in the bottom, and the iron is poured onto it. A bonding agent is then added so residual slag collects quickly and forms a covering layer, and the metal is allowed to stand for two or three minutes before casting.

Why bother? Left to themselves, the carbides in high chromium iron tend to precipitate as a continuous network. In effect you have buried a brittle skeleton through the matrix, and impact cracks travel along it: the part comes out hard and unforgiving. Rare-earth based modifiers work against that in three ways, which is well supported in the published literature and in production practice:

• rare earths segregate to grain boundaries and interrupt continuous carbide precipitation along them, so the network breaks up;
• the increased lattice distortion and vacancy density help carbon diffuse, which favours compact, rounded carbides over skeletal ones, and distributes them more evenly;
• cleaner, lower-energy grain boundaries raise impact toughness, with a useful gain in hardness alongside it.

Carbide refinement. A practice widely reported for hypereutectic alloys is to hold primary carbides below about 50 µm. High carbide volume is what gives excellent abrasion resistance, but coarse carbides leave too little toughness for impact and the part tends to fail by chipping. The control levers are inoculation combined with solidification rate, rather than alloy additions alone.

3. Moulding: pick the route per component

One assumption worth dropping is that a foundry runs a single process. A casing, a volute liner and a closed impeller have very different requirements for passage accuracy, batch size and wall thickness, so they usually travel different routes.

Lost foam pouring is commonly run with about 0.02 MPa of negative pressure, pulling metal into the cavity and drawing decomposition products away at the same time. It looks like a small number, but it has a direct influence on filling and on gas defects, which is why most plants trend it continuously rather than reading it off a gauge once per shift.

4. The digital layer

This is where foundry work has shifted most over the past two decades.

4.1 Solidification simulation

Where to put feeders, whether chills are needed, where the gates belong: these used to be settled by experience and repeated trial pours. The common approach now is to build the model first and run MAGMA, ProCAST or FLOW-3D over filling and solidification, mark the zones likely to concentrate shrinkage, then revise the running and feeding system against that picture. The trial-and-error cost moves from the floor onto the screen, and process sign-off typically comes through in about half the time with a visible gain in yield.

4.2 Printed sand

Binder jetting prints moulds and cores straight from CAD, with no pattern or core box, and the benefit shows up clearest where several cores used to be glued into one assembly. Public cases worth knowing:

• a water-technology manufacturer built an impeller core from four bonded segments, taking seven days; printed as a single core it was ready in two, at roughly 30% less cost on some codes;
• for one foundry customer, the improved dimensional repeatability collapsed grinding and balancing from two operations into one, taking machining from 2 hours 36 minutes down to 49 minutes per impeller;
• a multi-alloy foundry reports that over 90% of its impeller cores are now printed as single pieces, roughly halving lead time.

For slurry pumps the payoff sits in closed impeller passages. The passage surface carries the hydraulic performance: every joint in a glued core assembly is a chance for the vane outlet angle and the throat geometry to drift.

5. Pouring temperature and cooling

Pouring temperature commonly falls between 1380 and 1420 °C, chosen against section thickness and geometry, with thin-walled parts towards the top of that band and heavy section work towards the bottom. Too cold and you get metal that has already precipitated too much secondary carbide, with misruns and cold shuts; too hot brings coarse grains, higher shrinkage porosity, burn-on and more distortion.

Cooling should aim for directional solidification: the metal furthest from the feeder freezes first, the feeder freezes last, and the feeding path stays open throughout. Three levers are in regular use: chills positioned to pull heat out of local hot spots, padding added over heavy sections that would otherwise starve, and tilt pouring, where the flask is raised so metal enters at the low end and fills progressively. Plants running tilt pouring often work at 8 to 20 degrees, and report less shrinkage and better dimensional control.

Covering the filled feeder with an insulating or exothermic compound keeps it liquid longer, which is a small spend for a solid improvement in feeding.

6. Cleaning and inspection

High chromium iron cracks easily, so cleaning has its own rules. Feeder heads are not knocked off with a sledgehammer; heavy sections are preheated before cutting and go into stress relief or the specified hold afterwards, because steep gradients are what start the cracks. Grinding needs the same care in miniature: grinding burn is invisible on the surface and becomes a crack initiation site once the pump is running.

Inspection normally covers:

• spectrometer analysis at the furnace, recorded against the heat;
• hardness traverse through the section, surface and core, to see how uniform it really is;
• ultrasonic, magnetic particle or dye penetrant, selected to match how critical the part is;
• dimensional checks, including vane outlet width and blade-to-blade consistency on impellers.

Worth asking for. A single hardness figure is one point on one casting. What tells you something about the batch is the spread across multiple points, together with metallographic images from the test block cast alongside it. Ask for those two things rather than the headline number on a datasheet.

7. Traceability through the heat number

A disciplined foundry threads the heat number through everything: which raw material batches went in, charging sequence, power input and temperatures, the spectrometer result, which ladle and which mould received it, who ran shakeout and heat treatment, and finally the NDT report. Where the infrastructure allows, castings are laser-marked or tagged with a 2D code at shakeout and scanned at each step.

Two things come out of it. When something does go wrong, you narrow it to a heat and a step in minutes rather than turning the plant over. And over time the accumulated record starts to show relationships between parameter drift and defect types, which is what makes improvement possible.

8. Inside the EXCELLENCE PUMP foundry

All of the above comes down to one question for any supplier: which of these stages do you run yourself. Buying in castings is a legitimate route, but you can only control process detail you can see:

• casting and heat treatment in house, so wetted parts never change hands and every parameter from composition to quench and temper sits in our own documentation;
• our own foundry process and furnace-side discipline, with spectrometer readings, inoculation practice and pouring temperature recorded per heat;
• machining and assembly under the same roof, so cleaned castings go straight onto the line instead of accumulating damage in transit;
• a performance test loop on site, so whether a pump meets its design duty is measured rather than assumed;
• ISO 9001 certified process control and record retention;
• supplying domestic and export markets from the same process system.