Hand-considered hypotheses, 4001–4100
Steelmaking and ferrous processing, written one at a time. Each entry carries its own logical form, printed above the claim, and the forms are distinct rather than repeated: some assert a bound, some an inversion, some a hidden term, some an artefact that mimics a real effect. The subjects run from burden preparation and the blast furnace through converter and electric-furnace refining, ladle treatment and inclusion control, casting, rolling and heat treatment. None has been tested. Estimates are marked as estimates; no entry cites a paper or quotes a measured value as fact.
Ironmaking: burden, coke and the blast furnace
Aggregate fallacy
H4001. Control of a sinter strand follows the mean exhaust temperature, yet the bed burns through unevenly, so the loop follows the coarse patches and lets the fine ones finish late. Gas short-circuits the permeable patches first and the mean rises when those patches are done, not when the whole bed is. A thermocouple rake across the strand at two strand speeds on one blend would show edge and centre burn-through diverging while the mean holds setpoint.
Hidden term
H4002. Coke's reactivity test heats its sample faster than a blast furnace heats its burden, and the ramp itself cracks pore walls before the gas can attack them. That cracking adds a strength loss the test attributes to the gas. Heating one coke at three rates to the same temperature, with the same hold, should give three losses, and two cokes that reverse rank under different ramps would prove the term real.
Inversion of stated purpose
H4003. Center-coke charging is justified as deadman protection, but in a furnace with a clean deadman its main effect may be to lower the cohesive zone and shorten the dripping path, which can raise the pressure drop. The thinner ore layer at the centre does the work, not the gas distribution at the tuyeres. Two burden distributions matched in centre gas flow but differing in cohesive-zone height, located with shaft probes, separate the two.
Runaway threshold
H4004. Above some zinc input a blast furnace can no longer excrete the metal fast enough, because the condensation zone descends into the cohesive zone and the condensate returns to the hot zone instead of leaving in the dust. Retention then feeds itself. Feeding zinc at increasing rates at constant blast and burden, with the condensation isotherm mapped and dust and slag analysed, should produce a sharp rise in retention rather than a straight line.
Capacity bound
H4005. Alkali leaves a furnace in slag and dust, so the tolerable load is a slag capacity rather than a burden input. Raising basicity at constant alkali input ought to raise the slag's alkali and lower the dust's until the slag saturates. A basicity series on one furnace, with potassium and sodium measured in both streams, would place that saturation as a plateau.
Optimum claim
H4006. Granulation moisture has an optimum that moves with the ore's internal porosity: water first fills the pores and only the excess forms the films that bind granules. A high-porosity ore therefore needs more water than its specific surface alone suggests, and a two-ore blend can have an optimum neither ore shows. Granulating one blend at a series of moistures and measuring green permeability with fired strength at each would put the optimum where pore filling is complete.
Core-shell claim
H4007. Pellet strength follows the core's oxidation state rather than the rim's. The rim converts first and seals the core against further oxygen, so the core's hematite bridges are what finally hold the particles together. Firing two pellet sizes to the same surface temperature, then sectioning them to measure core oxygen against crushing strength, is the direct test.
Reversal
H4008. Hydrogen reduction outpaces carbon monoxide at low conversion, yet above some threshold the order may reverse. Iron formed in hydrogen sinters into a dense layer, while carbon monoxide deposits carbon that keeps the product porous. Reducing one ore in both gases at matched temperature, with rate measured at several conversion levels, would reveal a crossing rather than a constant offset.
Onset claim
H4009. Direct-reduced iron begins to stick at a metallisation that falls as the gas's hydrogen fraction rises. Metallic surfaces weld once oxide coverage drops below a critical fraction, and hydrogen exposes iron earlier than carbon monoxide does. Reducing beds to a series of metallisations in two gas mixtures, then tumbling and screening, should place the onset at different conversions for the two gases.
Hidden variable
H4010. Pressure drop across the stack is a poor witness to the deadman, since the cohesive zone supplies most of the resistance; a clean deadman and a blocked one can read the same at different zone heights. Following the drop against isotherm positions after a burden change separates them, and the deadman should then declare itself in the hearth's temperature. Shaft probes and hearth thermocouples on one campaign would show whether the two ever move together.
Proxy failure
H4011. Silicon serves as a furnace thermometer only while the slag's silica activity stays roughly constant. At high basicity that activity falls faster than temperature raises silicon, so a hotter furnace can tap a lower silicon. Tapping temperature plotted against silicon across a basicity series at one furnace would break the correlation at a basicity readable from the slag.
Identity claim
H4012. Salamander and elephant foot are one wear front at two stages, both starting where hot metal's flow stalls and carbon dissolves into an unsaturated boundary layer. The salamander is the stagnant zone that later freezes; the elephant foot is its eroded rim. Carbon in hot-metal samples from several depths, set against a flow field computed from tap practice, should resolve one front rather than two.
Unit trap
H4013. A percentage loss in the coke test hides the fines a furnace actually receives: a strong coke breaks into a few large pieces and a weak one crumbles, at the same percentage. Initial size also moves the number. Two cokes matched in percentage but not in initial size, with fines below ten millimetres (estimate) weighed, would show whether the ranking survives the change of unit.
Accumulation claim
H4014. Char from high coal injection does not all leave with the top gas; some settles in the deadman. Its particles are smaller than the raceway's voids but dense enough to be carried into the bed, where they lower permeability until the raceway shortens. Injecting coal at two burnout levels while tracking deadman temperature and pressure drop should make the deadman answer to burnout, not to injection rate.
Geometric claim
H4015. A wall channel can form with no gas-distribution bias at all. The burden's angle of repose leaves a thinner ore layer at the wall, and gas prefers that path for purely geometric reasons; the channel is built by charging before gas arrives. A cold model filled without gas flow and then fluidised at a uniform inlet would show the channel in the same place, which removes gas distribution as its cause.
Surrogate failure
H4016. Ferrous oxide predicts sinter strength only along one thermal path. At fixed oxide, strength follows the calcium ferrite's morphology, which changes with basicity and cooling rate, so the oxide is a surrogate for history rather than for the bonding phase. Firing a basicity series to the same oxide level, then measuring strength and ferrite aspect ratio by image analysis, ought to track morphology rather than the oxide.
Nucleation-limited claim
H4017. Pellet swelling counts reduction nuclei rather than the growth rate of the iron they produce. Each nucleus grows a fibrous whisker that pries the pellet apart, so few nuclei mean little swelling even at the same reduction rate. Pellets with and without a surface treatment that suppresses nucleation, reduced at matched rate, should differ in swelling while the rate stays put.
Ontology claim
H4018. The raceway behaves as a fluidised bed, not a cavity. Its boundary is the coke's terminal velocity, and voidage changes gradually rather than at a wall, so a probe's sharp depth records a transition midpoint. A cold model swept over gas velocities, with voidage mapped, would expose that gradual transition and no wall, which changes what a raceway model has to predict.
Mediation claim
H4019. Optical texture may be a bystander in coke's strength after reaction. Texture could act only through pore wall thickness, which the coal's dilation sets during carbonisation; two cokes with similar texture but different wall thickness should then differ in strength. Measuring both on a set of cokes and comparing strength at matched wall thickness separates the two.
Thermal-saturation claim
H4020. Checker brick, not flame temperature, bounds hot-stove blast temperature. The top course reaches a cyclic steady state in which outgoing gas is cooler than the flame, so the brick's stored heat per cycle is the limit. Lengthening the cycle at fixed flame temperature should raise blast temperature until a plateau, and a stove with different checker geometry should settle at a different level.
Converter and electric-furnace steelmaking
Partition claim
H4021. Lance height changes foaming as much as it changes jet momentum, and the decarburisation rate may follow the slag-metal area that foaming creates. Oxygen's partition between carbon and iron is decided at that interface, so a lance change that alters foaming can move the rate without changing impact. One lance height, two slags that foam differently, one oxygen flow: that is the separating experiment.
Sign-change claim
H4022. Phosphorus returns to steel at a temperature that follows the slag's ferrous oxide activity rather than its basicity. The reverse reaction consumes ferrous oxide, and a high-oxide slag buffers the oxygen potential that drives it. Sealed crucibles of slag and steel heated through the suspected range, with both phases analysed, should order the reversal temperatures by oxide content and not by basicity.
Self-limiting feedback
H4023. Foam height in the converter finds its own level. A thicker foam holds the slag longer, lets more ferrous oxide be reduced, and destabilises the film so that the foam drains; the loop settles instead of running away. Blowing at constant oxygen while measuring foam height and slag ferrous oxide at the same time should reveal the height oscillating about a level whose period belongs to the loop, not to lance movement.
Serial bound
H4024. Scrap, not the transformer, bounds electric-furnace power. The pile's voids screen the arc and heat transfer falls as the pile flattens, so two charges of different bulk density should need different power-on times at the same power. If the times follow bulk density and not the transformer's limit, the scrap is the bound.
Nucleation switch
H4025. The decarburisation switch from oxygen transport to carbon transport is a nucleation event, not a bulk-carbon threshold. Below a critical supersaturation, bubbles nucleate only at the slag-metal interface and the area collapses. Blowing at two total pressures with the same oxygen flow should move the switch, because the critical supersaturation depends on ambient pressure.
Stoichiometric bound
H4026. Chromium sets the carbon floor in argon-oxygen decarburisation, not the argon's dilution. Below a critical carbon, chromium's activity buffers the oxygen potential and decarburisation stalls, so a heat with more chromium should stall at a higher carbon at the same argon flow. Decarburising two chromium levels to the stall, with carbon and chromium sampled together, tests the bound.
Exponent claim
H4027. Vacuum decarburisation may follow the square root of pumping speed rather than pumping speed itself. The rate would then be set by growth of the carbon monoxide layer at the free surface, whose thickness scales with the diffusion length. Two vessels with pumping speeds in a known ratio, decarburising the same steel from the same carbon, should give rates in the ratio of the square roots if the exponent is real.
Equilibrium-gap claim
H4028. Turndown oxygen sits above the carbon-oxygen equilibrium because the slag's ferrous oxide is not equilibrated with the bulk metal, and the gap grows with the slag's mass. Comparing measured oxygen with the equilibrium value across heats of different slag volume, at one carbon and temperature, should make the excess follow slag mass rather than carbon.
Heat-recovery claim
H4029. Silicon's oxidation heat escapes the electric furnace in the slag. The silica enters the slag and leaves with the tap, so the apparent heat yield depends on basicity. Two heats with equal silicon input but different basicity should differ in heating rate per unit oxygen, and a heat balance on off-gas, slag and bath would show where the heat went.
Freeboard threshold
H4030. Slopping starts at a foam height that clears the vessel's freeboard by a margin set by lance height. Film drainage fixes the foam's collapse rate, and the jet breaks foam only within its reach. Raising the lance in steps at constant blowing rate and recording the first slop should map a boundary in the lance-height and foam-height plane, not one critical rate.
Surface-blocking claim
H4031. Surface oxygen, not nitrogen diffusion, slows nitrogen desorption. Oxygen occupies the sites where nitrogen molecules recombine, so a fully killed heat should degas faster than a rimmed one at the same initial nitrogen and pressure. Nitrogen decay after a vacuum step, in heats of different aluminium content, separates the surface effect from diffusion.
Vortex-scaling claim
H4032. A vortex, not slag viscosity, decides carryover at tapping. Its formation depends on tap-hole diameter and bath depth, so a deeper bath carries less slag at the same tap time. Two heats with equal slag volume but different bath depths, tapped at a constant rate, ought to order carryover by the depth-to-diameter ratio.
Droplet-population claim
H4033. Droplet population, not oxygen flow, sets the slag's ferrous oxide at turndown. Iron oxidises on droplet surfaces, and a finer emulsion oxidises more iron for the same oxygen, so two lance designs at one flow should give different ferrous oxide if their droplets differ. Quenched slag samples, with droplet size and ferrous oxide measured together, test the link.
Dissolution-limited claim
H4034. Lime dissolves through slag-side mass transfer, not through its own calcination. A soft-burned lime with open pores dissolves faster and lifts basicity earlier, and early basicity rather than the final value controls dephosphorisation. Two limes of different porosity, added at the same size and mass, with slag basicity measured against time, ought to differ.
Saturation claim
H4035. Phosphate capacity saturates, and lime beyond saturation buys nothing. Dicalcium silicate begins to precipitate and the phosphate's activity coefficient rises, so the partition should plateau and then fall. A basicity series at fixed temperature and ferrous oxide, with both phases analysed, locates the saturation.
Post-combustion mixing claim
H4036. Jet entrainment of carbon monoxide, not oxygen flow, sets the post-combustion ratio. A higher lance mixes with more of the vessel's gas and burns more carbon monoxide, so two lance heights at one flow should give different off-gas ratios. Off-gas analysis at both heights tests the mixing explanation.
Surface-area claim
H4037. Scrap surface area and slag superheat, not total oxygen, control melting in the converter. Light scrap melts earlier and cools the bath more sharply, so two scrap mixes of equal mass but different bulk density should give different temperature trajectories at the same blow. A bath thermocouple and a scrap weighbridge would expose the difference.
Radiation-shield claim
H4038. Foam in the electric furnace shields the wall; arc stability is a secondary benefit. The foam absorbs the arc's radiation and lowers the heat flux, so two foaming heights at the same arc power should give different wall fluxes. A panel meter on the wall at both heights, with power and tap weight matched, separates shielding from stability.
Ladle refining, degassing and inclusion control
Nucleation-site claim
H4039. Bubble residence time limits hydrogen removal, and open pores at the refractory set that time. A clean lining should degas faster than one coated with solidified slag at the same pressure. Hydrogen decay after a pressure step on two ladles of different coating, at matched circulation, would show the rate following the surface rather than the pressure.
Droplet-area claim
H4040. Droplet area in the vacuum's free space, not circulation, controls the degasser's decarburisation. Gas lift and vessel height set that area, so two campaigns with the same circulation but different snorkel immersion should differ. A deliberate change of immersion between campaigns, with circulation held, is the whole experiment.
Stream-area claim
H4041. Nitrogen pickup follows the stream's area per unit mass, not the tapping time. Absorption is interfacial, and the stream's area per unit mass grows as it narrows. One heat tapped through two nozzle diameters at the same tap time, with the same nitrogen in the alloy, should place pickup with computed stream area, not with time.
Source apportionment
H4042. Manganese and zirconium tag the two reoxidation sources: slag-derived inclusions carry manganese, refractory-derived ones carry zirconium. Casting one heat with a slag of known oxide content and a nozzle of known zirconia content, then typing inclusions by those elements, apportions the oxygen without a plant trial.
Degeneracy claim
H4043. The calcium-to-aluminium ratio is degenerate. The same ratio can be reached with different oxygen and sulphur, which also consume calcium, so heats matched in ratio but not in sulphur ought to modify differently. Plotting modified fraction against ratio and sulphur together would reveal which variable carries the effect.
Stratification claim
H4044. A solid slag cap, not steel buoyancy, fixes ladle stratification. The cap suppresses natural convection, so a heat with a solid cap holds a hotter bottom longer. Two ladles of equal superheat but different slag solidification temperatures, fitted with a thermocouple array, should stratify differently.
Residence-time claim
H4045. Mean residence time and dead volume are different variables. Flow rate sets the mean; the weir's geometry sets the dead volume, so two tundishes can share a mean and differ in dead volume. Tracer tests at two flow rates with one weir, computing both measures, would show which is the geometric invariant.
Product rule
H4046. Capacity and area enter sulphur removal as a product. A slag with twice the sulphide capacity but half the interfacial area removes at the same rate, so trials that vary one and not the other mislead. Two slags with capacities in a known ratio but different melting behaviour, stirred at the same power, would reveal which term controls.
Common-cause claim
H4047. Stirring and cleanliness share superheat as a common cause: a hotter heat needs more stirring and carries more inclusions. That alone would produce the correlation without any causal link. Selecting heats within a narrow superheat band and comparing cleanliness at two stirring powers should flatten it.
Aggregation claim
H4048. Alumina grows by collisions, not by nucleation. Nucleation supplies many small particles; Brownian and turbulent collisions build the large ones, so the large-particle fraction should follow stirring energy rather than aluminium addition rate. Two heats deoxidised at different rates but stirred alike should collapse onto one population curve.
Wettability claim
H4049. Contact angle at the bubble, not bubble size, limits flotation of small alumina. A particle that is not wetted enough cannot attach, and sulphur adsorbs at the interface and changes the angle. A high-sulphur heat should therefore remove inclusions more slowly at the same gas flow.
Competitive-reaction claim
H4050. Calcium spends itself on sulphur first. Calcium sulphide forms before the oxide is touched, and only excess calcium modifies spinel, so a high-sulphur heat needs more calcium for the same result. Two heats with equal calcium but different sulphur, with inclusions typed, should put modification with excess calcium.
Bimodal claim
H4051. Total oxygen and large-inclusion counts are separate populations. Large inclusions come from reoxidation during casting while total oxygen is set in the ladle, so two heats with equal total oxygen can differ by an order of magnitude in large inclusions (estimate). The large count should change across the tundish while total oxygen holds, which sampling both locations would reveal.
Recirculation claim
H4052. Clogging grows where recirculation delivers alumina to the wall, not where the wall is coldest. Particle arrival rate and wall shear decide the deposit, so a nozzle with a changed bore should clog elsewhere with no change in steel chemistry. One heat through two bores, with the deposit mapped, tests the recirculation explanation.
Solid-fraction claim
H4053. Solid fraction at the nozzle's temperature, not total inclusion load, decides clogging. Liquid calcium aluminate does not adhere; solid inclusions do. A heat whose inclusions are liquid at casting temperature should not clog even at high oxygen, which a comparison of two heats at different calcium-to-aluminium ratios would show.
Sectioning bias
H4054. Polishing removes soft inclusions before the microscope sees them. Relief hides them in the focal plane, so a polished section under-reports the soft population. One sample measured both by polished section and by electrolytic extraction exposes the deficit, which should grow with softness.
Thermodynamic-path claim
H4055. Slag basicity, not aluminium total, fixes silica reduction in the ladle. The slag sets silica activity and therefore the oxygen potential the aluminium sees, so a high-basicity slag should suppress reduction even at high aluminium. Two heats with equal aluminium but different basicity, with silicon pickup measured, should follow basicity.
Shape claim
H4056. Cluster shape, not mass, controls bubble attachment. A cluster has a longer contact line per unit mass and attaches more readily, so clusters leave faster than spheres of equal mass. Two populations of equal mass but different morphology in one stirred vessel should remove the clusters first.
Provenance claim
H4057. Zirconia in the inclusions points to the refractory, not to slag equilibrium. The inclusions carry zirconium from the nozzle while the slag's composition would predict a different alumina-to-silica ratio. Inclusions, slag and refractory from one heat cast through a new and an old nozzle should apportion the sources without a tracer.
Area claim
H4058. Slag-eye area, not gas flow, controls reoxidation. The eye's area is set by gas flow and slag thickness, and two stirring rates that produce the same area should produce the same reoxidation. Matching eye area by adjusting slag thickness at two gas flows, then measuring nitrogen pickup, tests the claim.
Continuous and ingot casting
Cooling-rate dependence
H4059. Break temperature belongs to the cooling rate as much as to the flux. The same flux can break hotter in a plant mould than in a laboratory crucible, so one flux measured at three cooling rates should give three values. That would make the quoted break temperature a property of the measurement as well as of the flux.
Decoupling claim
H4060. Mark pitch and mark depth answer to different controls. Pitch follows negative strip time; depth follows the flux's crystallisation, so two moulds with the same pitch can produce different depths. One steel cast at two oscillation settings chosen to hold negative strip time constant, with the flux changed, separates the two.
Misattribution
H4061. Mould friction, not secondary cooling, can thin the shell before the sprays are reached. A taper slightly too small makes the mould do the thinning. One heat through two tapers with secondary cooling held identical, and shell thickness measured at the mould's exit, separates taper from cooling if friction is the cause.
Precursor claim
H4062. Heat flux falls before the friction alarm on a sticker. Flux infiltration stops when the shell's shrinkage opens a gap at the meniscus, and the loss of flux shows in the mould's thermocouples. Replaying records for stickers and matched non-stickers should put the flux drop ahead of the alarm; the lead time is the experiment's output.
Process-window claim
H4063. Soft reduction has a window, not a monotonic benefit. The reduction must land where solid fraction is high enough to feed the centre but low enough for the shell to deform. One steel with reduction rolls at three positions and total reduction held constant should put the segregation minimum at one roll position.
Thermal-mixing claim
H4064. Stirring may refine structure by mixing heat rather than by breaking dendrites. Electromagnetic stirring flattens the strand's temperature field as it stirs, so the same benefit could arrive without fragmentation. Two frequencies with equal flow but different superheat, with structure and segregation both measured, separate the thermal effect from the mechanical one.
Interfacial-layer claim
H4065. The crystalline layer's thickness, not the flux's bulk properties, controls the shell's heat flux. Break temperature and mould extraction fix that thickness, and a thin layer passes more heat than a thick one, whatever the flux's viscosity. Two fluxes of different break temperature in one mould should give heat fluxes ordered by layer thickness, not by viscosity.
Inversion
H4066. Argon shrouding reverses above a flow. Bubbles coalesce and the gas layer breaks, so more argon can mean more clogging. Increasing argon in steps at one casting speed while measuring deposit growth should expose a minimum rather than a monotonic improvement.
Lubrication limit
H4067. Taper's optimum belongs to the flux's lubrication, not to the shell's shrinkage. Too much taper raises the friction the flux must carry; too little lets the shell detach. A taper series on one steel with one flux, reading friction force and shell surface, should put the optimum where the film is thinnest but unbroken.
Aspect-ratio claim
H4068. Ingot aspect ratio, not exothermic powder heat, controls hot-top feeding. Feeding is a gravity drain whose path lengthens with ingot height. Two ingots with equal hot-top volume but different heights, sectioned for pipe depth, should make pipe depth follow height, not powder mass.
Initiation-site claim
H4069. Plate cracks start at the slag rim, not at the hot face's centre. The rim imposes a thermal cycle shorter than the oscillation period where it touches the copper. Crack initiation mapped on a used plate, and compared between plates run with different rim contact, should track the rim.
Film-boiling instability
H4070. Corner cracks follow an unstable vapour film. The corner's geometry lets the film collapse and rewet in cycles, and the thermal cycling strains the shell. High-frequency corner temperatures at two water fluxes should reveal the oscillation amplitude falling as flux rises, with cracks following amplitude rather than mean temperature.
Path dependence
H4071. Width transitions leave a crack trail that remembers speed. Strain accumulates while the mould's taper lies between the two widths, so two transitions to the same final width at different speeds should leave cracks in different places. Mapping crack position against transition speed tests the path.
Inventory drift
H4072. Tundish slag drifts as the refractory dissolves, and the drift is an inventory effect. Dissolution adds silica during a sequence, so the last slab should carry more inclusions than the first. Slag sampled at the start and end of a sequence, with inclusions measured in the first and last slabs, should make the two track.
Causal ordering
H4073. The hook comes before the crack. Flux's carbon film is entrained at the hook, and the crack nucleates at the hook's root under bending strain, so the hook is a cause rather than a scar. Slabs cast with fluxes of different carbon content, with hooks and cracks mapped on the same sections, ought to confine cracks to hooks deeper than a threshold.
Entrainment threshold
H4074. A vortex air core, not slag depth, sets entrainment at the tundish nozzle. The core forms when the outlet's Froude number passes a value set by nozzle diameter, so two diameters at one slag depth should entrain at different heights. Water-model tests would show the critical height scaling with diameter.
Impingement claim
H4075. Jet angle, not casting speed, drives mould-level fluctuation. Penetration scales with the angle's sine and the fluctuation follows impingement, so two speeds with one nozzle and one speed with two angles should separate the variables. The second pair is the test.
Breakup-length claim
H4076. Stream breakup, not nominal continuity, starts reoxidation. The open stream breaks into drops at a length set by viscosity and nozzle diameter, and the drops entrain air even when the stream looks solid. Oxygen pickup measured along the stream at two diameters, with breakup length photographed, should begin at the breakup point.
Hot and cold rolling
Phase-lubrication claim
H4077. Scale phase, not scale thickness, may set hot-rolling friction. Wüstite is ductile at the roll bite's temperature and magnetite is not, so a thin wüstite scale can lubricate better than a thick mixed one. Two descaling practices giving different phase mixes at equal thickness, with forward slip measured, separate thickness from phase.
Coupling claim
H4078. Crown and edge drop answer to different roll terms. Bending sets crown; wear sets edge drop, so a mill can hold one and lose the other. Profile measured at two bending settings with wear fixed, then at two wear states with bending fixed, should separate them.
Confounding claim
H4079. Forward slip carries a tension term. Tension changes the roll's elastic flattening, so a tension sweep at constant reduction and speed should move the inferred friction even with the roll's surface unchanged. That makes the inference confounded, not merely noisy.
Lag claim
H4080. The coiler reads a temperature the coil's centre has already left behind. The centre keeps transforming after coiling, so the coiler's reading is not the temperature that sets properties. A centre thermocouple during cooling, with properties compared at two coiling temperatures and one centre hold, should put properties with the hold rather than the reading.
Non-linearity claim
H4081. Water flux and temperature drop part company above film boiling. Once the surface films, the heat-transfer coefficient falls, so doubling the water can add less cooling than expected. Strip temperature at three points along the table at two fluxes reveals the drop saturating.
Scale argument
H4082. Orange peel tracks grain size relative to sheet thickness, not roll roughness. A fine-grained sheet stays smooth even under a rough roll. Two sheets rolled to equal roughness but different grain size, with surface waviness measured, should make waviness follow grain size.
Buckling-wavelength claim
H4083. A flatness wave is a buckling wavelength, set by strip thickness and the residual stress's gradient. The same stress can give edge waves in one thickness and centre waves in another. Two thicknesses rolled to the same measured stress profile, with flatness mapped, should move the wave's location with thickness.
Dimensionality claim
H4084. Edge shear makes width spread three-dimensional. A formula calibrated on wide strip overpredicts narrow strip because it omits that shear. Two widths at equal reduction, with spread measured at the edge, should make the coefficient follow width-to-thickness ratio.
Model-failure claim
H4085. Roll-force formulas assume a friction that high strain rates remove. The roll's surface heats and the scale softens as rate rises, so measured force should fall below prediction. A strain-rate series at constant reduction is the test.
Neutral-point claim
H4086. Flow stress moves the neutral point. Forward slip is read as friction, but a stronger steel shifts the neutral point without any change at the roll's surface. Two steels of equal surface and different flow stress, rolled at one reduction, would reveal the shift.
Wear-mapping claim
H4087. Scale phase maps onto roll wear. Harder scale abrades the roll more, so a strip with mixed scale should wear the roll in bands that follow the phase. Rolling strips of different phase mixes, with the roll's profile measured after each campaign, should map the bands.
Radial-conduction claim
H4088. Radial conduction, not coiler water, cools the coil. Heat leaves through the wrap, so the edge cools faster than the centre and properties vary across the width even at uniform coiling temperature. Three radial thermocouples during cooling ought to record the edge cooling faster than the centre.
Emulsion-aging claim
H4089. Emulsion fines degrade the cleaning performance. Oil concentration can hold while iron fines accumulate, so a bath at the correct concentration still leaves residue. Strips rolled at the start and end of an emulsion's life, at equal concentration, should leave more residue at the end.
Spall-depth claim
H4090. Spall depth follows case depth. The spall runs along the case-core interface, so a deeper case gives a deeper spall at the same thermal fatigue. Two rolls of different case depth under one mill load, with spall depth measured, should put spall depth with case depth.
Heat treatment and quenching
Measurement-load claim
H4091. Hardness load shortens the case depth it measures. A higher load samples into the softer core, so the case appears shallower. One carburised sample at three loads, compared with the carbon profile from a companion, would reveal a load effect larger than process scatter.
Gas-surface decoupling
H4092. Shim carbon and surface carbon part company. The shim equilibrates with the gas while the steel's surface also loses carbon to soot, so a shim can read a potential the surface never sees. One furnace at two gas flows with one setpoint, both carbons measured, separates the gas reading from the surface reading.
Delamination claim
H4093. Cooling rate, not layer thickness, spalls the white layer. Layer and substrate expand differently, and a fast cool puts the layer in tension. Two nitrided samples of equal layer thickness cooled at different rates, with residual stress and spall count measured, separate the two.
Orientation claim
H4094. Quench orientation fixes distortion direction. The vapour blanket's asymmetry sets the thermal gradient, and the gradient sets the distortion, so identical parts quenched in two orientations should distort in opposite directions. That reversal is the test.
Boundary-chemistry claim
H4095. Boundary chemistry, not tempering temperature, embrittles. Embrittling elements segregate to boundaries established at austenitising, so a steel with more boundary area should embrittle less at equal phosphorus. One steel austenitised at two temperatures to two boundary densities, then tempered alike, should order embrittlement by boundary area.
Timing claim
H4096. The quench delay tempers the transition zone. Heat conducts inward between the heating's end and the quench's start, softening the zone behind the case. Two delays at one frequency and power, with the hardness profile mapped, should move the softened band with the delay.
View-factor claim
H4097. View factor, not gas flow, leaves the cold spot. The load's interior sees the elements through narrow channels, and convection is too weak to close the difference. Identical blocks at centre and surface, run with and without a forced-gas fan, should leave the centre's lag intact with the fan on.
Creep claim
H4098. Load weight creeps the bottom of a carburising stack. At austenitising temperature the steel is weak enough for the stack's own weight to flatten the lowest parts, so distortion should follow stack height rather than thermal gradient. One tray loaded at two heights, with bottom-part ovality measured, tests it.
Hold-time claim
H4099. Cryogenic wear resistance may come from carbides, not austenite. The long hold precipitates fine carbides as well as transforming austenite, and a short deep freeze transforms without the long hold. Quenched, short-frozen and long-frozen samples of one steel, with wear and austenite both measured, should decide which one wear follows.
Geometric-blanket claim
H4100. Vapour blanket lifetime follows part geometry. Vapour escapes most easily from upward-facing surfaces, so a concave face holds its blanket longer. Two parts of one alloy with different cavity orientations quenched in one oil, with hardness measured at the cavity's base, should make blanket lifetime follow geometry.
Status and plan
100 hypotheses, one logical form each. They were written individually rather than generated from a grid, and the form label is printed so that a reader can see the reasoning moves differ structurally and not only topically. The subjects are confined to ironmaking, steelmaking, casting, rolling, heat treatment and inclusion control, and to the physical and chemical mechanisms that connect them.
None has been tested here. A hypothesis list is worth the ideas it makes available; a test is a separate piece of work, and several of these could be settled with records a plant already keeps.
If one of these is worth a plant trial, it is H4062 — that the sticker breakout’s heat-flux drop precedes the friction alarm. It needs only mould thermocouple records that are already logged, and the lead time it yields is the experiment.
Incoprea Research · steelmaking and ferrous processing · hand-considered hypotheses 4001–4100 · 3 November 2026 · ~/Desktop/Nov3/