FIELD GUIDE / material selection
The final plain piece follows an excellent material choice into the moment it must become plate, tube, casting, forging, wire, or powder. It is where a selection chart meets tolerances, directionality, availability, and the actual factory.
Selection begins badly whenever a material name appears before the component it must become.
A chart can make that error feel disciplined.
It offers density, modulus, strength, temperature limit and price columns. It sorts them with admirable speed. It returns a material which appears to have won.
The purchase order is then asked to supply the winner as plate, tube, casting, forging, wire or powder, at the required dimensions, in a condition which will survive the actual route.
That last step is not administration.
It is where the material selection is finally tested against the component.
A design-oriented method separates requirement definition, screening, ranking, investigation of the shortlisted candidates, and the contextual constraints.
The separation matters because each stage answers a different question.
A database can identify a plausible material class.
It cannot establish whether a wall can be held, a bore can be controlled, a weld can be examined, a powder can spread, or a supplier can deliver a qualified lot.
A selection brief starts with function, interfaces, service scenarios, design life, expected volume, decision owner and a comparison basis.
The brief must give quantities and units.
A bracket which carries 8 kN, stays below 1.2 kg, operates from minus thirty to ninety degrees, and completes two hundred thousand duty cycles is a real screening problem.
“Lightweight and durable” is a preference waiting to become one.
Geometry belongs in that first brief, even where the design is still preliminary.
A fixed envelope with a tight positional allowance can rule out whole routes before the analyst has built a detailed model.
So can a fixed wall thickness, a small clearance, a sealing land, a bore, or an inaccessible inspection face.
A property chart assumes that a section can be resized.
The component very often says that it cannot.
The same is true of time.
A half-hour overload and a ten-year static exposure demand different evidence.
A cyclic requirement needs the stress range, the mean stress, the cycle count, the frequency and the spectrum.
A million cycles under a fully reversed stress ratio is not interchangeable with a million cycles under a tensile mean.
If creep governs, the relevant statement is a permitted strain after a stated number of hours at a stated temperature.
It is not a room-temperature tensile value.
And the product form must be able to carry the section and the condition on which that evidence rests.
The brief should also distinguish a hard constraint from an objective, and both from an unknown.
A candidate which cannot meet the continuous service temperature does not need an elaborate cost model.
A candidate which passes a preliminary temperature screen can still fail a stress limit, once the actual section has been fixed.
Geometry, material and process are tightly coupled, and a material choice can quietly preclude a great many manufacturing processes.
Plate, sheet, bar, tube, wire, casting, forging and powder expose different geometry to load, tooling, defects and inspection.
Calling them versions of the same material loses the information which makes a component possible.
The grade may remain constant while the form changes the directionality, the surface state, the tolerance field, the residual stress, the defect population, the joining access, and the amount of machining required.
Plate begins as a slab or sheet ingot and is hot rolled into a flat rectangular product, generally supplied in individual lengths.
The aluminium trade boundary sits at about six and a third millimetres, which explains its practical character. Plate is handled, cut, machined and formed as a discrete blank.
Steel definitions run plate from a quarter of an inch to well over a foot.
Neither figure is a universal material rule.
Both show why plate is not merely sheet which has become heavier.
Its thickness supports deep machining allowance, bolted bearing surfaces and heavy-section stiffness.
It also makes through-thickness temperature and residual-stress variation into material concerns.
Rolling leaves a longitudinal and a transverse direction. Mill scale, or a required surface preparation, can determine the next operation entirely.
So a plate quotation needs thickness, width, length, edge condition, flatness, camber, surface condition, rolling direction and mass.
The word plate has supplied none of them.
Sheet has the same rolled ancestry, reduced far enough to be efficiently coiled.
It continues through a continuous mill, enters coil, and receives one or several cold-reduction passes. The boundary between sheet and foil sits around a sixth of a millimetre.
Cold work raises dislocation density and can increase strength while reducing ductility, unless annealing or another heat treatment changes the condition.
A sheet candidate must therefore carry its condition as well as its nominal alloy.
For sheet, flatness, coil set, edge quality, surface finish and directional forming response can outweigh a bulk strength value entirely.
A record reading “3.0 mm sheet,” with no thickness tolerance and no sampling plan, cannot show that it fits a die or retains a required minimum section.
A test certificate with one reading is evidence of one reading, unless it also names the method, the instrument resolution and the sampled location.
Bar is solid stock, round or square or flat or hexagonal, made for machining, threading, bending or upsetting while retaining a stated straightness and surface condition.
It may be hot rolled, cold finished, extruded, peeled or drawn.
Its solid centre makes axial and torsional members straightforward.
That same centre adds mass, and turns an internal passage into a machining problem.
Tube places the material around a bore instead.
It can give stiffness, and pressure or fluid capability, per unit mass, by moving material away from the centroid.
It introduces ovality, collapse, local buckling, bend flattening and a seam question in exchange.
The record needs outside diameter, wall thickness, inside diameter, straightness, ovality, eccentricity and seam type.
For a circular tube the inside diameter is the outside diameter less twice the wall, which is a piece of arithmetic anybody can do.
It does not establish a measured wall minimum, which is the number that matters.
Route changes the remaining question.
Drawn-over-mandrel tube uses a drawbench, a die and an internal mandrel, and its advantage is control of inside diameter, wall thickness, inside and outside surface, and gauge quality.
Welded tube begins as strip, formed into a cylinder and joined along its length.
The seam is therefore a location, an inspection target, and sometimes a forming limitation.
It is not enough that the tube falls within an outside-diameter range.
Wire is a different long-product problem again.
Rod may be continuous-cast or rolled redraw stock. Drawing reduces it through successively smaller dies, establishing diameter and surface finish while cold working it.
A nick which is negligible on bar can control fatigue or breakage in fine wire.
A single one-millimetre strand has a calculated area of about eight-tenths of a square millimetre, and a strand count alone still does not establish a cable area, because the gaps and the individual tolerances remain.
Bend radius, coating, tensile condition, conductivity, strand construction and winding behaviour all belong in the form requirement.
Casting supplies the greatest primary freedom for enclosed passages, bosses, ribs and complex contours.
Its sequence is melt, prepare mould, fill, separate, cool, post-process.
The same freedom brings shrinkage feeding, porosity, inclusions, hot tears and local variation, because the part records how the metal filled and froze.
A pattern is deliberately larger than the required casting, to account for solidification shrinkage and machining removal.
So a casting record retains the process, the mould or die type, the melt or lot identity, the heat treatment, the machined datum surfaces and the examination method.
Forging deforms a billet, bar or preform with hammers or presses while the material remains solid, commonly at a temperature where it flows plastically.
Large sections are normally hot worked, because deformation resistance falls at elevated temperature.
Its distinguishing evidence is directional flow, and a geometry reachable by dies and press force.
Starting stock, forging method, deformation direction, heat-treatment condition, flash removal, machining allowance and die sequence belong in the record.
“Forged” alone is not traceability.
Neither route owns a universal strength claim.
A forged yoke may place continuous worked material around a load path. A cast yoke may provide a far more intricate internal passage.
The relevant comparison needs an alloy, a quality level, a section, a defect concern and a load case.
A blanket claim that one is stronger postpones the selection problem.
It does not solve it.
Powder is less a small solid than an engineered population, held in suspense before consolidation.
Particle-size distribution, sphericity, satellites, surface oxide, apparent density, flow and moisture together determine how it fills a die, spreads a layer, or enters a spray stream.
Size distribution, flowability, spreadability and powder-layer density are all connected to the quality of an additively manufactured part.
The record needs a distribution and its method.
It does not need a headline size.
Average powder sizes across alloys and processing conditions span an order of magnitude, and two lots sharing a nominal figure need not share a distribution or a flow behaviour.
Powder reuse adds a cycle count, a blending ratio, a sieving event and a changed distribution or chemistry. A storage label may also need an inerting or desiccation state.
These are lot-history constraints.
The finished part inherits every one of them, through the consolidation route.
A material property never arrives alone.
Metals change with heat treatment, cold work, section thickness and weld region.
A modulus measured at room temperature, dry, is not the same data object as the same nominal modulus measured hot, after humid conditioning.
The same discipline applies to a certificate. Mechanical values need a stated product condition, direction, test method, gauge length and specimen location.
The authority must also match the field.
A drawing or product standard controls dimensions and tolerances. A mill test report identifies one lot’s reported results. A calibration record supports the measurement system.
A laboratory report supports a measured microstructural or powder value.
A familiar grade name does not join those evidence types into one guarantee.
This is also where a selection chart can manufacture false margin.
Steel and an aluminium alloy differ in density by nearly a factor of three, which implies quite different sections for the same stiffness.
But that comparison does not decide whether the necessary plate thickness is available, whether a tube wall survives bending, or whether a welded detail retains the required property.
Specific strength and specific modulus are screen coordinates.
They are not manufacturing specifications.
Environmental evidence has the same form dependence.
A temperature swing across an expansion mismatch can produce fit loss or stress even where both materials meet their individual temperature ratings.
A saltwater immersion, a condensing coastal enclosure and an intermittent chloride cleaning are three different environments.
A daily condensation cycle can matter a great deal more than an annual rainfall figure.
The form determines the crevices, cut edges, coatings, seams, machined surfaces and inspection access at which those conditions become material behaviour.
Manufacturing and supply constraints enter during screening, not after the material has been declared.
A highly complex geometry with a tight tolerance can make a nominally cheap casting require extensive machining, changing both the cost and the material feasibility.
Annual volume, lot size, target cycle time, permissible scrap rate, surface condition, source count, qualified-region coverage, minimum order quantity, lead time, purchase currency and price-update date are all decision inputs.
They are not fixed properties of an alloy or a polymer family.
Shortlisting is the point at which generic values are replaced by grade, condition, product specification and process-route evidence.
Two candidate components are comparable only after both satisfy the same load, displacement and service-life requirement.
The procedure calls for a ranked shortlist rather than a forced early winner, and the usual range is two to five candidates.
Sensitivity cases at eighty percent, one hundred percent and a hundred and twenty percent of nominal load, and over a range of service-life assumptions, reveal the candidates which won only because one provisional number was allowed to behave like a fact.
The release review is cross-functional.
Design owns the performance. Manufacturing owns route feasibility. Quality owns verification. Procurement owns source and lead-time evidence. The accountable manager owns the residual risk.
More than seventy percent of a product’s cost is determined during preliminary design, which is precisely when its requirements are least precise.
For high-consequence or regulated work, the final gate adds a prototype build, defined tests, production-representative samples, and controlled approval of suppliers and process changes.
Five prototype pieces are not evidence of a stable process at five thousand units a year.
So the final selection is not a material label followed by an unexamined line on a purchase order.
It is a controlled statement of a component: grade, condition, form, dimensions, tolerances, direction, surface state, route, joining and inspection needs, permitted sources, evidence, and the circumstances which trigger a re-selection.
That statement is a good deal less elegant than a chart winner.
It is also the one a factory can make, a quality system can inspect, and a fielded component can be expected to survive.
AT THE BENCH
A property chart screens material classes under stated conditions; it does not establish that a component can be supplied, formed, joined, inspected, and qualified. The final record needs the grade, condition, product form, product specification, process route, dimensional controls, and evidence source. A chart position can remain attractive while the required geometry, tolerance, surface condition, or available supplier removes the candidate.
A selection brief states component function, interfaces, service scenarios, design life, expected volume, decision owner, and comparison basis. It separates non-negotiable constraints, objectives, and unknowns. A bracket requirement such as 8.0 kN, 1.2 kg, minus 30 to 90 °C, and 200,000 duty cycles is testable; a request for something lightweight and durable is not.
Geometry eliminates a candidate when a fixed envelope, clearance, wall, tolerance, or interface has no credible compatible route. An envelope of 120 mm by 60 mm by 8 mm with a 0.10 mm positional allowance can narrow the field before a detailed model exists. A 2.0 mm wall and a 0.50 mm clearance also change which sections, forming routes, and inspection methods are feasible.
Plate and sheet are both rolled flat products, but their handling, delivery state, and formability differ. The Aluminum Association uses 6.35 mm as an aluminium trade boundary that helps distinguish discrete plate from coiled sheet, and 0.152 mm as a lower sheet limit before foil. Plate supports deep machining allowance and heavy-section stiffness; sheet makes flatness, coil set, edges, surface, and forming direction central controls.
A tube record needs outside diameter, wall thickness, inside diameter, ovality, eccentricity, straightness, and seam condition because it carries a controlled bore. For a circular tube, inside diameter equals outside diameter minus twice wall thickness: a 50.0 mm outside diameter with a 2.0 mm wall has a nominal 46.0 mm inside diameter. That calculation does not replace measured wall minima or a tolerance method.
Yes. Formed welded tube begins as strip that is shaped into a cylinder and joined along its length, so the seam can be a locating feature, inspection target, and forming limitation. A usable record retains weld location, treatment, and non-destructive examination result. Seamless tube still needs its piercing, extrusion, or drawing route recorded when known because route affects dimensional control and surface condition.
Cold rolling changes more than thickness. It raises dislocation density and can increase strength while reducing ductility unless annealing or another heat treatment changes the condition. Two sheets with the same nominal alloy can therefore respond differently in bending or deep drawing. The record must retain condition, rolling direction, thickness, surface state, and the test condition attached to any mechanical value.
Casting and forging begin with different routes and different constraints. Casting creates complex contours, bosses, ribs, and enclosed passages by filling and solidifying a mould or die, but shrinkage, porosity, inclusions, hot tears, and local variation require control. Forging plastically deforms solid stock and creates directional flow, but requires die reach, press force, draft, flash handling, trimming, and machining. Neither form has a universal strength ranking.
Powder quality depends on particle-size distribution, measurement method, morphology, apparent density, tap density, flow, moisture, and surface-sensitive assay, not one nominal diameter. NIST identifies size distribution, flowability, spreadability, and powder-layer density as variables connected to additively manufactured part quality. A reported 5 µm to 50 µm average-size range from an atomizer does not prove that another producer’s 50 µm powder flows or spreads the same way.
A nominal value names the intended size, a specified value states an acceptance limit, a measured value records an observation, and a calculated value is derived from stated inputs. A 25.4 mm nominal bar, a 25.0 mm permitted minimum, a 25.18 mm reading, and an area calculated from that reading are four different records. Confusing them can turn a purchasing label into an unsupported design margin.
Manufacturing feasibility belongs in screening, alongside temperature, environment, and mandatory properties. A complex geometry with a plus-or-minus 0.05 mm tolerance can make a nominally inexpensive casting require machining, altering cost and feasibility. NIST notes that material choice can preclude many manufacturing processes and that geometry, material, and process are tightly coupled. Detailed tooling and conversion costing follow after the candidate field is smaller.
Release replaces generic chart data with grade, condition, product specification, process-route, supplier, and test evidence. Design confirms performance; manufacturing confirms route feasibility; quality confirms verification; procurement confirms source and lead-time evidence; and an accountable manager accepts residual risk. For high-consequence work, prototype builds, defined tests, production-representative samples, and controlled approval of suppliers and process changes form an additional qualification gate.
Occasional updates on new articles, materials, and the mechanisms behind them.