Materials Science · Processing, structure and useful properties

Material properties and component performance

A material modulus is not a bracket deflection. Turning it into a prediction requires geometry, loads, connections and a model of deformation. A strength value similarly needs a defined failure criterion and a relevant stress assessment.

Useful background: Atoms, bonding, structure and defects; Units, loads, stress and strain. Each worked case explains the quantities it uses.

Back to the chapter

After reading, you can

Connect property data to a specified component model and assessment.
Predict load-induced motion in a simple link-and-attachment model using material, geometry and connection data.
Use local and overall displacement observations to distinguish explanations for a model discrepancy.
Assess the same component under different stated conditions without transferring a short test to untested service.

From a value to a working part

An alignment link connects a loaded piece of equipment to a rigid frame. Its material has a measured elastic modulus, yet the equipment can move because the link stretches and because its attachment also deforms. To decide whether the equipment stays aligned, we need the movement of the complete load path, not the modulus number alone.

Material response relates quantities such as stress and strain in a specified state, direction and environment. Component performance asks whether the actual part and its connections meet a requirement under the relevant loads and conditions. The geometry and boundary conditions turn one description into the other. Boundary conditions specify how a component is supported, restrained or loaded; an attachment assumed perfectly fixed is one such model choice.

Use a deliberately simple component. A uniform straight link has original length L = 100 mm and cross-sectional area A = 10 mm². One end is connected to a rigid frame through a compliant attachment; the other carries a slowly applied axial tensile force F = 200 N. Compliant means that a finite force produces some movement. The requirement is a load-induced movement of the loaded end no greater than 0.100 mm relative to the frame. All numerical inputs and observations below are original invented teaching data.

Model the link and its attachment separately

For the uniform link, engineering stress is σ = F/A and average axial strain is ε = δ_m/L, where δ_m is its own extension. With a small linear-elastic response σ = Eε, substitution gives δ_m = FL/(EA). E is the effective Young’s modulus for the material direction and conditions being modelled. The earlier stiffness lesson derives this relation; here the new task is to place that link in an assembly.

Represent the attachment by an ideal linear spring with stiffness k_j, defined by F = k_j δ_j. Thus δ_j = F/k_j, where δ_j is the link’s frame-end movement relative to the frame. The spring symbol need not represent a literal coil. A real connection can deform through its members and contacts, and may have sliding or settling that a simple elastic spring does not capture. A measured effective k_j describes the particular attachment and test, rather than an intrinsic material modulus.

The link and attachment lie along one load path, in series. With negligible inertia and no other axial loads, they carry the same force F. The attachment moves the frame-end of the link by δ_j; the link then lengthens by δ_m between its own ends. The loaded end therefore moves δ_total = δ_j + δ_m = F/k_j + FL/(EA). The two extensions add, rather than the two stiffnesses. This is the displacement compatibility of the assembly.

Check the dimensions: E in N/mm² multiplied by A in mm² gives a force, so FL/(EA) is in mm. k_j is in N/mm, making F/k_j also a length. Compliance means movement per force. Here δ_total/F = 1/k_j + L/(EA), in mm/N. The reciprocal is the overall axial stiffness, in N/mm. Adding GPa to N/mm would mix material and component quantities.

Assume aligned tension, a uniform effective modulus, small deformation, a rigid frame and load introduction that adds no other movement or bending. Both elements are assumed linear and recoverable through 200 N in the stated short tests. The measured link length excludes the compliant attachment. A bracket loaded transversely would need a bending model; the original bracket question below should not be solved by silently relabelling this axial equation.

Original schematic of the axial teaching model, not to scale. A rigid frame on the left connects through one ideal linear attachment spring of stiffness k_j to a uniform link with modulus E and original area A. The link length between its own ends is L = 100 millimetres. A 200 newton tensile force acts to the right at the loaded end. There is one series load path and, under static equilibrium with no other axial loads, the same force passes through both elements. Attachment movement relative to the frame and extension between the link ends add to the loaded-end movement. The drawn spring represents an effective attachment response, not a required physical coil, and the schematic does not draw actual displacement magnitudes.

Carry the test conditions into the component prediction

The table supplies matched material and attachment inputs for a cool test at 20 °C and a warm test at 100 °C. At each temperature the assembly is first equilibrated without the tensile load, then the 200 N load is slowly applied and the reading is taken after 10 s. E and k_j are supplied for that same short-test regime and configuration. Their numerical changes are stipulated observations, not predictions for a named alloy or connection.

Use L = 100 mm and A = 10 mm² in both calculations, neglecting their small thermal dimensional changes for this simplified comparison. Importantly, each displacement is measured from the unloaded reference at that same temperature. This assesses the extra movement caused by the applied load. It does not calculate the absolute change in equipment position on heating from 20 to 100 °C. That question would also need thermal expansion, restraint and the actual temperature-dependent dimensions.

At 20 °C, the link extends 0.020 mm and the attachment contributes another 0.020 mm. Their total is 0.040 mm, within 0.100 mm. At 100 °C, the link extends 0.025 mm and the attachment contributes 0.100 mm. The total becomes 0.125 mm and misses the requirement. The link-only calculation would have predicted a pass at both temperatures because it omitted the larger contribution.

In each test the link’s average stress is 200/10 = 20 MPa. The chosen strains, 0.020/100 = 0.00020 and 0.025/100 = 0.00025, are small and consistent with the stipulated linear range. This checks the deformation model’s internal consistency. It does not supply a failure strength, allowable load or safety factor for a real assembly.

Original invented short-test inputs and predictions at F = 200 N. Load-induced movements use an unloaded reference at the same temperature; nominal L and A are held fixed in this simplified model.
Quantity	20 °C test	100 °C test
E	100 GPa	80 GPa
Attachment stiffness k_j	10 000 N/mm	2000 N/mm
Link extension δ_m	0.020 mm	0.025 mm
Attachment movement δ_j	0.020 mm	0.100 mm
Total loaded-end movement	0.040 mm	0.125 mm
Meets 0.100 mm limit?	Yes	No

Original calculated comparison at 200 newtons, with nominal link length 100 millimetres and area 10 square millimetres at both temperatures. A shared horizontal scale shows load-induced movement from 0 to 0.150 millimetre, referenced to the unloaded state at each temperature. At 20 degrees Celsius, the solid blue link contribution is 0.020 millimetre and the hatched attachment contribution is 0.020 millimetre, totalling 0.040 millimetre. At 100 degrees Celsius, the link contribution is 0.025 millimetre and the attachment contribution is 0.100 millimetre, totalling 0.125 millimetre. A long-dashed vertical line marks the maximum permitted total movement of 0.100 millimetre. The cool total lies below the limit, whereas the warm total exceeds it. The hatched segments distinguish attachment motion beyond colour. This graph does not show absolute position changes caused by heating.

Close the loop with evidence

Check that the data represent the finished material state and environment. Compare component predictions with suitable tests, including the relevant connections and manufacturing variability. Differences can reveal a faulty boundary condition, an overlooked joint or inappropriate material data rather than a universal failure of the material.

Now test a separate actual assembly in the invented cool-condition investigation below. It is nominally the same design, but at 200 N its loaded end moves 0.060 mm rather than the model’s 0.040 mm. Suppose independent material testing remains repeatable at E = 100 GPa. Repeatability is useful evidence about that test; it does not establish that every attachment and dimension in the component is as modelled.

If one assigns all 0.060 mm to stretching the 100 mm link, the apparent modulus would be FL/(Aδ_total) ≈ 33.3 GPa. That fit treats motion of a different measurement interval as material strain. It is not a new material measurement. Even before the independent material result, the total displacement alone leaves competing explanations: E = 100 GPa with k_j = 5000 N/mm, or E = 50 GPa with k_j = 10 000 N/mm, both give 0.060 mm at 200 N.

Add a measurement between the link’s own ends. The table’s 0.020 mm local extension at 200 N gives ε = 0.00020 and E = 20/0.00020 = 100 000 MPa = 100 GPa. It distinguishes the two proposed explanations: the lower-modulus explanation would predict 0.040 mm local extension. Subtracting 0.020 mm from the 0.060 mm overall movement leaves 0.040 mm outside that link gauge.

If the rigid-frame and single-attachment model is valid, this remaining movement corresponds to an effective k_j = 200/0.040 = 5000 N/mm. It does not identify a microscopic cause within the connection. Check attachment motion, frame rigidity, alignment, dimensions and measurement references rather than declaring that a particular bolt slipped. Real residuals also need measurement uncertainty and variation assessed; the exact invented readings merely make the competing explanations visible.

Separate original invented cool-condition component observations. Each row is referenced to the unloaded condition. Local link extension and total loaded-end motion use different, explicitly defined endpoint pairs.
Force (N)	Link extension between its own ends (mm)	Loaded-end motion relative to frame (mm)
0	0	0
100	0.010	0.030
200	0.020	0.060

Improve the part that controls the observed movement

A higher modulus reduces only the link term. In the nominal cool model, doubling E from 100 to 200 GPa halves the link extension to 0.010 mm, but leaves the 0.020 mm attachment movement. The total falls from 0.040 to 0.030 mm, a 25% reduction rather than 50%. A large change in one material property can give a smaller change in complete-component performance.

In the warm model, doubling area to 20 mm² gives δ_m = 0.0125 mm, but the attachment still moves 0.100 mm at 200 N. The total 0.1125 mm still exceeds the limit. Within this model, increasing area alone cannot achieve a total no greater than 0.100 mm: the attachment already consumes that entire movement allowance, and the finite link still stretches. This does not mean that a real redesign leaves connection stiffness unchanged; that independence is an explicit condition of the calculation.

Choose an investigation or change that matches the question. A local strain measurement tests material response; an attachment displacement measurement tests the assumed restraint; component loading tests the combined response. A thicker link, a stiffer connection, a different load or a changed operating condition affect different terms. The later selection lessons optimise designs under multiple constraints; here the aim is to know which model input and observation justify a proposed change.

A passed short test has a defined reach

The 10 s calculation says nothing by itself about hours under load, repeated cycles, impact, chemical exposure or damage. Time-dependent deformation under sustained stress is called creep; changes in contact or material state can also make an initially adequate component drift. Those effects require appropriate evidence or an extended model. A service-condition label should include the loading history and environment, not just a single nominal force.

For an assembled component, retain the actual load position and direction, dimensions, manufacturing state, connections and measurement reference. A repeatable coupon modulus can be the right material input while an assumed perfectly fixed support is wrong. Conversely, a well-characterised attachment cannot make an unrelated room-temperature coupon represent a changed hot material state.

Use the chain processing → structure → property → component model → checked performance. Each arrow needs a justified link. The present model deliberately isolates one axial movement requirement; bending, buckling, stress concentrations, failure endpoints, thermal position changes and durability remain separate questions when the real use introduces them. A disagreement is useful because it tells us which assumptions to investigate, rather than which property to adjust until the numbers agree.

Worked examples

The load point moves more than the link stretches

At 200 N, calculate the nominal cool and warm loaded-end movements, including the attachment. Compare both with the 0.100 mm requirement.

Use 1 GPa = 1000 N/mm², so E is 100 000 N/mm² at 20 °C and 80 000 N/mm² at 100 °C. Keep F = 200 N, L = 100 mm and A = 10 mm².
Cool link: δ_m = (200 × 100)/(100 000 × 10) = 0.020 mm. Cool attachment: δ_j = 200/10 000 = 0.020 mm. Add them: δ_total = 0.040 mm ≤ 0.100 mm.
Warm link: δ_m = (200 × 100)/(80 000 × 10) = 0.025 mm. Warm attachment: δ_j = 200/2000 = 0.100 mm. Their total is 0.125 mm > 0.100 mm.
The warm total exceeds the limit by 0.025 mm, or 25% of the permitted 0.100 mm. The attachment is 0.100/0.125 = 80% of the warm total. These percentages have different denominators and answer different questions.

The cool model passes at 0.040 mm; the warm model fails at 0.125 mm. A link-only prediction misses both the attachment contribution and its change between the supplied conditions.

Work through the case

A bracket deflects more than a model predicts, although its modulus measurement is repeatable. What should be investigated?

Check load position and restraints.
Inspect joint movement and actual section dimensions.
Then examine material-state assumptions and measurement scope.
Choose observations that distinguish those explanations. For the axial link example, separate local link extension from whole-assembly movement; for the bracket, use the appropriate bending and connection model rather than this axial equation.

Investigate the whole model–component match rather than adjusting modulus solely to force agreement.

One total motion can hide different causes

At 200 N the separate cool assembly moves 0.060 mm overall. Verify the two proposed material/attachment explanations, then use the measured 0.020 mm local link extension to distinguish them.

For E = 100 GPa and k_j = 5000 N/mm, the link extends 0.020 mm and the attachment moves 200/5000 = 0.040 mm. The total is 0.060 mm.
For E = 50 GPa and k_j = 10 000 N/mm, the link extends 20 000/(50 000 × 10) = 0.040 mm and the attachment moves 0.020 mm. This also totals 0.060 mm. Overall motion alone does not separate the two inputs.
The local reading is 0.020 mm across L = 100 mm, so ε = 0.00020. With σ = 200/10 = 20 MPa, E = σ/ε = 100 000 MPa = 100 GPa. It supports the first explanation under the stated axial and uniformity assumptions.
The remaining 0.060 − 0.020 = 0.040 mm is outside the link gauge. Assigning it to the one effective attachment gives 5000 N/mm only if the rest of the assumed load path and reference are correct. Check those assumptions before naming a physical defect.

Both proposed input pairs fit the same overall movement. The added local measurement selects the 100 GPa link response and directs the investigation toward attachment or other out-of-gauge motion.

Transfer to an attachment change

Keep the warm link at E = 80 GPa, L = 100 mm and A = 10 mm² under 200 N. A redesigned attachment is separately demonstrated to have k_j = 5000 N/mm under the same conditions. What total movement does the model predict?

The unchanged link still extends 0.025 mm. The new attachment term is 200/5000 = 0.040 mm.
The predicted total is 0.025 + 0.040 = 0.065 mm, within the 0.100 mm movement limit. The nominal difference below the limit is 0.035 mm.
This change targets the larger original warm contribution. The comparison assumes that the new connection does not alter load alignment, effective link length or material state. Demonstrate the complete changed assembly, rather than treating its separately measured k_j as a full service approval.

The model predicts 0.065 mm, satisfying this exact-data short-test movement requirement. Other component requirements and the changed assembly still need their own evidence.

Try the reasoning

Can a specimen test establish the stiffness of every assembly?
Show the answer and reasoning
No.
Assemblies add geometry and connections. In the worked model, E helps predict δ_m, while k_j contributes separately to δ_total. A specimen modulus alone therefore does not determine loaded-end motion.
Why test manufactured parts as well as ideal specimens?
Show the answer and reasoning
To assess effects introduced by the actual process and assembly.
Those effects may be absent from idealised property tests. Component observations retain connections, actual load paths and dimension variation. Separate local and overall readings can distinguish an assembly effect from a changed material response.
In the nominal cool model, double E to 200 GPa while keeping force, geometry and attachment stiffness unchanged. Does total movement halve?
Show the answer and reasoning
No. It becomes 0.030 mm, which is 25% below the original 0.040 mm.
The link term falls from 0.020 to 0.010 mm. The attachment still moves 0.020 mm, so add them. Only one contribution halved; the complete load path did not become twice as stiff.
In the nominal warm model, reduce the tensile force to 100 N without changing the geometry, material state or connection. Predict link stress and total movement, staying within the stipulated linear range.
Show the answer and reasoning
Stress becomes 10 MPa and total movement becomes 0.0625 mm, within the 0.100 mm limit.
Both terms scale with F: the link contributes 0.0125 mm and the attachment 0.0500 mm. The load is part of the performance question; satisfying the movement limit at 100 N does not establish it at the required 200 N.
Double the warm link area to 20 mm² but hold the original 2000 N/mm attachment stiffness and 200 N force fixed. Calculate total movement. Could increasing only link area ever meet the 0.100 mm limit in this model?
Show the answer and reasoning
The total is 0.1125 mm and fails. No finite positive area can achieve that limit with the unchanged attachment term.
The link term is 0.0125 mm, while 200/2000 = 0.100 mm is already used by the attachment. Any finite link adds positive extension. This conclusion is conditional on the connection being unaffected by the area change.
A calculation uses the separate assembly’s total 0.060 mm movement as the extension of its 100 mm link at 200 N. What apparent E results, and why is it not a measured material modulus?
Show the answer and reasoning
It gives about 33.3 GPa, but includes motion outside the link gauge.
FL/(Aδ_total) = 20 000/(10 × 0.060) ≈ 33 333 MPa. The matched local reading gives 100 GPa instead. A reproducible total movement is still a measurement of the assembly interval, not automatically the link’s own strain.
A new aligned link has L = 80 mm and A = 8 mm². At 160 N it extends locally by 0.016 mm, while the loaded end moves 0.056 mm relative to the rigid frame. Under the same two-element model, find E and the effective attachment stiffness.
Show the answer and reasoning
E = 100 GPa and k_j = 4000 N/mm.
Stress is 160/8 = 20 MPa; local strain is 0.016/80 = 0.00020, giving E = 100 000 MPa. The outside-gauge movement is 0.056 − 0.016 = 0.040 mm, so k_j = 160/0.040 = 4000 N/mm. This inference still needs the assumed rigid frame, alignment and single effective connection.
The warm test satisfies the movement requirement after a connection change. Does its displacement, zeroed at 100 °C before loading, establish that equipment position changed by less than 0.100 mm when heated from 20 °C?
Show the answer and reasoning
No. It establishes only load-induced movement from the warm unloaded reference in this test.
The reference changed with temperature. Absolute position on heating also depends on thermal expansion, restraints and temperature distribution. A correctly calculated answer to one displacement question does not answer the other.
The modified warm assembly moves 0.065 mm after 10 s. Can that one result establish alignment after a day under load or after many load cycles? Name an appropriate next question.
Show the answer and reasoning
No. Ask how total and local movements evolve under the relevant sustained or repeated loading and environment.
The short linear model does not represent creep, settling, damage or cyclic changes. A service-representative observation or validated extension of the model is needed; the initial pass remains evidence only within its tested conditions.

What to carry forward

Properties become performance through a model and a design.
Use component evidence to check the connection.
In the stated series model, δ_total = FL/(EA) + F/k_j; material and attachment contributions have different inputs.
Measure the interval named in the model. Overall motion alone can leave several material/connection explanations possible.
Carry reference state, geometry, load, direction, temperature and timescale into a performance claim. A short movement check does not supply an entire service qualification.

References

Teaching resources and research supporting this lesson’s explanations and examples.