Defensible Turbomachinery Structural Analysis Workflow
End-to-end turbomachinery structural workflow from requirements and operating envelope through loads, model hierarchy, static and thermal response, dynamics, life, extreme events, clearance, verification, correlation and final engineering evidence.
1. Define the engineering decision
Start with the decision the analysis must support: blade strength, disc life, overspeed qualification, tip clearance, casing distortion, shaft fatigue, mount load, HCF resonance, blade-off survival or containment. Identify the applicable programme, customer requirement, certification basis and service-life objective. The required evidence and allowable method follow from that decision. Building a detailed finite-element model before defining the governing limit state usually produces more output but weaker engineering evidence.
2. Freeze the configuration and material basis
Confirm the controlled geometry revision, blade and disc mass properties, casing split-line arrangement, shaft interfaces, bearing locations, mounts and attached hardware. Record material grade, product form, heat treatment, coating and surface condition. For hot components, prepare temperature-dependent modulus, thermal expansion, strength, fatigue and creep data as required. For rotating hardware, density and radial position deserve the same configuration control as geometry because they directly set centrifugal load.
3. Build the operating-envelope matrix
Define normal speed range, pressure ratio, torque, thrust, steady and transient temperature fields, start-stop schedules, dwell time, overspeed, vibration conditions and credible abnormal events. Preserve physical simultaneity: maximum speed, pressure and thermal gradient may not occur at the same instant. Map each event to its number of cycles or duration and to the failure modes it is intended to assess. This matrix becomes the backbone of static, life, dynamic and fault analyses.
4. Establish the load chain
Trace each load to its source. Rotation creates centrifugal body force and blade pull; CFD creates aerodynamic pressure and harmonic forcing; thermal analysis creates metal-temperature fields; the fluid system creates pressure and thrust; rotordynamics provides bearing and orbit loads; connected equipment creates mount and interface reactions. Check units, coordinate systems and sign conventions at every transfer. The analysis should make it possible to follow each load from source through the component to the final support reaction.
5. Perform independent mechanics checks
Before detailed FEA, calculate the quantities that can be predicted directly: blade root centrifugal force, disc stress scale, pressure resultant, shaft torsion, thrust, free thermal growth, mass and centre of gravity. These calculations establish the correct order of magnitude and create independent targets for model verification. They also reveal wrong speed units, density, rotation axis or load scaling before local stress fields make the model look deceptively sophisticated.
6. Choose the modelling hierarchy
Use the least complex model that captures the governing mechanism. Beam or rotor models are efficient for shaft-line and bearing-system behaviour; axisymmetric models are powerful for early disc and casing sections; cyclic sectors suit tuned periodic rotors; shells suit global casing distortion; solids resolve thick hubs, roots and local contacts. Submodels refine the critical geometry without making the full machine expensive. Full-annulus or explicit models are reserved for genuinely asymmetric events such as mistuning, blade-off or containment.
7. Establish static and thermo-mechanical operating states
Solve the steady or transient structural state with centrifugal load, pressure, aerodynamic load, torque, thrust, preload, contact and temperature as required. Apply loads in the physical sequence when contact or plasticity makes path dependence important. Review stress, strain, contact pressure, radial and axial growth, casing distortion and interface reactions. Keep the contributions of rotation, pressure and temperature understandable so that the governing mechanism remains visible.
8. Evaluate modal behaviour and excitation
Use the appropriate pre-stressed operating state for modal analysis. Generate speed-dependent blade or rotor frequencies and track mode identity through the speed range. Add engine-order, vane-passing, blade-passing and other physically supported excitation lines to the Campbell diagram. Candidate crossings then progress to forced-response analysis using the correct spatial harmonic, phase and damping. Where blade-to-blade scatter matters, assess mistuning rather than relying solely on a perfectly tuned cyclic model.
9. Assess static strength, fatigue, creep and fracture
Use acceptance methods matched to the failure mechanism. Static yielding and overspeed may require elastic or elastic-plastic assessment; blade and disc LCF require cyclic strain ranges from start-stop states; HCF requires alternating stress from forced response combined with the correct operating mean stress; high-temperature dwell can require creep or creep-fatigue assessment. Known flaws or damage may require fracture or remaining-life methods. Do not reduce all structural integrity questions to one maximum von Mises stress.
10. Assess clearance and functional performance
Build a controlled clearance budget combining disc and blade centrifugal growth, rotor and casing thermal expansion, casing ovalisation, shaft/bearing motion, vibration allowance, tolerances and long-term creep where relevant. Use one datum and one sign convention. Repeat through transient startup and shutdown, because the minimum gap may occur before thermal equilibrium. Treat alignment, seal position and interference retention as structural acceptance criteria where they govern before material strength.
11. Assess extreme rotor events separately
Overspeed, rotor burst, blade-off and rub are different physical limit states and should not be collapsed into one generic fault case. Overspeed may require plasticity and permanent-set assessment; blade-off requires asymmetric transient rotor-bearing response; containment requires calibrated high-rate material and contact modelling; rub requires nonlinear contact, friction and possibly thermal feedback. Use the analysis method and acceptance criterion appropriate to each event and preserve its actual sequence and duration.
12. Close the loop with verification
Verify mass, centre of gravity, inertia, centrifugal resultant, reactions, pressure force, torque, CFD mapping, thermal mapping, contact load sharing, mesh convergence, modal trend and nonlinear energy behaviour. Cross-check interfaces between the blade, disc, shaft, casing and rotordynamic models. Repeat selected cases at changed speed or mesh density to exercise the setup rather than only inspecting one final result. Record verification as part of the engineering evidence, not as informal analyst notes.
13. Correlate with test and service evidence
Use spin-rig, modal, strain-gauge, blade-tip-timing, clearance, dimensional and engine data to validate the quantities the model predicts. Correlate frequency and mode shape before forced-response stress, and thermal state before thermo-mechanical life. If the model is updated to improve correlation, document what changed and why. Service inspection, blade-frequency surveys and dimensional growth measurements can also validate long-term creep, wear or fatigue assumptions.
14. Quantify uncertainty and robustness
Identify the inputs that materially control the result—damping, material properties, friction, heat-transfer coefficient, support stiffness, blade mass, frequency scatter or manufacturing tolerance—and vary them within credible bounds. A design that passes only at one precise nominal assumption is less robust than one with stable margin across production and operating variability. Convert important sensitivities into drawing controls, operating limits, inspection requirements or test acceptance criteria where possible.
15. Integrate the engineering conclusion
Bring static strength, life, dynamics, clearance, extreme-event response and interface loads together for the same hardware configuration. Identify the governing component, location, load case and failure mechanism. Keep safety factors, material allowables, statistical knock-downs and model conservatism traceable so the same uncertainty is not counted twice. Where one discipline owns a downstream decision—such as foundation design or aerodynamic clearance—provide the simultaneous interface data needed rather than isolated scalar maxima.
16. Report a traceable qualification case
The final report should connect requirement → configuration → load → model → verification → result → acceptance criterion → margin → conclusion. Include assumptions, limitations, model revisions, software version, material sources, mesh evidence, damping basis, contact definition, test correlation and unresolved actions. State clearly what changes would invalidate the assessment: speed, blade mass, temperature, material, repair geometry, bearing stiffness or mount condition. The goal is a reusable technical baseline rather than a one-off set of plots.
A defensible turbomachinery structural analysis is an evidence chain. Model complexity is valuable only when it improves the accuracy, verification and traceability of that chain.