From Explicit Animation to Defensible Engineering Evidence

The Purpose of an Explicit Simulation

The purpose of an explicit simulation is not to produce an animation. It is to reproduce the governing physics with sufficient fidelity to answer an engineering question. The animation is a visualisation of the calculation — a convenient way to see what the model computed — but it is not the evidence. The evidence is the verified physics behind the animation: the correct mass and velocity, the correct contact and load transfer, the correct material response, the correct failure behaviour, the verified energy balance, the demonstrated mesh independence, and the correlation with physical test. A convincing animation can be produced by a model with any of these wrong — a visually impressive result is not evidence that the model is correct. The discipline of converting an explicit simulation into defensible engineering evidence is the discipline of checking every link in the chain from the physical event to the engineering conclusion.

AN EXPLICIT SOLVER CAN CALCULATE A VISUALLY IMPRESSIVE ANSWER TO A BADLY POSED PROBLEM JUST AS EFFICIENTLY AS IT CAN CALCULATE A USEFUL ONE.

The Complete Chain — From Physical Event to Engineering Conclusion

Every defensible explicit impact analysis follows the same chain: the physical event is defined, the mass and velocity are established, the momentum and energy are checked, the structural configuration is defined, the contact is defined, the material models are selected, the failure models are defined, the mesh and formulation are selected, the stable timestep is checked, the mass scaling is justified, the hourglass and artificial energy are checked, the energy balance is checked, the mesh and parameter sensitivity is performed, the results are compared with test or physical evidence, the limitations are documented, and the engineering conclusion is made. Each step depends on the ones before it, and skipping or accepting a step without verification breaks the chain. The ordered list below is the chain that every analysis should follow.

1. 01 — DEFINE THE PHYSICAL EVENT — What is being analysed? What is the impact scenario, the bodies involved, the boundary conditions, the duration, the expected response? The analysis must answer a defined question about a defined event.
2. 02 — ESTABLISH MASS & VELOCITY — The mass and velocity of every body in the event must be correct. These define the kinetic energy, the momentum and the load. A wrong mass or velocity makes everything downstream wrong.
3. 03 — CHECK MOMENTUM & ENERGY — The initial momentum and kinetic energy must be consistent with the physical event. These are the quantities that the simulation must conserve and transfer correctly.
4. 04 — DEFINE STRUCTURAL CONFIGURATION — The geometry, the thicknesses, the joints, the connections and the boundaries must represent the physical structure. Simplifications must be justified.
5. 05 — DEFINE CONTACT — The contact definitions — what touches what, with what stiffness, friction and damping — must represent the physical interfaces. Contact energy must be small.
6. 06 — SELECT MATERIAL MODELS — The material models must represent the physical material behaviour at the strain rates, stress states and temperatures of the event. Calibration sources must be documented.
7. 07 — DEFINE FAILURE / DAMAGE MODELS — The failure and damage models must represent the physical failure mechanisms. The criteria, the parameters and the regularisation must be appropriate and calibrated.
8. 08 — SELECT MESH & FORMULATION — The mesh density, the element formulation and the integration scheme must resolve the deformation, the stress waves and the failure. The choice must be justified, not defaulted.
9. 09 — CHECK STABLE TIMESTEP — The stable timestep must be understood and reported. The timestep must be small enough to resolve the highest-frequency response of interest.
10. 10 — JUSTIFY MASS SCALING — Any mass scaling must be quantified and justified. The added mass must not change the inertial response in the frequency range of interest.
11. 11 — CHECK HOURGLASS / ARTIFICIAL ENERGY — The hourglass and artificial energies must be checked as a fraction of total energy. They must be negligible or the formulation must be revised.
12. 12 — CHECK ENERGY BALANCE — The total energy must be approximately conserved. The transfer from kinetic to internal must be explainable. Artificial energies must not dominate.
13. 13 — PERFORM MESH / PARAMETER SENSITIVITY — The engineering quantities must be shown to be stable under mesh refinement and parameter variation. The failure path must not change with mesh.
14. 14 — COMPARE WITH TEST / PHYSICAL EVIDENCE — Where test data exists, the event history — force, velocity, strain, timing, damage — must be compared, not just the final deformation.
15. 15 — DOCUMENT LIMITATIONS — What the model assumes, what it does not capture, what remains uncertain and what the correlation does and does not prove must be documented.
16. 16 — MAKE THE ENGINEERING CONCLUSION — The conclusion must be supported by the verified physics, qualified by the documented limitations, and stated with the confidence that the evidence supports — no more, no less.

The Animation Is the Visualisation. The Evidence Is the Verified Physics.

The chain above converts a calculation into evidence. The animation shows the result of the calculation; the evidence is the body of verification — energy balance, mesh sensitivity, material calibration, test correlation — that proves the calculation is physically credible. An analyst who presents an animation without the verification evidence is presenting a visualisation, not an engineering argument. An analyst who presents the verification evidence alongside the animation is presenting a defensible engineering conclusion. The distinction is the difference between a picture and a proof.

THE ANIMATION IS THE VISUALISATION. THE ENGINEERING EVIDENCE IS THE VERIFIED PHYSICS BEHIND IT.

Explicit Verification Checklist

The checklist below is the comprehensive verification that every explicit impact analysis should pass before its results are used for an engineering decision. The checklist is organised by category — physics, numerics, response, validation and conclusion — and each item is a question that must be answered, not a box to be ticked.

• PHYSICS — Mass correct? — Every body in the model has the correct mass
• PHYSICS — Velocity correct? — Initial velocities match the physical event
• PHYSICS — Energy correct? — Initial kinetic energy matches mass × velocity
• PHYSICS — Momentum correct? — Initial momentum matches the physical event
• PHYSICS — Geometry correct? — Structure, thicknesses, connections represented
• PHYSICS — Material condition correct? — Material models, calibration sources, rate dependence documented
• NUMERICS — Stable timestep understood? — Timestep reported and justified; highest frequency of interest resolved
• NUMERICS — Mass scaling quantified? — Added mass reported; inertial effect in frequency range of interest assessed
• NUMERICS — Hourglass energy checked? — Hourglass / artificial energy as fraction of total; negligible or revised
• NUMERICS — Contact behaviour checked? — Contact energy small; no artificial penetration; no excessive contact stiffness
• NUMERICS — Mesh sensitivity performed? — Engineering quantities stable under mesh refinement; failure path checked
• NUMERICS — Failure model sensitivity performed? — Failure parameters varied; effect on conclusion assessed
• RESPONSE — Force history credible? — Peak, shape, duration physically plausible and consistent with test where available
• RESPONSE — Acceleration history credible? — Deceleration consistent with mass, stiffness and contact
• RESPONSE — Deformation mode credible? — Mode matches physical expectation and high-speed video where available
• RESPONSE — Failure location credible? — Failure occurs where the physics predicts, not where the mesh happens to be weak
• RESPONSE — Failure timing credible? — Failure occurs at the physically expected time in the event
• RESPONSE — Energy absorption credible? — Internal energy consistent with observed deformation and failure
• VALIDATION — Physical test available? — If yes, correlate; if no, document the absence of direct validation
• VALIDATION — High-speed video available? — Frame-by-frame comparison of deformation mode and timing
• VALIDATION — Strain data available? — Strain-time histories at matched locations compared
• VALIDATION — Force data available? — Force-time history overlaid and compared
• VALIDATION — Damage extent available? — Internal damage (ultrasonic / X-ray) compared to FE damage extent
• VALIDATION — Residual velocity available? — Post-impact velocity compared where perforation / rebound occurs
• CONCLUSION — What aspect is actually validated? — State explicitly which quantities have test support
• CONCLUSION — What remains model-dependent? — State explicitly which quantities rely on the model without direct test support
• CONCLUSION — What uncertainty remains? — State the range of outcomes from sensitivity studies and the residual uncertainty

Evidence Types and Their Role

Different types of evidence validate different aspects of the model. The table below maps the principal evidence types to what they validate, what they do not validate, and their role in the engineering argument. No single evidence type is sufficient; the argument is built from the combination.

The Explicit Simulation Surrounded by Evidence

The diagram below shows the explicit simulation at the centre, surrounded by the evidence that supports it — energy balance, mesh sensitivity, material data, test correlation, high-speed video, strain data, force history and residual velocity. The simulation alone is a calculation; surrounded by evidence, it becomes an engineering argument. The absence of any evidence type weakens the argument; the absence of all evidence types reduces it to an unverified prediction.

EXPLICIT SIMULATION SURROUNDED BY EVIDENCE

                    ┌───────────────┐
                    │  ENERGY       │
                    │  BALANCE      │
                    └───────┬───────┘
            ┌───────────────┼───────────────┐
     ┌──────┴──────┐  ┌─────┴─────┐  ┌──────┴──────┐
     │   MESH      │  │  EXPLICIT │  │  MATERIAL   │
     │ SENSITIVITY │  │ SIMULATION│  │   DATA      │
     └──────┬──────┘  └─────┬─────┘  └──────┬──────┘
            │               │               │
     ┌──────┴──────┐  ┌─────┴─────┐  ┌──────┴──────┐
     │   TEST      │  │  FORCE    │  │  STRAIN     │
     │CORRELATION  │  │ HISTORY   │  │   DATA      │
     └──────┬──────┘  └─────┬─────┘  └──────┬──────┘
            │               │               │
     ┌──────┴──────┐        │        ┌──────┴──────┐
     │ HIGH-SPEED  │        │        │  RESIDUAL   │
     │   VIDEO     │        │        │  VELOCITY   │
     └─────────────┘        │        └─────────────┘

  The simulation alone = a calculation.
  Surrounded by evidence = an engineering argument.

Documentation of Limitations

A defensible engineering conclusion is accompanied by a documented statement of limitations. What does the model assume? What does it not capture? What uncertainty remains from the sensitivity studies? What aspect of the conclusion is validated by test and what remains model-dependent? What conditions are outside the validated range? Stating these limitations is not a weakness — it is the honesty that makes the conclusion defensible. A conclusion presented without limitations is a claim; a conclusion presented with its limitations is evidence. The analyst who states what the model does not know is more credible than the one who implies the model knows everything.

Making the Engineering Conclusion

The engineering conclusion is the final step: the answer to the question that the analysis was commissioned to answer, supported by the verified physics and qualified by the documented limitations. The conclusion must be stated with the confidence that the evidence supports — no more, no less. If the model is fully validated by test, the conclusion can be stated with confidence for the validated conditions. If the model is verified numerically but not validated by test, the conclusion is a prediction with stated uncertainty. If the model has unexplained energy, mesh-dependent failure or no correlation, the conclusion is not supported — and the honest answer is that the analysis does not yet provide defensible evidence. The discipline of stopping short of an unsupported conclusion is what separates engineering evidence from a calculation.

Key Takeaways

• The purpose of an explicit simulation is to reproduce physics, not to produce an animation.
• Every analysis follows the same chain — from physical event to engineering conclusion — and every link must be verified.
• The animation is the visualisation; the evidence is the verified physics behind it.
• No single evidence type is sufficient — the argument is built from energy balance, mesh sensitivity, material data and test correlation together.
• A defensible conclusion is stated with the confidence the evidence supports, qualified by documented limitations.