You Already Know How to Think About the Inputs
What Every Engineer Should Know About FEA
A Guide for Engineers and Designers Who Commission Simulation Work
PART 2
Drop Simulation
What Goes Into the Model
Joseph P. McFadden, Sr.
The Holistic Analyst
2026
You Already Know How to Think About the Inputs
Part One established the foundation: your system has a nature, the simulation is a window into that nature, and you are a partner in the investigation — not a passive recipient of results. This part opens the door to the model itself, not to turn you into an analyst, but to give you a genuine understanding of what goes into building one.
Consider what you would need to specify before running a physical drop test in a real laboratory. You would need to know what you are dropping, from what height, onto what surface, and in what orientation. You would need to know how the product is configured — whether optional modules are attached, whether the battery is installed, whether any doors or covers are in their normal closed position.
You already know how to think about the inputs to a simulation. A drop simulation is a virtual drop test. It needs the same information. The difference lies in how precisely that information must be defined, and in what happens when some of it is uncertain or unavailable. This part walks through each category of input — what it is, why it matters, and where your knowledge as the product expert directly shapes the quality of the model.
Geometry: The Shape of Your Product in the Model
The analyst begins with your CAD model. That is the starting point — but the CAD model and the simulation model are never the same thing, and understanding why is important.
A CAD model is built to communicate design intent and to support manufacturing. It may contain thousands of small features: tiny fillets, logo embossments, text details, small holes, and fine chamfers at part edges. Each of those features exists in the real world. In the simulation, some of them influence the structural response and some do not. The analyst makes informed decisions about which features to carry forward and which to simplify, based on engineering judgment about what is actually controlling the load path and the stress distribution.
This process — sometimes called geometry cleanup or idealization — is not a shortcut. It is part of doing the work correctly. A simulation that attempts to model every detail of a complex product would be impractical to build and prohibitively expensive to run, without necessarily providing better answers than a well-considered simplification. The skill is in knowing which details to preserve and which to remove without changing the answer that matters.
Where your knowledge is essential:
Are there features in your design that you know are structurally significant — a rib that carries primary load, a boss that is historically near the limit, a snap fit that must survive the event intact? Identify those features to the analyst before the geometry is prepared. Do not assume they will be recognized from the CAD file alone. Your knowledge of the design's structural intent makes the model better.
The Mesh: Dividing the Structure for Computation
Once the geometry is prepared, the analyst converts the continuous structure into something the simulation can compute with. This is called meshing. The structure is divided into thousands — sometimes hundreds of thousands — of small, discrete pieces called finite elements. Each element is a mathematical unit that the computer solves. Assembled together, they produce a picture of the complete structural response.
Think of a mosaic. From a distance, the image looks continuous and smooth. Up close, you can see the individual tiles. A finer mosaic — smaller tiles, more of them — resolves finer detail. The simulation works the same way. A finer mesh captures stress gradients and geometric transitions more accurately. But more elements mean more computing time. The analyst balances resolution against practicality, using a finer mesh where it matters most and a coarser mesh where the results are relatively insensitive to element size.
In practice, regions of the model where stress is expected to concentrate — around corners, at holes, along thin walls, at interfaces between different thicknesses — will have a finer mesh. Regions far from the structural action may be resolved more coarsely. This is a deliberate engineering choice, and it is one of the factors that distinguishes an experienced simulation engineer from one who is simply running software.
Materials: Characterizing How Your Design Is Actually Made
Materials are where the model becomes most deeply connected to the physics of your specific design. Every material in your product — every plastic housing, metal bracket, adhesive layer, seal, and gasket — must be characterized by properties the simulation can use.
The Basic Properties
At minimum, the simulation needs to know how stiff a material is — how much it resists deformation under load — and how dense it is, which determines how much mass each part contributes to the dynamic response. These two properties govern most of the structural behavior during a drop event.
For materials that may be loaded beyond their elastic limit during the event, the simulation also needs to know their strength — the stress level at which they begin to yield or crack — and how they behave after that threshold is crossed. A housing that deforms plastically during a drop stores and dissipates energy differently than one that remains elastic throughout. Getting that distinction right requires having the right material data in the model.
Rate Dependence: The Behavior That Surprises Most Engineers
Here is something many engineers are surprised to learn. Many materials behave differently under high-speed loading than they do under slow, quasi-static loading. The rate at which strain accumulates in the material — the strain rate — can significantly change its apparent stiffness and strength. Some plastics are tougher under impact conditions than their standard data sheet properties suggest. Others become more brittle. The behavior of the material during a drop event is not necessarily what you would observe in a slow mechanical test.
A well-built drop simulation uses material properties that are appropriate for the loading rates associated with drop events. When those properties are available — from the material supplier, from published literature, or from coupon-level impact testing — the model can represent the material behavior much more accurately. When rate-dependent properties are not available, the analyst uses the best available characterization and notes that as a limitation. That limitation is real, and it belongs in the conversation between analyst and engineer.
Where your knowledge is essential:
What materials are actually in your production product? Are there regions where the material is reinforced — glass-filled nylon, for example — and is the fiber orientation from the molding process known to affect performance? Has this design been through a previous revision that addressed a specific material failure? Have you observed behavior in drop testing that was different from what was expected? That observational knowledge is data. Share it.
The Drop Conditions: Setting Up the Event
Beyond the product itself, the simulation requires a precise definition of the event. This is where the partnership between analyst and engineer becomes most consequential.
Impact Velocity
The drop height specified in your test requirement determines the velocity at which the product strikes the floor. This is a straightforward physics calculation — the analyst computes the impact velocity from the drop height, accounting for the effect of gravity over the fall distance. What matters is that the correct height is used, and that if there are multiple drop heights in your specification, the analyst understands which cases are being evaluated and why.
Impact Surface
The surface the product lands on matters more than it might seem. A rigid floor — the most common assumption in drop simulation — brings the product to a stop more abruptly than a surface with any compliance. That abruptness determines how the impact pulse is distributed over time, which in turn affects the peak forces and stresses in the structure. Most simulations model a rigid surface, which represents the worst-case loading scenario and is consistent with hard-floor test standards. If your product is used in environments where softer surfaces are more realistic, that is worth discussing with the analyst — it changes what the model is actually representing.
Impact Orientation
A product can land on a flat face, on an edge, or on a corner. Each orientation loads the structure differently and may reveal different vulnerabilities. The analyst typically evaluates the orientations most likely to be structurally significant — those that load the design hardest at its most vulnerable locations. Identifying those orientations, however, requires knowledge of the product that no CAD file can fully convey.
Where has this design — or a similar one — failed in drop testing before, and in what orientation? Which faces or corners are most likely to be impacted in real-world use? Are there orientations that your test requirement specifies but that you know are unlikely in the field, or conversely, orientations that are not in the requirement but that you have seen produce failures? These are questions only you can answer. Bring them into the conversation before the analysis plan is finalized.
Contact: How Parts Interact During Impact
Contact is one of the most technically demanding aspects of a drop simulation — and one of the most physically important. During a drop event, multiple contact interactions occur simultaneously: the outer housing contacts the floor; internal components contact the housing as they continue moving while the shell decelerates; the two halves of a housing contact each other at their interface; fastened or bonded surfaces are loaded in ways their connections must manage.
All of these interactions must be defined in the model. The analyst specifies which surfaces can come into contact with which others, how friction behaves at each interface, and how much those surfaces can push against each other before something gives. Getting this right is a significant part of what separates an accurate simulation from a misleading one.
The contact between a plastic housing and a concrete floor at the moment of impact is a genuinely complex physical interaction. The model approximates it. That approximation is typically good enough to produce useful results — but it is worth knowing that it is an approximation, and that the assumed friction and contact stiffness values have an influence on the answer.
Where your knowledge is essential:
How are the internal components actually attached to the housing? Are there features that limit their travel — bosses, ribs, or stops — that need to be in the model? Are there interfaces in the assembly that are known to slip, separate, or behave unpredictably under impact? These details are frequently the difference between a model that predicts reality and one that misses the most important failure mode.
The Complete Picture: A Model Frozen Before Impact
With all of those inputs defined, it helps to form a mental image of the model at the moment before the simulation begins.
Your product is represented as a mesh of thousands of small elements. Every element has been assigned material properties — stiffness, density, strength, and rate dependence where available. Contact relationships are defined between every surface pair that will interact during the event. The floor surface is in place. The product's orientation at impact is set, and its velocity at the moment of contact has been calculated from the drop height.
The simulation begins to advance in time. Contact is established as the housing meets the floor. Forces develop and propagate through the structure in waves, moving through the housing, into the internal components, across the interfaces between them. Stress builds rapidly in some regions and remains low in others. Some parts of the structure deform elastically and recover. Others may cross into plastic deformation and remain permanently changed. The event plays out over a few milliseconds of simulated time, and the record of everything that happened — at every element, at every time step — is what the analyst now reads.
Every input to the model carries uncertainty. That uncertainty does not undermine the simulation's value — it defines the conversation you need to have about what the results mean.
What This Means for You
The quality of a drop simulation's results is directly proportional to the quality of its inputs. Geometry that faithfully represents the load-carrying features of the real design. Material properties that characterize how the actual production materials behave under impact conditions. Drop conditions that reflect the scenarios that matter for your product. Contact definitions that match the real mechanical relationships between parts.
Many of those inputs rely on information that only exists in one place: your knowledge of the product. The manufacturing process, the field failure history, the design intent behind features that may look minor in a CAD file, the use scenarios that are not captured in any formal specification — all of it is input to the model, even when it does not take the form of a number in a data field.
When you sit down with the analyst to review the model setup — ideally before the analysis runs rather than after it concludes — bring that knowledge. Look at the geometry simplifications and flag anything that strikes you as structurally important. Review the material assignments against what is actually in production. Confirm that the drop orientations being evaluated represent the scenarios you are most concerned about. Ask what assumptions were made, which of those assumptions carry the most uncertainty, and how sensitive the results are likely to be to those assumptions.
This is not second-guessing the analyst. This is what a genuine partnership looks like. The analyst has built the best model possible with the information available to them. You are the source of the information they could not get anywhere else.
Coming Up in Part 3
Part Three moves from inputs to outputs. The model has been built. The event has played out in the simulation. Now comes the most consequential part of the process: reading what the structure is telling you. We will walk through stress, displacement, and acceleration results in detail — not just what they measure, but how to look at them, what questions they should prompt, and how to connect the model's predictions to the design decisions in front of you.
— End of Part 2 —
© 2026 Joseph P. McFadden, Sr. | The Holistic Analyst | McFaddenCAE.com
Freely shared for the engineering community. Not for resale.
