How impact speeds are calculated from conservation of momentum
Updated in 2026 · 5 minute read
From the marks left after the collision you can establish how fast each vehicle was travelling, without measuring a single crush depth. The calculation is a system of two equations and presents no difficulty: everything rests on what is measured on the scene diagram — the approach and departure headings of each vehicle, and the path they travelled before coming to rest.
The two equations
The momentum of the pair is the same immediately before and immediately after the impact. It is a vector quantity, so there are two equations, one per axis:
mA·vA·sin θA + mB·vB·sin θB = mA·v′A·sin φA + mB·v′B·sin φB
Where θ are the approach headings —before the collision—, φ the departure headings —after— and v′ the departure speeds. All angles measured against one and the same reference axis: the north of the diagram or the axis of the road, but the same one for all four. Two equations and two unknowns, vA and vB.
The departure speeds are not measured: they are calculated from the post-impact runs to rest, with the same formula used for speed from skid marks. That is the whole chain: marks → departure speeds → impact speeds.
A worked case
A right-angle intersection between two passenger cars, with the data exactly as it comes off the diagram and the post-impact runs:
| Datum | Vehicle A | Vehicle B |
|---|---|---|
| Mass | 1,470 kg | 1,325 kg |
| Approach heading | 0° (east) | 90° (north) |
| Departure speed, from the marks | 31.5 km/h | 35.0 km/h |
| Departure heading | 23.8° | 47.8° |
| Impact speed obtained | 50.0 km/h | 40.0 km/h |
The collision consumes 47 % of the pair's kinetic energy, which is what a severe impact looks like and serves as a first check: if the departure split assumed gave more energy after than before, the data are not compatible and it is back to the diagram.
What has to be measured on the diagram
- The approach heading of each vehicle, before the collision.
- The departure heading of each one, from separation towards its final position.
- The post-impact run of at least one of them, to rest, with the friction that applies — split into segments if it crosses different surfaces.
- The masses, with occupants and load.
- All angles against the same reference axis.
No crush measurement is needed, and that is the advantage: it is the alternative route when the damage is not documented, when the vehicle is no longer available, or when an independent second route is wanted to check what the damage-based calculation gave. If the two routes agree, the reconstruction is solid. If they do not, there is something to review — and knowing that is already a result.
The forward check
It is the reverse route, and the one simulation programs follow: start from assumed speeds, solve the impulse of the collision and let the two vehicles run until they stop, to see whether they end up where they actually did.
It serves above all to check the contact geometry. If, when the simulation is run, the vehicles stop far from their real resting positions, the contact point or the contact plane assumed are not those of the case. Of the controls that govern that simulation, the one that weighs most on rotation is the impulse centre; restitution weighs least.
Where this calculation is done
- Desktop reconstruction programs — paid licence, and built for the expert who works with them daily. They solve momentum and also simulate the full dynamics in 2D or 3D.
- Module 03 of this calculator — in the browser, with nothing to install.
Over four of the RICSAC tests, and with the momentum-from-trajectories approach, module 03 gives 8.6 % mean error in speed. Four tests: not a large sample, and that is how the figure should be read. The detail is in how this calculator is validated.
Running it
Module 03 opens with a blank scaled plan, one metre per square, so the vehicles can be placed and measured without loading anything. If the scene diagram exists, it is loaded as an image and calibrated by clicking two points a known distance apart —a graphic scale, the width of the carriageway, the gap between two lines—: from then on everything measured on the plan is in real metres.
The vehicles are placed at rest, the approach and departure headings are marked, the post-impact runs are entered, and it returns the impact speeds of both. Afterwards, if wanted, the forward check is run on that same plan.
Frequently asked questions
- What is needed to calculate speeds from conservation of momentum?
- The masses of the two vehicles, the approach and departure heading of each one measured against the same reference axis, and the post-impact run of at least one of them to its resting position, with its coefficient of friction. No damage measurement is needed.
- What is the advantage over the damage-based calculation?
- It does not need crush measurements, so it works when the damage is not documented or the vehicle is no longer available. And because it starts from different data, it allows a cross-check: if the two routes agree, the reconstruction is solid.
- Which axis should the angles be measured against?
- Any axis, as long as it is the same one for all four headings: the north of the diagram or the axis of the road. What enters the equations are the directions relative to each other, not their absolute value.
Sources
The method and the data it has been checked against come from published work.
- Neades, J. (AiTS) and Smith, R. (De Montfort University). The Determination of Vehicle Speeds from Delta-V in Two Vehicle Planar Collisions, appendices B and E. The data for the twelve full-scale RICSAC tests (NHTSA, 1978), with measured impact speeds, come from there.
The checking of these methods against documented cases is detailed in how this calculator is validated.