# Total wheel loads: aero, transfer and estimation from data

> Each tyre’s vertical load is the sum of static weight, lateral and longitudinal transfer, aero load, banking/grade and bump content. You can estimate it from accelerations and setup data, or measure it via damper potentiometers or load cells, as long as you know what each method misses.

Area: Weight Transfer & Balance · Starting level: L2 Practitioner · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/wt-wheel-loads-aero-estimation

## L1 · Foundation

"The principal task of the race engineer is to balance the vertical loads acting on the tyres" (Segers ch10). Each wheel's load is a sum of contributions:

$$F_{z,i} = F_{z,i}^{static} \pm \Delta F_{lat} \pm \Delta F_{long} + F_{aero,i} + F_{bank/grade} + F_{bump}$$

**Worked example: GT3 outside front (left front in a right-hander)**, 180 km/h, trail braking at 0.8 g with 1.4 g lateral. $m = 1300$ kg, 45% front, $h = 0.45$ m, $L = 2.70$ m, $t = 1.65$ m, front LLTD 49%, $C_LA = 1.5$ m², aero balance 40% front.

| Contribution | LF (N) | RF (N) |
|---|---|---|
| Static (585 kg / 2) | 2869 | 2869 |
| Longitudinal (1700 N total, half per side) | +850 | +850 |
| Lateral front share (49% of 4870 N) | +2386 | −2386 |
| Aero (2297 N × 40% / 2) | +459 | +459 |
| **Total** | **6564** | **1792** |

The outside front carries **3.7×** the inside front's load. Through tyre load sensitivity, the front pair is far from its ideal-grip state, which is why entry balance is so sensitive to LLTD, brake bias and aero balance under pitch.

**Aero load** is $\frac{1}{2}\rho C_LA V^2$, split front/rear by the aero balance (which moves with ride height and pitch). Orders of magnitude for $C_LA$: GT3 ~1–2 m² (BoP and configuration dependent), LMP2/Hypercar ~3–4 m², F3 ~2–3 m², F2/Super Formula ~3.5–4.5 m². At 250 km/h, each 1 m² of $C_LA$ is ~2.95 kN of downforce.

## L2 · Practitioner

**Estimating wheel loads from data (Segers §10.4).**

*Method 1: from accelerations (simulation-style math channels).* Use logged $a_y$, $a_x$ and speed, plus setup constants (masses, CG, RCs, roll stiffness split, $C_LA$ and balance). This is a model, not a measurement. It is good for understanding the order of magnitude and for corner-by-corner comparison, but it inherits every error in the constants.

*Method 2: from damper potentiometers.* Convert travel to wheel force for each element:
- spring: $F_{wheel} = k_s\,x_{damper}\,MR$ (MR = damper/wheel travel, the academy convention, so wheel rate = $k_s MR^2$), plus preload. Segers defines MR the other way round (wheel/damper; §10.4 Eq. 10.22, also used in Eq. 9.1): invert it when you use his equations;
- bump rubber/packer: non-linear $F(x - x_0)$ once the gap $x_0$ is closed;
- ARB: proportional to the left–right travel difference, opposite sign on each side;
- damper: from a velocity–force map (separate low/high speed bump/rebound).

The sum gives the **sprung (elastic) load** change at each wheel. Add the unsprung and geometric transfer by calculation, because the pots do not see them.

*Method 3: pushrod/strain-gauge load cells.* Direct, high resolution and independent of rate assumptions. But again, **not** the unsprung, geometric or wheel-generated aero forces (open-wheel tyres and wheel fairings generate lift/downforce outside the sprung path).

**Total download channel.** Sum the four wheel loads (or load cells) minus static weight: you get the aero load (plus banking and vertical-curvature effects). Plotted against $V^2$ on straights and steady corners, the slope gives $\frac{1}{2}\rho C_LA$, an on-track $C_LA$ estimate. The front/rear split gives aero balance vs speed. On an aero car the total-load trace looks like the speed trace squared; on a GT4 it barely moves.

**Modal view.** Decompose the four damper travels into heave, pitch, roll and warp (Segers §10.5). Each mode has its own stiffness, which gives clean per-mode forces and shows warp (diagonal) load from kerbs or chassis twist.

## L3 · Advanced

**Error budget of each method.**
- **Acceleration-based**: CG height (±10–20 mm typical uncertainty unless measured by tilting, Segers' raised-axle method), roll-centre migration, $C_LA$ and aero balance changing with ride height, fuel burn. Use it for relative comparisons and to fill in components the sensors can't see.
- **Pot-based**: motion-ratio non-linearity (MR changes through travel by 5–15% on many pushrod layouts), spring-rate tolerance (±3–5%), packer and bump-rubber curves (temperature dependent), damper hysteresis and friction. Zeroing: pots must be zeroed at static ride height with the correct fuel and driver mass.
- **Load cells**: temperature drift, crosstalk from bending, calibration across the load range. They are the most accurate for the sprung path.

**Tyre spring as a load sensor.** Ride-height lasers measure to the ground, while pots measure suspension travel. Their difference, after geometry, is **tyre deflection**, and with a known tyre vertical rate (a few hundred N/mm for race slicks, pressure dependent) it is a load estimate that includes *everything* (unsprung, geometric, wheel aero). Segers §10.7 discusses tyre spring rates. It is noisy but a useful cross-check.

**Consistency checks you should automate:**
- Sum of wheel loads on a straight at constant speed = $mg + \frac{1}{2}\rho C_LA V^2$. A drift over a stint = fuel burn (good) or a sensor problem.
- Left–right sum difference ÷ $a_y$ should be stable (≈ $2\,m h/t$ × the share seen by the sensors, because one side gains $\Delta W$ while the other loses it). A change between runs without a setup change means damper/packer/bar link issues (Segers §9.6 troubleshooting).
- Front LLTD from load cells vs setup calculation, binned by $a_y$ and speed. Divergence at high speed = packers/bump rubbers or aero-induced ride-height effects.

**Using wheel loads for engineering decisions:** inside-wheel load near zero (TCR rear, GT4 inside front on kerbs) shows LLTD saturation. Minimum front load in high-speed braking zones guides bias and anti-dive. The outside-front load peak during trail braking sets the tyre's worst-case thermal input.

## L4 · Expert

**From wheel loads to tyre-level understanding.** Combine estimated $F_z$ per wheel with $a_y$/$a_x$ and a mass-distribution model to estimate per-axle lateral force (moment balance with yaw acceleration $I_z\dot r$). Then normalise: axle $F_y / \sum F_z$ is the **axle friction utilisation**. Plotting front and rear utilisation through a corner directly shows which axle limits the car in each phase, a data-driven pair analysis (RCVD Ch. 7). It is the most direct objective balance metric available without tyre force transducers.

**Wheel force transducers** (WFT, 6-axis hubs) remove most of the guesswork but are expensive, heavy and usually test-only. In series that allow them in testing, a single WFT test can calibrate your pot- and load-cell-based models for the season.

**Dynamic content.** Bumps and kerbs create high-frequency load variation. The **standard deviation of tyre load** (or of pushrod load), band-passed around wheel-hop frequency (~10–20 Hz), is a mechanical-grip metric. Because of load sensitivity, a tyre with large load variation produces less average force than a steady one. That is the quantitative basis for damper and spring tuning for grip (see the dampers domain).

**Limits:** none of these methods separate *where on the tyre* load is carried (camber, pressure distribution). And all of them assume the vehicle model (masses, geometry) is right. After a crash repair, re-verify CG, motion ratios and roll centres before trusting historical comparisons.

## Key formulas

**Wheel load sum**

$$F_{z,i} = F_{z,i}^{static} \pm \Delta F_{lat,i} \pm \Delta F_{long,i} + F_{aero,i} + F_{bank/grade,i} + F_{bump,i}$$


**Aero load per axle**

$$F_{aero,f} = \beta_a\,\tfrac{1}{2}\rho\,C_L A\,V^2, \quad F_{aero,r} = (1-\beta_a)\,\tfrac{1}{2}\rho\,C_L A\,V^2$$

- \beta_a: aero balance (front fraction)
- C_L A: downforce coefficient × area (m²)

**Spring force at the wheel from damper pot**

$$F_{spring@wheel} = k_s\,x_{damper}\,MR$$

- k_s: spring rate (N/mm)
- x_damper: spring/damper travel from static (mm)
- MR: motion ratio (damper travel / wheel travel); Segers uses the inverse

**On-track downforce coefficient**

$$C_L A \approx \dfrac{\sum F_{z,i} - m g}{\tfrac{1}{2}\rho V^2}$$


## Trackside takeaways

- Wheel load = static + lateral + longitudinal transfer + aero + banking/grade + bump content.
- In trail braking a GT3 outside front can carry ~3–4× the inside front’s load.
- Pots and load cells see only the sprung/elastic path; add unsprung, geometric and wheel aero by calculation.
- Sum of wheel loads vs V² on straights gives an on-track C_LA and aero balance estimate.
- Automate consistency checks: they catch sensor, packer and bar-link problems before they mislead setup work.

## By car category

### GT3

Damper pots are standard and load cells are less common in customer teams, so pot-based wheel-load channels are the norm. Aero is moderate, so mechanical transfer dominates wheel loads in most corners.

### LMP2

Aero load can exceed the car’s weight at top speed, so total-load channels track V² closely and packer engagement makes per-wheel loads strongly non-linear. Load-cell or pot-plus-packer models are essential.

### FIA F2

High downforce on a light car: the aero term dominates at speed. Wheel-generated aero forces (exposed tyres) are not seen by pushrod load cells, a known source of discrepancy in total-load channels.

### GT4

Low downforce: wheel loads are almost purely static plus transfer, so acceleration-based channels are already quite accurate and a good teaching ground.

### Hypercar (LMH/LMDh)

Heavy cars with high downforce and regulated aero windows: total download estimates and aero balance vs ride height are tracked closely against the homologated aero map and BoP constraints.

### One-make Cup

Limited sensors (often no pots on some one-make cars, depending on series rules) mean acceleration-based wheel-load estimates and driver feedback carry more weight. Knowing the car’s CG and roll distribution from the manufacturer data is valuable.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §10.4 Total Wheel Loads
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §10.5 Determining Wheel Loads with Modal Analysis
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §10.6 Measuring Wheel Loads with Suspension Load Cells
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §10.7 Tire Spring Rates
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 18 Wheel Loads
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 7 Steady-State Pair Analysis (Axle utilisation view)
