Physics I course sequence · Interactive lesson
Estimation and Model Checks
Four cheap tests stand between a calculation and a result worth using: size, units, limiting cases, and physical plausibility. Each takes seconds, and each catches a mistake the others miss.
UPI-01.08 · Lesson 8 of 9 in this unit. Position is not a completion score.Build the model
An estimate predicts a scale. A model check tests the result independently.
An estimate is a deliberately rough calculation aimed at the right power of ten. A model check is a fast test — units, a limiting case, or a known benchmark — that a candidate result must pass before it is believed.
- Estimate sizePredict a power of ten.
- Check unitsMatch dimensions on both sides.
- Test limitsRecover a trusted simple case.
- Check plausibilityCompare with physical evidence.
A failed check is a reason to investigate the result, assumptions or benchmark. Passing all four is not proof.
Order of magnitude
Choose the nearest power of ten on a logarithmic scale.
q = Q/U > 0
n = round(log₁₀ q)
Q ≈ 10ⁿ U
For power, use U = 1 W. The boundary between 10³ W and 10⁴ W is 10³⋅⁵ W ≈ 3162 W.
Fermi decomposition
Split an unfamiliar quantity into factors you can estimate or bound.
Q ≈ q₁ × q₂ × … × qₙ
Human height ≈ 1.8 m → nearest-power scale 10⁰ m, or 1 m. A scientific-notation exponent is not always the nearest-power exponent.
What an order-of-magnitude claim actually says
Choose a reference unit before taking a logarithm: q = Q/U is a numerical value. For a positive q, this lesson uses n = round(log₁₀ q), so Q is classified by the nearest power 10ⁿ U on a logarithmic scale. Values from 10(n − ½) U up to, but not including, 10(n + ½) U share that classification; exact ties go to the higher power. This is a rounding convention, not a guarantee that an estimated value lies near the unknown true value. State assumptions, bracket uncertain inputs, and investigate a large disagreement with the finished calculation.
Worked estimate: cyclist drag power
Start with bounded factors, then calculate the scale.
P = Fᵈv = ½ρCᵈAv³
| Air density ρ | 1.2 kg/m³ |
|---|---|
| Drag area CᵈA | 0.3 m² |
| Speed v | 8 m/s |
P = ½(1.2)(0.3)(8³)
≈ 92 W ≈ 9.2 × 10¹ W
Nearest-power scale: 10² W. This is power to overcome air drag, not the rider’s total metabolic power.
Break the unknown into factors you can bound
Split an unfamiliar quantity into factors you can estimate. For a cyclist, take air density ρ ≈ 1.2 kg m⁻³, drag area CdA ≈ 0.3 m² and speed v ≈ 8 m s⁻¹ in still air. P = ½ρCdAv³ gives about 92 W, so the nearest-power scale is 10² W. This is aerodynamic power only; rolling resistance and other losses require extra terms. Multiplicative errors can offset or reinforce each other—cancellation is not guaranteed. For a positive bracket, √(qlow qhigh) is the multiplicative midpoint; retain the bracket to communicate uncertainty.
Units are the cheapest lie detector
Every added term and both sides of an equation must have matching dimensions.
Speed
[v] = LT⁻¹
Acceleration
[a] = LT⁻²
Energy
[E] = ML²T⁻²
The arguments of sin, ln and exp must be dimensionless: sin(θ), ln(x/L) and exp(−t/τ). A dimensionless numerical factor or wrong sign can still survive a unit check.
Units are the cheapest lie detector
Both sides of an equation must have the same dimensions, as must all terms added together. Speed has dimensions L T⁻¹, acceleration L T⁻² and energy M L² T⁻². The arguments of sin, ln and exp must be dimensionless: in exp(−bt/m), b therefore has units kg s⁻¹. Check the symbols before inserting numbers. A wrong dimension exposes a fault, but a correct dimension does not detect a wrong sign or a missing dimensionless factor.
Run all four checks: a car at 30 m/s
Use CᵈA = 0.8 m² and ρ = 1.2 kg/m³.
Calculate & check units
P = ½(1.2)(0.8)(30³)
= 12,960 W ≈ 1.3 × 10⁴ W
[P] = (kg m⁻³)(m²)(m³ s⁻³) = kg m² s⁻³ = W.
Compare the size correctly
Pcar / Pcyclist
= (0.8 / 0.3)(30 / 8)³
≈ 141
Both drag area and speed change. Using the speed factor alone would miss a factor of about 2.7.
Test a limit
At fixed area, v → 0 gives P → 0. Doubling v multiplies P by 8.
Check plausibility
A roughly 13 kW aerodynamic load is a plausible scale. Other road loads and drivetrain losses still need to be included.
Run the four together on one result
For a car with drag area 0.8 m² at 30 m s⁻¹ in still air of density 1.2 kg m⁻³, aerodynamic power is ½(1.2)(0.8)(30³) = 12,960 W, or about 1.3 × 10⁴ W. Its dimensions are (M L⁻³)(L²)(L³ T⁻³) = M L² T⁻³. Compared with the cyclist, both area and speed change: the factor is (0.8/0.3)(30/8)³ = 140.625. Zero speed gives zero power; doubling speed at fixed drag area gives eight times the power. A roughly 13 kW aerodynamic load is a plausible scale, not proof of total cruising consumption or of the drag coefficient chosen.
Limiting cases
Recover a known model within the stated assumptions.
Ideal Atwood machine
a = (m₁ − m₂)g / (m₁ + m₂)
m₁ = m₂ → a = 0.
m₂ → 0 → a → g.
Swap masses → reverse the sign.
Small-angle approximation
sin θ ≈ θ
Use radians and |θ| ≪ 1. At θ = 0.1 rad, relative error is about 0.167%.
Binomial approximation
(1 + x)ⁿ ≈ 1 + nx
First order as x → 0 for fixed n. Check neglected terms; large n may need a much smaller x.
Limiting cases audit the structure
Test limits only where the model or its stated idealisation applies. For a frictionless, massless-pulley Atwood machine, a = (m₁ − m₂)g/(m₁ + m₂), with positive acceleration chosen toward the heavier first mass. Equal masses give zero; m₂ tending to zero gives g; exchanging masses reverses the sign. For angles in radians, sin θ ≈ θ when |θ| is small. For a fixed exponent n, the first-order binomial approximation is (1 + x)ⁿ ≈ 1 + nx as x tends to zero; if n is large, a correspondingly smaller x may be needed. A failure may reveal an algebra error or an attempt to use a model outside its domain.
Plausibility anchors
Context matters: these are approximate reference scales.
Plausibility is a physical claim, not a feeling
Compare with physical scales under stated conditions: g ≈ 9.81 m s⁻² near Earth, walking around 1.4 m s⁻¹, atmospheric pressure around 10⁵ Pa, water density around 10³ kg m⁻³, and air density around 1.2 kg m⁻³ near sea level. Sustained human mechanical power is roughly 10² W and varies strongly with the person and task. A 1500 kg car at 30 m s⁻¹ carries 675,000 J, comparable to raising it about 46 m against gravity. Apparent contradictions need a defined quantity: an ordinary energy-conversion efficiency cannot exceed one, but a heat pump’s coefficient of performance is not that efficiency. Report only the precision supported by the inputs.
Interactive drag-power lab
Predict what will happen, then change speed or drag area.
Air density fixed at 1.2 kg/m³.
Minimum positive inputs: 0.001 m/s and 0.000001 m²; exact zero is also supported.
Drag area = drag coefficient × frontal area.
Nearest-power band: [31.6228, 316.2278) W. Doubling speed at this drag area multiplies power by 8. The displayed digits help you inspect the calculation; rough inputs do not justify that measurement precision.
Model: P = ½ρCᵈAv³, for constant drag area and still air. Aerodynamic power only: no rolling resistance, hills, acceleration or drivetrain losses. It is not a fuel-consumption prediction.
Explore the original two-slider model
Original ranges: 4–35 m/s and 0.2–1.0 m². The graph’s 1 W floor is below all results within those ranges; it is not a replacement for zero power. Shading is a classification band, not an uncertainty interval.
At fixed drag area, doubling speed multiplies power by eight. The nearest-power band changes when P crosses a boundary 10(n + ½) W.
DRAG POWER P92 W
log₁₀ P1.96
ORDER n2
P ÷ 10ⁿ0.92
Live interpretationDRAG POWER P: 92 W. log₁₀ P: 1.96. ORDER n: 2. P ÷ 10ⁿ: 0.92
Catch the common trap
Passing these checks is not proof.
- Unit checks cannot detect a missing ½, 2π, efficiency factor or sign.
- A limit tests one part of a model, not every regime.
- A nearest-power band is not a guaranteed uncertainty range.
- Fermi-factor errors can reinforce; they do not have to cancel.
Quick check
Five questions · practice for this visit
Review or reset practice
Refreshing clears these answers. Existing saved progress is unchanged.
Practice & worked examples
Explain what each check establishes—and what it does not.
Easy · Order of magnitude
A finished calculation returns 4.7 × 10³ W. State its order of magnitude, and give its nearest-power classification band.
Show worked solution →
- Take the base-10 logarithm of P/(1 W): log₁₀(4.7 × 10³) = 3 + log₁₀ 4.7 = 3 + 0.672 = 3.672.
- Round to the nearest integer: n = 4, so the claim is P ≈ 10⁴ W. The decade boundary sits at √10 × 10³ = 3.2 × 10³ W, and 4.7 × 10³ W is above it.
- The nearest-power classification covers 103.5 to 104.5, that is 3.2 × 10³ W up to 3.2 × 10⁴ W — anything from 3.2 kW to 32 kW has the same nearest-power classification. This is not a guaranteed uncertainty interval.
Order 10⁴ W, with a classification band from roughly 3.2 × 10³ W to 3.2 × 10⁴ W.
Medium · Delivery van drag estimate
A delivery van has drag area CdA = 1.2 m² and cruises at 25 m s⁻¹ through air of density 1.2 kg m⁻³. Estimate the power it spends pushing air aside, then run the unit, size and plausibility checks on the answer.
Show worked solution →
- Fermi decomposition: P = ½ρCdAv³, and every factor is bounded — ρ = 1.2 kg m⁻³, CdA = 1.2 m², v = 25 m s⁻¹.
- Prefactor: ½ × 1.2 × 1.2 = 0.72 kg m⁻¹. Speed cubed: 25³ = 1.5625 × 10⁴ m³ s⁻³.
- P = 0.72 × 1.5625 × 10⁴ = 1.125 × 10⁴ W, so P ≈ 1.1 × 10⁴ W — two figures is all the inputs earn.
- Units: (M L⁻³)(L²)(L³ T⁻³) = M L² T⁻³, the dimensions of power. Size: log₁₀(P/(1 W)) ≈ log₁₀(1.125 × 10⁴) = 4.05, so the order is 10⁴ W. Plausibility: about 11 kW is a reasonable scale for this aerodynamic load; it is not the total cruising power, and an engine rating alone cannot validate it.
P ≈ 1.1 × 10⁴ W, order 10⁴ W; units, size and plausibility all pass.
Hard · Linear-drag model
An object falling with linear drag obeys v(t) = (mg/b)(1 − e(−bt/m)). Audit the formula with a unit check and two limiting cases, then put numbers on it for m = 0.145 kg and b = 0.020 kg s⁻¹.
Assume release from rest, downward-positive velocity, constant m, b and g, and drag force −bv. This ideal model requires a regime where linear drag applies.
Show worked solution →
- Units: the exponent bt/m must be a pure number, so [b] = M T⁻¹, that is kg s⁻¹. Then [mg/b] = (M · L T⁻²)/(M T⁻¹) = L T⁻¹, a speed — the prefactor is dimensionally allowed to be v.
- Limit t → ∞: e(−bt/m) → 0, so v → mg/b, a constant. That is terminal velocity, where the drag bv has grown to balance the weight mg.
- Limit bt/m ≪ 1: expand e(−x) ≈ 1 − x, so v ≈ (mg/b)(bt/m) = gt. Early on, before drag matters, the formula collapses onto free fall.
- Numbers: mg = 0.145 × 9.81 = 1.42245 N, so mg/b = 1.42245/0.020 = 71 m s⁻¹, and the time constant is m/b = 0.145/0.020 = 7.25 s.
- Model check: 71 m s⁻¹ and 7.25 s follow from the assumed linear law. These scales do not establish that linear drag remains valid for a real object at that speed. Compare with measurements and the relevant flow regime.
[b] = kg s⁻¹; v → mg/b = 71 m s⁻¹ as t → ∞; v → gt as t → 0. Terminal speed 71 m s⁻¹, drag time constant 7.25 s.
Calculated values are vterminal = 71.1225 m/s and τ = 7.25 s. Extra digits show the arithmetic, not measurement precision. For b → 0 at fixed t, the limit is gt; b = 0 must be handled separately.
Summary: estimate first, then test the result
State assumptions, estimate a scale, check units, test a limit, and compare with physical evidence. Investigate disagreements and report only justified precision.
Open all six formulae and their conditions
Order of magnitude
q = Q/U > 0; n = round(log₁₀ q); Q ≈ 10ⁿ U
Nearest-power convention. U is the stated reference unit; this is not an uncertainty interval.
Fermi decomposition
Q ≈ q₁ × q₂ × … × qₙ
Estimate or bracket each factor. Errors can reinforce as well as cancel.
Unit check
[LHS] = [RHS]; [v] = LT⁻¹; [a] = LT⁻²; [E] = ML²T⁻²
Match every added term. Arguments of sin, ln and exp are dimensionless.
Aerodynamic power
P = ½ρCᵈAv³
Constant density and drag area, still air; excludes other loads and losses.
Limiting cases
a = (m₁ − m₂)g/(m₁ + m₂)
Ideal Atwood machine: equal masses → 0; m₂ → 0 gives a → g.
Plausibility anchors
g ≈ 9.81 m/s²; ρair ≈ 1.2 kg/m³; ρwater ≈ 10³ kg/m³
Approximate contextual benchmarks, not exact values in all conditions.