Presets
Playback
Structure
Haul ramp
Sandstorm and earthquake
Playback
Camera
Layers
Quality
Model check
The science
Stress in a pyramid. The area at height z is proportional to (H - z)^2, so the weight above a level is rho g A(z) (H - z)/3 and the average vertical stress is
sigma(z) = rho g (H - z) / 3
exactly one third of what a straight tower of the same height carries (rho g (H - z)). For limestone of 2300 kg/m3 and the Great Pyramid's 146.6 m the base stress is about 1.1 MPa, far below the 20-40 MPa strength of the stone; the same stone could in principle carry a pyramid 3 sigma_c / (rho g) high, about 4 km.
Hauling a block. A sledge on a slope of angle theta needs F = m g (sin theta + mu cos theta). Damp sand packs and lets the sledge ride on a smooth surface: friction drops from about 0.55 on dry sand to about 0.25 when a few percent of water are present (as measured for dune sand and consistent with the tomb painting of Djehutihotep, where a man pours water in front of a sledge). The model uses mu(w) = 0.2 + 0.35 exp(-w / 1.6%). Each worker is assumed to pull 0.35 kN; the page counts workers per block, teams, round-trip time and blocks per day.
Windblown sand. The friction speed from the wind at 10 m is u* = 0.4 u10 / ln(10 m / z0) with roughness z0 = 1 mm. Grains move when u* > u*t = 0.1 sqrt((rho_s - rho_a) g d / rho_a) (0.23 m/s for 250 micron sand). Above it the Bagnold flux is q = 1.8 sqrt(d/D) (rho_a / g) u*^3 (1 - (u*t/u*)^2) kg per metre of width per second, so doubling the wind multiplies transport by about eight. Half of the flux is assumed trapped at the wall, piling up as a wedge at the repose angle (33 degrees): area = h^2 / (2 tan 33).
Wind load. Drag on the projected triangle, F = 1/2 rho_a v^2 Cd B H / 2, acting at H/3, against the weight times half the base: the safety factor against overturning is huge (hundreds) even in a severe storm; sand burial and abrasion, not toppling, are the storm's real risk.
Masonry stability. Stone carries almost no tension, so the resultant of weight and horizontal load must stay inside the middle third of the base (eccentricity e <= B/6). Wind adds e = F (H/3) / W; an earthquake with horizontal acceleration a adds inertia a W at the centroid (H/4), so e = a H / 4. The peak edge pressure on the square base is W/B^2 (1 + 6e/B), and a course of blocks slides when a/g > mu_s (stone on stone, about 0.6). At e = B/6 the lee edge just unloads and the windward edge carries twice the average pressure; beyond that a crack opens. A real pyramid is far inside these limits because it is so broad for its height.
Honest simplifications. A solid homogeneous pyramid (no chambers), stress is the average over the cross-section (the real stress field peaks near the centre of the base); the ramp is a single straight embankment (real builders probably used several designs); friction depends only on moisture; the sand wall is a repose wedge against a vertical wall, not a real dune migration; there is no saltation feedback or suspended dust.
Checks run in the page. Summed slab weights against rho g (H - z)/3; the sliding threshold tan(theta) = mu; the 250 micron threshold speed; u*^3 scaling; the repose-wedge area by numerical integration; the Great Pyramid base pressure. The height of the wind-drag resultant (H/3) by integration; the zero-pressure edge at e = B/6; the kern margin against the overturning factor.
What you see in 3D
The pyramid is built from forty stepped courses; their colour shows the average compressive stress against the stone strength (pale is safe, red is near failure). The dashed outline is the final shape. The sledge, the workers, and the sand wall on the windward face are drawn enlarged so that they stay visible; the numbers use the true dimensions. Everything comes from the same model as the 2D page.