Stockpile Blending — Chevron Stacking & Bridge Reclaiming (2D)
A chevron stacker lays fluctuating-grade ore into a pile in horizontal layers over time; a bridge reclaimer then slices vertically through every layer at once and averages them, turning a noisy feed into a steady output — with input and output coefficient-of-variation computed live from the real simulated data.
This 2D companion drives the real math behind stockpile homogenization instead of orbiting a decorative 3D plant scene: a genuinely noisy, drifting ore-grade feed is sampled while a chevron stacker lays it down in horizontal layers across the pile, then a bridge reclaimer cuts a vertical slice through every completed layer at once and averages them. Because each column of the pile mixes material deposited at many different moments, the reclaimed output is measurably steadier than the raw feed — and both the input and output coefficient of variation (CV = standard deviation ÷ mean) are computed live from the actual sampled numbers, not a scripted "improvement" figure, so pushing the layer count or feed variability sliders visibly changes the real blending ratio.
Feed grade is generated as a slow sinusoidal drift plus a damped random walk. Each stacking pass samples this signal once per pile column and writes it into that layer; the reclaimer averages every layer at a column when it sweeps through. Input/output CV and the resulting variance-reduction ratio are recomputed every frame from the real sample arrays.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install