Mass balance: delivered = bunker + hopper + belt + recycled + burned, checked in tonnes at every tick. Handling capacity is the minimum of the crane (bite × cycles per hour) and the belt (v / spacing items per second); if arrivals exceed it the bunker level rises until trucks queue (a simple single-server queue with utilisation ρ = λ/μ).
Energy: each burned item of mass m and lower heating value H gives thermal energy m·H and electricity E = η m H / 3.6 (kWh) with generating efficiency η = 0.28. Heating values here are MJ/kg: organic 4, paper 12, plastic 32, residual 9, glass and metal 0.
Sorters detect each item with probability equal to the accuracy plus a material bonus (magnet +8 %, plastic +4 %, glass +2 %, paper −4 %); the benchmark compares measured detection fractions with this expectation using a binomial standard error √(p(1−p)/n).
Simplifications: items are point-like pieces with fixed mass per type, the crane is a fixed-cycle server (28 s real time at 1x, an estimate of the 3D crane), moisture and ash are ignored, and plant time runs 8x faster than the wall clock. Educational model of a waste-to-energy plant, not a design tool.