Every small mirror (heliostat) tilts to bisect the angle between the incoming sunlight and the direction to the central receiver on the tower — the same "angle of incidence = angle of reflection" law a bathroom mirror obeys. Each mirror's useful output is reduced by the cosine of the angle between its face normal and the sun (the cosine efficiency): mirrors facing the sun almost edge-on contribute little, which is why real CSP fields are laid out asymmetrically around the tower rather than in a plain square grid.
normal = normalize(sunDir + towerDir)
P_field = Σ DNI · cos(normal·sunDir) · A_mirror
dT/dt = (P_field − P_loss − P_extracted) / thermalMass
P_elec = η_turbine · P_extracted (only once T ≥ operating threshold)
- Sun elevation — lower sun angles worsen everyone's cosine efficiency at once, exactly like early morning or winter reduces a real plant's output.
- Heliostat field — more mirrors mean more collection area, but each added ring sits at a worse average angle.
- Direct irradiance (DNI) — the raw solar resource; only the direct beam concentrates, so haze or high humidity (lower DNI) hurts CSP far more than it hurts flat-panel PV.
- Turbine draw — how aggressively the plant pulls stored thermal energy off the receiver/salt loop to spin the turbine; draw too hard and the receiver cools faster than the sun can reheat it.
- Pass cloud — briefly zeroes the direct beam, showing the thermal mass of molten salt buffering the turbine through a short outage — the core reason CSP can dispatch power PV cannot.
Real-world relevance: this is the "power tower" CSP design (e.g. Ivanpah, Gemasolar) — heliostats replace the parabolic trough's single curved mirror with many small tracking flats, and the receiver's molten-salt thermal mass is what lets a CSP plant keep generating for hours after sunset, unlike photovoltaic panels.