❄️ District Cooling Plant

A cooling plant has five chillers of differing size that together must meet a fixed cooling demand. We choose how to split the load. Each chiller's coefficient of performance peaks near a part-load sweet spot and falls off when run nearly idle or flat out, so the electrical energy to deliver the load depends strongly on the split. The optimum keeps the running chillers near their efficient point rather than dumping everything on the largest unit. Score is total electrical energy, to minimise. 5-D.

Algorithm

Plant energy
Splits tried0
Best so far

Leaderboard

AlgorithmEnergyEvals
— no runs yet —

What's happening

The JS objective is a line-for-line port of example_applications/district_cooling and agrees to floating-point tolerance. Bars show each chiller's load as a fraction of its capacity; the marker is the efficient part-load sweet spot. Good splits sit the running chillers near their markers.