Why inverter compressors save energy at part load — the hidden physics
By DigiEntropy Engineering · 2026-07-18 · 9 min read
A variable-speed compressor and a fixed-speed one have almost the same COP at any single operating point. So where do the savings come from? Not the compressor — the coils. Here is the physics, with real polynomial data.
About this post — Authored by an AI assistant using DigiEntropy's polynomial corpus, the universal compressor predictor, and the project's chart-generation tooling. Charts are produced by Python scripts that read the same database the live site queries; tables and formulas are pulled from the same engineering modules. Findings reflect the project's current dataset and methodology — send corrections or deeper questions to admin@digientropy.com. A user sent us a sharp question. In our Operation Simulation tool, they had put a variable-speed (inverter) compressor next to a fixed-speed one on the same cold room. The inverter came out using more energy over the day. That is against everything the brochures promise — so either the brochures are wrong, or the model was. We went and checked, as a refrigeration engineer would rather than a marketer. The answer turned out to be a genuinely interesting piece of physics that most people get slightly wrong, and it is worth writing down. The uncomfortable starting point Take any compressor's published performance polynomial and evaluate its cooling capacity (Q) and power (P) at a fixed evaporating and condensing temperature, at 100%, 75% and 50% capacity. Do the division. The COP is the same at every capacity, to three decimal places. That is not a rounding artefact — it is how the data is built. A variable-speed compressor at half speed moves half the refrigerant and does half the work, so Q and P both halve and Q/P does not move. Here is a real Bitzer-family part-load set from our database, plotted two ways. The grey line is the world most sizing tools quietly assume: the evaporating and condensing temperatures are pinned by a fixed "approach" to the room and the ambient, so they never move, and the part-load COP is flat. In that world an inverter is exactly as efficient as an on/off machine at any given moment — and then it loses, because a real inverter carries a drive that wastes about 2% of the power as heat. That is precisely the "inverter uses more energy" result our user saw. The old version of our model made that fixed-approach assumption. It was not lying; it was incomplete. Where the savings actually live: the coils, not the compressor The green line is what a real plant does. The trick is that the evaporator and condenser are fixed pieces of hardware. Each one can only move heat in proportion to the temperature difference across it: $Q = UA \cdot \Delta T$ At design load the evaporator needs, say, a 7 K difference between the room air and the boiling refrigerant. At half load only half as much heat is…