COP: efficiency at one point
COP = cooling capacity (kW) ÷ power input (kW)
A chiller delivering 600 kW of cooling while drawing 100 kW has a COP of 6.0. The figure is meaningful only when the operating conditions are stated: chilled water temperatures, condensing temperature and refrigerant all affect it. A COP quoted without conditions is a marketing number.
| Machine type | Typical COP at design condition |
|---|---|
| Air-cooled scroll | 2.8–3.2 |
| Air-cooled screw | 3.0–3.4 |
| Water-cooled screw | 4.5–6.0 |
| Water-cooled centrifugal | 6.0–6.8 |
| Magnetic-bearing centrifugal | 6.3–7.0 |
IPLV: efficiency across the year
IPLV — integrated part load value — is a weighted average of efficiency at four load points, using weights that reflect how buildings typically operate.
IPLV = 1 ÷ (w₁/A + w₂/B + w₃/C + w₄/D)
Where A, B, C and D are the COP values at 100%, 75%, 50% and 25% load, and w are the weighting factors.
| Load point | AHRI 550/590 | GB/T 18430.1 (China) |
|---|---|---|
| 100% | 1% | 2.3% |
| 75% | 42% | 41.5% |
| 50% | 45% | 46.1% |
| 25% | 12% | 10.1% |
The weights are heavily concentrated at 50%–75% load because that is where most chillers spend most of their hours. This is the single most important fact about IPLV: it reflects reality, whereas COP reflects a condition the plant may only see on three afternoons a year.
Note that the two standards use slightly different weights. When comparing machines, confirm that both figures were calculated on the same basis — mixing an AHRI IPLV with a GB/T IPLV is not a comparison.
A worked example
Two chillers, both 600 kW:
| Load point | Machine A COP | Machine B COP |
|---|---|---|
| 100% | 6.2 | 5.8 |
| 75% | 6.0 | 6.4 |
| 50% | 5.4 | 6.8 |
| 25% | 4.6 | 5.9 |
| IPLV (AHRI weights) | 5.52 | 6.50 |
Machine A has the better full-load COP; machine B has the better annual efficiency, by a margin of about 18%. In a building running 4,000 equivalent full-load hours a year at 600 kW, that difference is roughly 66,000 kWh — a number that dwarfs any plausible difference in purchase price.
Why part-load efficiency varies so much
At 50% load, a fixed-speed screw compressor may unload by sliding the slide valve, which is efficient, or by cycling, which is not. A centrifugal machine at low load approaches surge and needs anti-surge control, which costs efficiency. A machine with a variable-speed drive can slow the compressor instead of unloading it, holding efficiency much further down the curve.
| Technology | Part-load behavior |
|---|---|
| Fixed-speed screw with slide valve | Good to 50%, then efficiency drops |
| Multiple compressors, staged | Good, provided staging is set up correctly |
| Variable-speed screw | Strong across the range |
| Centrifugal with VSD | Strong, but surge control needed at low load |
| Magnetic-bearing centrifugal | Strong, no oil system losses |
What to specify
- COP at the project’s design conditions, not at generic catalog conditions.
- IPLV at the standard weights applicable in your market, so the figures are comparable.
- The load points, not just the weighted result — a single good IPLV can hide a poor value at 25%, which matters for a plant with long low-load nights.
- The compressor technology and control strategy, because these determine whether the catalog curve is achievable in practice.
Where the numbers mislead
IPLV assumes a generic building. A data center running at 90% load for 8,000 hours a year gains almost nothing from a good IPLV; it needs the highest full-load COP. A hotel with a long low-load night benefits enormously from a good 25% point. Match the metric to the load profile.
Catalog values are at clean conditions. Fouling, refrigerant charge loss and degraded water treatment all reduce real efficiency. A machine that meets its IPLV on the test stand may not meet it in year three.
Auxiliary power is excluded. Cooling tower fans, condenser water pumps and chilled water pumps can add 15%–25% to plant power. Optimizing the chiller while ignoring the auxiliaries is optimizing the wrong half of the bill.
Frequently asked questions
Which is more important, COP or IPLV?
For a plant that runs at part load most of the year, IPLV. For a plant at steady near-full load, COP. Most HVAC plants are the former; most process plants are closer to the latter.
Can IPLV be higher than COP?
Yes, and it usually is, because part-load efficiency is often better than full-load efficiency. A high IPLV with a low COP is not a contradiction — it describes a machine optimized for the way buildings actually run.
Is a higher IPLV always better?
Only when the load profile matches the weighting. Check the individual load points against your expected operating hours rather than relying on the composite figure.
How is IPLV measured?
On a test stand, at specified part-load ratios with defined condensing conditions, following the applicable standard. Field performance will differ because field conditions differ.
Does refrigerant choice affect the figures?
Yes, indirectly: different refrigerants have different theoretical cycle efficiency, and charge limits or phase-down rules may restrict which machines are available in a given market.
What efficiency should I expect in a tender?
Require the supplier to state COP and IPLV at the project conditions, together with the standard used and the individual load points. A tender that quotes a single COP figure without conditions cannot be evaluated.
Longkang Environmental Equipment Co., Ltd. quotes chiller efficiency at the project’s operating conditions and provides both COP and IPLV with the calculation basis, so the figures can be compared rather than merely read. Phone: +86 400-878-0280 | +86 130-0200-1280
Related reading