Heat pumps · Chillers · Sludge drying systems
Industrial cooling · Process

Chiller Retrofit Payback: Baseline and Sensitivity

Every chiller retrofit proposal ends with a payback number, and the same scope can produce 3.2 years in one proposal and 8 years in another. The difference is rarely equipment price. It is usually the baseline, the investment boundary, or whether the calculation is static or discounted.

This article sets out the three definitions, then works through an 800 kW plant to show how the baseline and savings are built. It closes with a sensitivity table for tariff, running hours and load factor, because those three variables decide whether the project is credible.

1. Define three things before calculating

Baseline. Savings = pre-retrofit annual electricity minus post-retrofit annual electricity. The first term must come from measurement, not from nameplate power multiplied by running hours. Actual load factor is often 50–60% of nameplate, so a nameplate-based baseline can double the apparent saving.

Investment boundary. Replacing only the chiller gives the longest payback. Including pump variable-speed drives, group control, cooling-tower optimization and water-side cleaning raises total investment by 15–25%, but can raise savings by 30–40%, which shortens payback. The boundary must be stated, or two proposals are not comparable.

Static versus dynamic. Static payback = investment ÷ annual net saving. It ignores the time value of money. Dynamic payback discounts annual cash flows and is typically 20–30% longer. Present both: static for ranking options, dynamic for financial approval.

Basis Method Use Common bias
Static Investment ÷ annual saving Comparing options Overstates benefit; ignores discount and degradation
Dynamic Discounted cash flow to zero NPV Financial decisions, loans Result depends on discount rate
Whole-life Includes maintenance, replacement, residual value Comparing equipment over its life Many assumptions; easy to manipulate

2. Build the baseline from meters

If sub-metering exists, take 12 consecutive months of chiller-plant electricity and subtract non-cooling loads. The 12 months must cover both the cooling season and the shoulder season. If a month is missing, use the same month from the previous year rather than a seasonal average.

For an older plant without sub-meters, estimate with the load-factor method:

Annual electricity (kWh) = installed cooling capacity (kW) × equivalent full-load hours (h) × average load factor ÷ plant efficiency.

Equivalent full-load hours come from operating records. Commercial buildings are often 1,800–2,600 h; offices 1,500–2,000 h; data centers above 6,000 h; industrial process cooling can reach 4,000–6,000 h depending on shifts. The average load factor is the mean of the hourly load over the cooling season; without records, use 55–65% of installed capacity.

The most common error is treating installed capacity as actual cooling. In a plant with multiple chillers, actual delivered cooling is often only about 60% of installed capacity. Using the installed figure overstates the baseline and therefore the saving.

3. Calculate the four measures separately

High-efficiency chillers. The saving depends on the difference in integrated part-load value (IPLV), not the nominal full-load COP. An old plant may have an IPLV of 4.0–4.5; a new Tier 1 machine under GB 19577-2024 may reach 6.0–6.5. The saving is approximately 1 − old IPLV ÷ new IPLV.

Pump variable-speed drives. Take two 45 kW chilled-water pumps and two 45 kW condenser-water pumps running 3,200 h/year. Do not use rated power × a savings percentage: the pumps usually run at 60–70% of rated power and not at full speed all year. Use measured current to establish the baseline. The annual saving is often 25,000–30,000 kWh.

Group control and water-temperature reset. Optimized staging and supply-temperature reset can reduce plant energy by 6–10%. This overlaps with chiller replacement, so the two savings cannot simply be added. Apply an overlap factor of about 0.8.

Cooling tower and water quality. New fill, even water distribution and condenser cleaning can lower the condensing temperature by 1–2 K and improve chiller efficiency by 3–6%. The investment is small and the effect is quick, yet it is often omitted.

Measure Investment (CNY) Annual saving (kWh) Annual saving at 0.95 CNY/kWh Static payback
High-efficiency chillers 640,000 106,000 101,000 6.4 years
Pump VSDs 80,000 26,000 25,000 3.2 years
Group control and reset 60,000 20,000 19,000 3.2 years
Cooling tower and cleaning 30,000 12,000 11,000 2.7 years
Total with overlap 810,000 131,000 124,000 6.5 years

4. Worked example: an 800 kW plant

A commercial complex has two 400 kW screw chillers, eight years old. Pre-retrofit measurements: 3,200 cooling hours, 58% average load factor, plant efficiency 4.2 including pumps.

Baseline electricity = 800 × 0.58 × 3,200 ÷ 4.2 = 353,219 kWh/year.

Replace with Tier 1 chillers at a plant efficiency of 6.0: 800 × 0.58 × 3,200 ÷ 6.0 = 247,253 kWh/year. Chiller saving = 105,966 kWh.

Add pump VSDs, group control and cooling-tower improvements. After an overlap factor of 0.8, total saving is 130,966 kWh. At 0.95 CNY/kWh, annual saving is 124,418 CNY.

Investment: chillers 800 kW × 700 CNY/kW = 560,000; installation and piping 80,000; pump VSDs 80,000; group control 60,000; cooling tower and cleaning 30,000; total 810,000 CNY.

Static payback = 810,000 ÷ 124,418 = 6.5 years. Chiller-only payback = 640,000 ÷ 101,000 = 6.4 years. Dynamic payback at a 6% discount rate is about 8.5 years.

The counterintuitive result is that the chiller-only option does not have a shorter payback. Chillers are 79% of the investment but the pumps, group control and tower measures are 21% of investment and contribute about 35% of the savings. The combined package costs 27% more and saves 24% more, so payback is similar, while operating stability is better.

5. Sensitivity: three variables decide the answer

Variable Value Annual saving Static payback
Tariff 0.75 CNY/kWh 98,000 8.3 years
Tariff 0.95 CNY/kWh 124,000 6.5 years
Tariff 1.15 CNY/kWh 151,000 5.4 years
Running hours 2,400 h 93,000 8.7 years
Running hours 3,200 h 124,000 6.5 years
Running hours 4,000 h 156,000 5.2 years
Load factor 45% 97,000 8.4 years
Load factor 58% 124,000 6.5 years

Tariff and running hours affect the result linearly. Load factor does not: at low load, the new chiller’s part-load advantage is larger, but the absolute cooling delivered is smaller, so the saving falls. “Lower load means it always pays to replace” is true only when the efficiency gap is large enough.

6. Where the payback can be misleading

Load growth. If cooling demand rises 20% after the retrofit, savings calculated at the new load are higher while the baseline stays at the old load, so payback is overstated. State “based on current load” in the proposal.

Remaining life. If the old chiller has less than three years left, the replacement should be justified as an early replacement, not as an energy project. Compare it with the cost of continuing to repair.

Subsidies and tax. Grants, green-factory scoring and accelerated depreciation can change the actual payback, but they should be shown separately rather than buried in the base case.

Demand charges. Under a two-part industrial tariff, lower maximum demand reduces the capacity charge, often 20,000–50,000 CNY/year for an 800 kW plant. This is frequently omitted.

Water side and terminals. If the retrofit stops at the chiller and the terminals cannot accept a higher supply temperature, the reset savings will not materialize. The scope should include a terminal-capacity review.

Common questions

What payback is reasonable?

Five to seven years for commercial projects and four to six years for industrial projects is generally acceptable. Above eight years, implement low-cost measures first — cleaning, VSDs and group control — and defer chiller replacement until the end of life.

Can savings be calculated from nameplate power?

No. Nameplate is a rated condition, while actual load is often 50–60% of rating, and the difference between old and new chillers is mainly part-load performance. Use measured load and IPLV.

Is chiller-only replacement always a longer payback?

Not necessarily. Chillers dominate investment, so the chiller-only option has a smaller investment and a smaller saving. Its payback is often similar to the combined package, but the combined package gives better stability and more savings.

Why present both static and dynamic payback?

Static is easy to compare across options; dynamic reflects the cost of capital. A loan-financed project or one with an internal return target should use dynamic. A cash-funded project with rising tariffs may use static as a screening figure.

What if the savings do not materialize?

Check three items first: water-side flow against design, condenser-water temperature against design, and whether the group-control sequence is actually in service. These usually account for most of the shortfall.

Are there measures that do not require chiller replacement?

Yes. Condenser and evaporator cleaning, cooling-tower distribution, pump VSDs and supply-temperature reset can reduce energy by 10–18%, often for less than 100,000 CNY, with a two-to-three-year payback.

What baseline?

A measured 12-month baseline with the same boundary. Use bills and a logger. A baseline built after the retrofit cannot prove the saving.

What discount rate?

Use the owner’s cost of capital, not a marketing rate. Test 5%, 8% and 10%. A payback that only works at 3% is fragile.

Related reading

LONGCOM can build a metered baseline, compare retrofit options and calculate static and dynamic payback for a chiller plant. For a project review, contact 400-878-0280 or 13002001280.

Related reading

COP and IPLV guarantee · Chiller plant group control · Water chiller sizing