How to calculate trigeneration payback: formulas and pitfalls

Trigeneration (CCHP) payback is calculated in two steps. First, we compare the annual costs of two options: the grid, electric chillers and a boiler house versus a gas engine generator set with an absorption chiller. The difference in costs is the annual savings. Then we divide the difference in investment by it: the result is the simple payback period in years.

This page covers the formulas, the pitfalls that spoil a feasibility study, and the data that must be collected before the calculation. Ready results for six facilities are on the trigeneration feasibility study page, and you can calculate your own facility in the calculator.

How are the savings calculated, in plain words?

The company's model is built from six steps.

Step	What we calculate	How
1	Cooling per year	peak cooling times full-load hours
2	Baseline costs	electricity for the facility and the chiller (annual cooling divided by COP) at the tariff, plus water and service
3	Heat for the absorption chiller	annual cooling divided by the machine's COP, about 0.75
4	Trigeneration costs	gas engine fuel (output divided by 40% efficiency and 9.3 kWh/m³) plus service, water, absorption chiller, minus the boiler gas we no longer burn
5	Savings	baseline costs minus trigeneration costs
6	Payback	difference in investment divided by the savings

If the engine heat is not enough for all the cooling, the shortfall is covered by an electric chiller powered by the same gas engine, there is more heat, and the calculation is repeated until it converges. The share of cooling from waste heat in the six models came out at 80 to 100%.

The model assumptions are listed below, in the section "What assumptions does the model use?".

Where the heat for cooling comes from

From one megawatt of gas, a gas engine gives about 400 kW of electricity and about 450 kW of heat, and about 150 kW is lost. That is the 40% electricity, 40–45% heat and 80–85% fuel utilization. A hot water absorption chiller with a COP of about 0.8 turns 450 kW of heat into roughly 360 kW of cooling. The model uses a more conservative COP of 0.75.

From 1,000 kW of gas, a gas engine produces about 400 kW of electricity and about 450 kW of heat; about 150 kW is lost. An absorption chiller with a COP of 0.8 turns 450 kW of heat into about 360 kW of cooling.

How much heat goes to the cooling tower

An absorption chiller rejects more heat to the cooling tower than an electric chiller: 2.45 kW per 1 kW of cooling for a single-effect machine, 1.72 for a double-effect one, about 1.20 for a chiller with a COP of 5. That is why the cooling tower in a scheme with an absorption chiller is almost twice as large and evaporates noticeably more water. This is included in the water and investment lines.

At 1,000 kW of cooling, an absorption chiller takes in 1,430 kW of heat and rejects 2,430 kW to the cooling tower. An electric chiller with a COP of 5 rejects about 1,200 kW at the same capacity. That is why the cooling tower in an absorption scheme is almost twice as large.

What comes from electricity and what from cooling?

The main part of the savings comes from on-site electricity. Gas costs RUB 9/m³, a cubic meter holds 9.3 kWh, an engine with 40% efficiency turns them into electricity, and the gas per kilowatt-hour costs RUB 2.42. Add RUB 1.3 for service, and your own power costs RUB 3.72 versus RUB 10 from the grid. This difference applies to the entire output.

Against this background, cooling is secondary in money terms but essential to the concept. It gives the engine heat a use all year round: in winter it heats the building, in summer it goes to the absorption chiller. And it takes the electric chiller load off the grid, which also reduces the capacity that would otherwise have to be connected. That is why the calculation for a business center gives 6.8 years: the electrical load is small, and cooling is needed only 1,200 hours a year.

Which chiller to compare against: COP 3 or 5.5?

This matters more than any tariff. If you take an old chiller with a seasonal COP of 3 as the baseline, trigeneration looks impressive. If you take a modern one with free cooling (cooling with outdoor air), with a COP of about 5.5, the advantage on cooling is noticeably smaller.

In a model of a gas-fired absorption chiller without a gas engine, 5,000 kW of cooling versus a chiller with a COP of 3, this shows in the range of savings when a parameter changes by ±20%:

Parameter	Range of savings, RUB million/year
Baseline chiller COP	42.5
Electricity tariff	35.2
Load hours	28.4
Absorption chiller COP	7.1
Gas price	6.8

Compared with a modern packaged chiller rather than an old one, the same gas-fired absorption chiller pays off in about 11 years, and about 6.8 years if the grid connection fee is taken into account. That is why in our six calculations the baseline chiller is a fair one: COP 5.5 for the data center, 3.5–4 for the rest. Trigeneration holds up even against such a chiller, because the main savings come from electricity, not from cooling. But if you are comparing only an absorption chiller with an electric chiller, without a gas engine, ask the contractor what COP is used for the baseline.

How many load hours are needed?

For an absorption chiller, payback is usually acceptable from 3,000–4,000 hours of cooling a year; below 2,000 hours it does not pay off. For trigeneration with exhaust heat recovery, ORNL names a year-round base heat load and at least 4,500 hours a year. In the company's model, the load is set as equivalent full-load hours: 6,130 for the data center, 3,456 for the dairy plant, 2,200 for the shopping mall, 1,200 for the business center.

How are projects screened in international practice?

The US DOE and EPA programs for combined heat and power have long used screening: questions that weed out facilities before any feasibility study.

What is checked	Threshold or guideline
Heat load	high and steady throughout the year
Operating hours	more than 3,000 h (ORNL, 2022); more than 5,000 h is attractive (DOE and EPA, 2015)
Electricity price	above $0.06/kWh (ORNL); at least $0.07/kWh (DOE, 2015)
Heat utilization	more than 50% of the available heat over a year: the project makes sense (DOE, 2007)

The spark spread is also calculated: the difference between the grid electricity price and the cost of on-site power including gas and maintenance. For trigeneration, absorption cooling in summer evens out loads and reduces the peak and demand charges. On the Austin project (2004, ORNL report), the estimated payback was 7 years, and the actual efficiency was 74% against the design 79%.

What pitfalls spoil a feasibility study?

Feasibility studies from different years repeat the same mistakes. The calculator is built to rule them out.

Pitfall	What it breaks
Payback as an investment ratio instead of years	the number does not answer "how many years"
Negative payback when the absorption chiller costs less to install	reads as "never" but means "immediately"
One-time benefit added to annual savings	the payback looks deceptively short
Different operating hours for electricity and gas	savings are distorted
No winter electricity for the chillers	the baseline is understated, trigeneration looks worse
Gas engine electricity not credited, but gas counted	trigeneration looks unprofitable
One machine versus three with different total capacity	the comparison is unfair
Tariffs from RUB 7.5 to 12 in different files of the same project	results cannot be compared

What affects the result most?

The order is as follows, from the strongest.

The baseline: what COP the chiller has and whether there is spare electrical capacity.
Equivalent load hours. Below 2,000 hours an absorption chiller does not pay off; above 4,000 the gain grows linearly.
The electricity tariff and the grid connection cost. The tariff applies to every kilowatt-hour; the connection is one-time but can outweigh the difference in machine prices.
The heat source: waste heat from a gas engine or purchased heat and gas.
The gas price. While gas is below RUB 11/m³, it has about five times less influence than the tariff.
Absorption chiller COP, service life, discount rate, tariff growth.

Discounted payback is 1.3–1.6 times longer than simple payback.

What payback range is honest?

A single number in a feasibility study always looks more convincing than it is. We show a range: the estimated savings at a tariff of RUB 10 and gas at RUB 9, and a conservative case at a tariff of RUB 7.5 and gas at RUB 12.

Facility	Estimated payback	Conservative payback
Data center	2.1 years	4.1 years
Dairy plant	2.2 years	4.1 years
Pharmaceutical plant	2.3 years	4.2 years
Shopping mall	2.9 years	5.2 years
Greenhouse complex	2.0 years	3.9 years
Business center	6.8 years	12.5 years

In our experience, Russian customers are prepared to consider a payback of up to 3–3.5 years, about 4 years at most, and consider seven years unattractive. If the conservative number works for you, the project is worth doing. If only the estimated number works, it is risky.

What assumptions does the model use?

The same for all six calculations and for the calculator: grid tariff RUB 10/kWh, gas RUB 9/m³, water RUB 50/m³, gas engine efficiency 40% electrical and 45% thermal, gas engine service RUB 1.3/kWh. Investment guidelines: RUB 78,000 per kW of electrical capacity, RUB 65,000 per kW of cooling (a plant with absorption chillers, cooling towers, pumps and controls) and RUB 30,000 per kW of cooling for the baseline chiller. These are the company's 2026 model guidelines; they are replaced by the price from a commercial proposal. The payback is simple, without discounting, tariff growth or standby gas engines.

What data are needed for a real calculation?

To request a selection, collect the data using the questionnaire. While even one item is empty, it is too early to send the request.

Block	What is needed	What happens if it is wrong
Chilled water	profile by hour and month, temperatures, redundancy	capacity shortfall at peak or running at 20%
Heat source	temperatures, flow, year-round demand, exhaust back pressure	capacity shortfall, the engine loses power
Gas engine	required, minimum and maximum output, grid-parallel or island mode	wrong generation capacity
Cooling tower and water	wet bulb, noise, make-up water analysis	the cooling tower cannot keep up in summer, scale
Electrical	spare capacity, connection price, tariff, gas price	lost months

Three questions define the economics: the actual tariff, the load hours and profile, and the cost of an hour of downtime. Selection of gas engines and Thermax and NCT equipment is based on this questionnaire. NCT cooling towers, chillers and dry coolers are described on the cooling towers and dry coolers pages.

If you are ready to calculate your facility, submit a request for a feasibility study or start with the calculator. Typical reasons for failure are described on the trigeneration mistakes page.

Sources

ORNL and DOE CHP TAP, CHP Project Screening, 2022: https://bptraining.ornl.gov/wp-content/uploads/2022/12/Session-2-CHP-Project-Screening-2022-12-01-FINAL.pdf
DOE Midwest CHP TAP, Combined Heat and Power Resource Guide, 2015: https://www.energy.gov/sites/default/files/2019/01/f58/CHPGuide2015.pdf
DOE and Midwest CHP Application Center, CHP Resource Guide for Hospital Applications, 2007: https://www.energy.gov/sites/prod/files/2013/11/f4/chp_hospital_guidebook_2007.pdf
ORNL, Berry et al., Advanced absorption chiller converts turbine exhaust to air conditioning, 2005: https://webarchive.library.unt.edu/web/20080920092112id_/http://www.ornl.gov/sci/engineering_science_technology/cooling_heating_power/pdf/ISHPC-095-2005.pdf

Frequently asked questions

Will trigeneration pay off if the site still has enough electrical capacity?

If the new load does not need additional electricity and the cooling demand is small, the project may not be justified. If the connected capacity will not be enough in the future, the scheme can pay off. A grid connection costs from RUB 20,000 to 100,000 per kilowatt, so look at the margin in the grid connection specifications three to five years ahead. And check the limits: at a tariff below RUB 5 or gas above RUB 11, the calculation must be rechecked separately.

Does the payback calculation include the cost of the cooling tower, pumps and piping?

In our model, yes: RUB 65,000 per kilowatt of cooling covers the machine, cooling tower, pumps and controls. What is more often missed is operation. For the machine alone, the specification gives 5–10 kW per MW; with all the auxiliaries, 46–130. Make-up water for the cooling tower of a 5 MW double-effect absorption chiller is about 12.5 m³/h; for a single-effect hot water machine, evaporation alone is about 18 m³/h, plus blowdown. Ask for these lines separately in the commercial proposal and check that the water and pump electricity are included too.