Chiller Retrofit Savings: Measure the Plant, Not Just the Compressor

Most chiller retrofit proposals arrive with one number — a lower kW per ton of refrigeration (kW/TR) for a new compressor drive, chiller or controls package — and a payback calculated from it. Whether that saving reaches the electricity bill depends on what the number leaves out: the equipment it was measured across, the hours the plant really spends at part load, what the pumps and cooling towers are doing, and whether anyone measured the plant before and after.

This guide covers what to measure before investing, how to compare options on one basis, and how to verify the result. It quotes no saving percentage, because there is no honest one without your plant's own baseline. To run the arithmetic on your figures, use the chiller retrofit savings calculator.

Draw the energy boundary before comparing any numbers

"Efficiency" means different things at different boundaries, and a retrofit that improves one can worsen another — a chiller selected for colder leaving water, say, or pumps pushed harder to deliver a new flow.

• Compressor or chiller boundary — electrical input to the chiller divided by the cooling it delivers. This is what a datasheet or a compressor-drive proposal usually quotes.
• Chilled-water plant boundary — chillers plus primary and secondary chilled-water pumps, condenser-water pumps and cooling-tower fans. This is the boundary your electricity bill sees.
• System boundary — the plant plus the air side (AHU and fan-coil fans), which matters when a retrofit changes supply-water temperature or flow.

Keep the units distinct: cooling output in TR or kilowatts of refrigeration (kWr; 1 TR = 3.517 kWr), electrical input in kW, energy in kWh. The coefficient of performance is the same ratio the other way up: COP = 3.517 ÷ kW/TR. Setting a plant kW/TR against a compressor COP, or a rated figure against a measured one, compares nothing.

Plant versus compressor energy boundary checklist

Build the part-load profile, because that is where the money is

A chiller plant rarely runs at design load. Constant-speed chillers do not draw full power whenever they run — they unload through inlet guide vanes, slide valves or compressor staging — but they generally do so less efficiently at low load and low lift than a machine designed for variable speed. So the hours spent at each load and condenser-water temperature decide what a compressor drive, a new chiller or a sequencing change can save.

• Start with records. For desktop data collection, BEE's retrofit guidance for existing commercial buildings lists the chiller design specifications, the chiller daily log book, plant operating hours, at least a year of electricity bills and LT log-book data, and any energy-management-system data.
• Log across seasons. Cover at least a hot, dry period and a humid monsoon period: cooling towers work against the ambient wet-bulb, so the monsoon is when condenser water runs warmest.
• Make a bin table. Hours in each band of cooling load (for example 0–25%, 25–50%, 50–75% and 75–100% of capacity) against bands of condenser-water temperature. Each option's plant kW in each bin, multiplied by the hours, gives its annual kWh.

Process-cooling plants need the same exercise against the production schedule: one shift and three shifts with a weekend shutdown are different profiles at the same installed TR.

Pumps and cooling towers: the part a compressor quote leaves out

Older plants often run chilled-water and condenser-water pumps at constant speed, throttle flow with valves and switch tower fans on and off. Drives on pumps and fans can cut their energy where flow can genuinely vary: centrifugal pump and fan power falls steeply with speed — the "cube law" in the Australian Government's equipment guide to motors and variable-speed drives, under which a 20% speed reduction can roughly halve the power required. In a chiller plant the limits are practical:

• Chillers have minimum evaporator and condenser flows set by the manufacturer.
• Condenser pumps lift water to the tower inlet, so part of their head is static and does not fall with speed.
• Colder condenser water helps the compressor but costs tower-fan energy; the optimum is a plant-level control decision.
• Low chilled-water ΔT — return water barely warmer than supply — forces extra pumping and extra chillers on line. It usually starts on the air side, in coils, valves or controls, and no plant-room retrofit fixes it.

BEE's Energy Conservation and Sustainable Building Code 2024 (ECSBC 2024) sets requirements for pumps and cooling-tower fans as well as chillers: the code, like the bill, treats the plant as the unit of efficiency.

Which retrofit fits which plant

Use the measured data to shortlist options; the last column is the engineering check each needs before money is committed.

Any option that adds drives adds non-linear load: check the effect on the transformer, on a DG set that feeds the plant, and on power-factor correction before ordering.

Turn the measured baseline into a payback

BEE's retrofit guidance names simple payback, net present value (NPV) and internal rate of return (IRR) as the usual tests. Each is only as good as the energy figure underneath:

1. kWh saved per year = the sum over the load bins of hours × (baseline plant kW − proposed plant kW), at the same boundary.
2. Rupees saved = kWh saved × the marginal energy rate, using the time-of-day rate for the hours when the saving occurs and the unit your tariff bills (kWh, or kVAh where it bills kVAh). Count a demand-charge saving only if the plant sets your monthly maximum demand and the retrofit lowers it in that interval.
3. Payback = total cost (equipment, electrical, civil, rigging, controls and any temporary cooling) ÷ rupees saved per year; then NPV or IRR over the equipment's remaining life.

Rerun it with hours and tariff a little worse than expected; if the payback only works at the optimistic end, the case is not made. The chiller retrofit savings calculator follows this method and shows the sensitivity; for plant size, the chiller capacity calculator gives a first estimate of TR.

Measurement and verification after the retrofit

A saving is real only if it can be shown after the work, adjusted for changes in weather and production. The International Performance Measurement and Verification Protocol (IPMVP), published by the Efficiency Valuation Organization, offers four options:

• Option A — retrofit isolation with key-parameter measurement (for example chiller kW across logged loads), estimating the rest. Cheaper, less certain.
• Option B — retrofit isolation with all parameters measured: electrical input and cooling output before and after. The natural fit for a plant that can be metered as a block.
• Option C — whole facility, from the utility meter; workable only when the saving is large relative to total consumption.
• Option D — calibrated simulation, where no baseline can be measured.

Agree the M&V plan before the purchase order: boundary, meters and calibration, baseline period, adjustment for wet-bulb and cooling load, and reporting period. A supplier's performance guarantee should be written against that plan; without an agreed baseline it cannot be tested.

Codes and approvals that can apply to a replacement

• ECSBC 2024 (BEE) covers commercial and office buildings with a connected load of 100 kW or more or a contract demand of 120 kVA or more, where the state has notified it. Under its section on additions and alterations, existing systems serving the space need not comply, but new equipment must meet the requirements for that equipment; for chillers the code refers to BEE star levels and full- and part-load rating under IS 16590. It does not cover equipment using energy for manufacturing processes unless it says so. Check your state's notification and the building's applicability.
• Electrical work. New drives, starters, cables or panel changes are alterations. Under Regulation 45(6) of the CEA (Measures relating to Safety and Electric Supply) Regulations, 2023, they may not be connected until approved by the Electrical Inspector or self-certified, depending on the voltage your state has notified.

Secured Engineers' HVAC and mechanical team designs and builds chiller plants; to have a proposal or your own data reviewed, request a scope review.

Set the measurement baseline before an efficiency project starts

Before work removes the old plant or changes its controls, assign custody of the baseline evidence. Name who exports the original meter and trend files, retains calibration records, documents missing intervals and confirms the operating calendar. Preserve raw records alongside any cleaned dataset so a reviewer can trace a reported figure back to its source. Agree access for the owner after the supplier has left.

Keep a dated change register during implementation and the comparison period. Production shifts, occupancy, setpoint changes or an unrelated pump replacement may affect the result. State who proposes an adjustment, which evidence supports it and who reviews it; leave unquantifiable effects and uncertainty visible. Agree how disputed or unavailable data will be handled before acceptance. This turns the measurement plan described above into a practical evidence handover rather than a spreadsheet only one vendor can explain.

What to confirm before the next decision

• Name the custodian of raw meter, trend and calibration records.
• Log simultaneous changes with dates and supporting evidence.
• Agree data access, adjustment review and treatment of missing information.

Ask for this review output: A baseline evidence agreement naming raw-data custodians, meter access, adjustment records, missing-data treatment and the review process for simultaneous operational changes.

Discuss the consultancy scope

Share the planned retrofit and available meter records. Request a baseline and verification plan before installation begins. Request a scoped MEP review. Disciplines, inputs and deliverables are agreed before an appointment; installation and authority decisions are separate.

Background guidance: US DOE FEMP: baseline evidence and measurement responsibilities. These inform the review principles; local requirements, project applicability and any necessary approvals must be confirmed by the responsible specialists.

Sources and how to verify

• Bureau of Energy Efficiency, Energy Efficiency Solutions for Existing Buildings — Commercial (four climates), prepared by TERI for BEE, February 2025 (data-collection parameters; payback, NPV and IRR), and Energy Conservation and Sustainable Building Code 2024 (sections 2, 3.2.5, 6.3.2–6.3.4). Both on BEE's publications page.
• Australian Government, Motors and variable speed drives (energy.gov.au equipment guide).
• Efficiency Valuation Organization, International Performance Measurement and Verification Protocol (IPMVP).
• Central Electricity Authority, CEA (Measures relating to Safety and Electric Supply) Regulations, 2023, notified 8 June 2023: Regulation 45.

Confirm locally: whether your state has notified ECSBC 2024 and whether your building is in scope, your current tariff order's energy and time-of-day rates, and the chiller manufacturer's position on converting an existing machine.