The Leaking Chiller, the Budget Cut and the Phase-Down
Two 20-year-old 500-ton centrifugal chillers serve a 30,000 m² office campus. Each holds 800 lb of R-134a. The 2024 service log shows one machine lost 88 lb, and no refrigerant records exist for 2023. The CFO has cut the maintenance budget 8% and deferred a USD 1.1 million chiller replacement by two years. Refrigerant prices are up 40% since 2022 and supply is tightening under the national HFC phase-down. The chief engineer's position is that 'it still makes cold water, so it's fine.' Summer peak load is 780 tons, and the plant has tripped twice on the hottest afternoons. The annual sustainability report cites a 12% emissions reduction that excludes refrigerants.
1. What are your first 30–60 day actions on the leaking machine, and why can none of them wait for the capital decision?
Read the worked answer
1. Treat it as a compliance event, not a maintenance preference. An 88 lb loss on an 800 lb charge is an 11% annual leak rate. For comfort-cooling equipment above the applicable charge threshold, the leak-rate trigger is 10%; crossing it obliges the owner to repair the leak and verify the repair with a follow-up test, with records retained. That obligation runs on its own clock and does not pause for a capital review.
2. Reconstruct the records. The missing 2023 log is a standalone finding. Establish a refrigerant log for both machines - charge, additions, recovery, service events, technician - and put it under one named owner.
3. Find the leak, don't just top it up. Leak detection survey on the leaking machine: seals, gaskets, tube joints, purge unit. Topping up without repair is the most expensive way to run a chiller once refrigerant prices are rising 40% over two years.
4. Quantify the emissions for the report. Working from the charge and leak rate, the lost refrigerant is a material quantity of CO2-equivalent emissions that is currently invisible in the 12% claim. The exclusion is a disclosure problem, not just an accounting one - and fixing it now is cheaper than restating later.
5. Put the plant risk on record. Two summer trips at 780 tons peak against a two-machine plant means there is no redundancy at design conditions. Write it up as an operational risk with a probability and a cost, and route it to the CFO in those terms.
2. Build the repair-versus-replace case the CFO asked you to defer. What does the analysis contain, and what would change your recommendation?
Read the worked answer
Baseline the options over the same horizon (e.g. ten years, discounted):
- Option A - repair and continue. Leak repair and verification, plus rising refrigerant cost for any residual loss, plus the probability-weighted cost of an unplanned summer failure (emergency rental chiller, overtime, tenant impact, possible business interruption). Include the price trend, not today's price: refrigerant cost is the one line in this budget that is reliably going up.
- Option B - replace now. USD 1.1 million, but sized against the actual 780-ton peak, not the original connected load. Right-sizing alone can reduce capital and part-load energy. Include lower refrigerant charge and lower-GWP refrigerant, part-load efficiency, and the maintenance differential on 20-year-old machines.
- Option C - repair, then replace in two years. This is what the CFO has effectively chosen. Cost it honestly: two more years of refrigerant, two more years of failure risk, and the likelihood that the two-year deferral becomes a four-year deferral once the capital line moves.
What would change the recommendation: a condition assessment showing the machine is sound and the leak is a single accessible seal; a load profile showing the plant is grossly oversized and a smaller replacement changes the economics; or a confirmed low-GWP retrofit path that extends life at low cost.
The point to make to the CFO: the deferred maintenance cut and the capital deferral are the same decision seen twice. The 8% maintenance cut reduces the leak-detection and service frequency that keeps the refrigerant in the machine, so the deferral raises the very cost it is meant to avoid.
3. How would you embed this into routine operations so it does not resurface as a crisis next summer?
Read the worked answer
- A refrigerant management procedure with a named owner: who logs charge additions, who signs off service, what triggers a leak survey, and where records live. Records must survive staff turnover.
- Leak-rate tracking as a live metric, calculated at every service event rather than annually, with an escalation trigger well below the regulatory threshold - an 11% annual rate should never be discovered a year late.
- Maintenance scope tied to the equipment, not the budget line. When an 8% cut is applied across the board, protect the tasks that carry compliance and refrigerant containment (leak surveys, purge maintenance, water treatment) and take the cut from discretionary work. Present that trade-off explicitly rather than absorbing it silently.
- A plant replacement roadmap with condition assessments, refrigerant transition options and a trigger-based decision rule (e.g. replace when annual refrigerant plus failure-risk cost exceeds a stated share of replacement cost), so the decision is not re-litigated from scratch each budget cycle.
- Redundancy and peak-load review: document the 780-ton peak and the two-machine exposure, and check whether load reduction (retro-commissioning, schedule repair, setpoint discipline) can restore margin without capital.
- Reporting integrity: include refrigerants in the emissions inventory and state the boundary. A 12% reduction that excludes the largest single source of high-GWP emissions will not survive scrutiny, and correcting it proactively is far cheaper than defending it.