The HVAC Retrofit That Competes With Itself
You manage a 240,000 sq ft multi-tenant office built in 1988. The chiller plant is two 600-ton constant-speed centrifugal machines, one of which failed last summer; the surviving unit carried the building at reduced capacity through a 96°F week. Engineering proposes replacing both with three 400-ton variable-speed magnetic-bearing chillers at $2.1M, projected to cut HVAC energy 28%. The building's blended electric rate is $0.11/kWh, and metered plant consumption last year was 3.9 million kWh. The asset manager has a five-year hold and a 9% discount rate. Two anchor tenants (38% of rent) have lease expirations in 26 months with no renewal options. The capital committee wants a recommendation in two weeks and has asked specifically whether to size for N+1 redundancy or for peak efficiency.
1. Quantify the simple payback of the proposed retrofit and assess whether the 28% savings figure is a sound basis for the capital decision.
Read the worked answer
Annual plant energy is 3.9M kWh × $0.11 = $429,000. A 28% reduction is $120,120/year, giving a simple payback of $2.1M ÷ $120,120 ≈ 17.5 years - roughly three and a half times the hold period. That alone fails most capital hurdles, but the deeper problem is the savings baseline.
The 28% figure was almost certainly modeled against a two-chiller plant running normally. Last year the plant ran on one machine at reduced capacity, so the "before" case was artificially cheap - a single chiller running flat-out at partial load is inefficient, but the building also shed load. The model may therefore overstate savings in a normal year and understate the cost of the failure mode.
The right move is to separate the two questions the committee conflated: reliability capital (avoiding another single-point failure) and efficiency capital (kWh reduction). Reliability spending should be justified by expected loss - tenant disruption, emergency rental chillers, lease leverage - not by a 17-year energy payback. Ask for a load-duration curve and an IPMVP-aligned baseline before any number goes to committee.
2. The committee asks whether to size for N+1 redundancy or peak efficiency. Recommend an approach, accounting for the lease rollover risk.
Read the worked answer
These are not mutually exclusive if the plant is modular. Three 400-ton variable-speed machines totaling 1,200 tons, with any two able to carry the design load, is effectively N+1 without a dedicated standby machine - the failure of one unit leaves 800 tons, which likely covers the building except on design-day afternoons. That is the configuration to argue for.
But the lease risk changes the sizing. With 38% of rent expiring in 26 months and no renewal options, you may be sizing for tenants who leave. If the anchors vacate, load drops sharply and three chillers become three badly-part-loaded machines - variable-speed helps, but a 400-ton screw at 10% load is still poor. Conversely, if you replace with two larger machines to chase peak efficiency, you have rebuilt the single-point failure that caused last summer's crisis and handed departing tenants a reason to demand concessions.
Recommendation: modular N+1 (three units), sequenced by a plant controller, with the capital case built on reliability and lease retention rather than energy payback. Pair the request with a tenant-retention strategy - early renewal discussions before the capital decision - so the load profile is known rather than assumed.