Chiller Plant Sizing: Sum of Zone Peaks vs. Coincident Block

The coincident block, 410 tons, is the correct basis. Summing zone peaks assumes every zone peaks in the same hour, which is impossible when east and west exposures peak roughly six hours apart. The block is the maximum of the hourly sum, which is what the plant actually sees.

Verification before accepting 410 tons:

• Confirm the block is the maximum hourly sum, not an average of zone peaks or a single "typical" hour from the report.
• Confirm the outdoor-air treatment. Because OA is handled at each AHU, the block hour already includes each zone's OA share once. The double-counting error appears only if someone adds a separate central OA load on top of the block.
• Confirm the block hour is plausible against the exposure mix: with equal east/west glass and a 8-to-5 schedule, the block should occur mid-afternoon, when the inter-zone diversity is not yet extreme.
• Confirm the design basis. The report was run at 1% and is being applied at 0.4%; the load will be roughly 3-5% higher on that basis alone, so the number to compare is closer to 425-430 tons.
• Confirm internal and pickup loads (people density, plug loads, morning pulldown) are in the block, not added afterward by hand.

Diversity = 410/520 = 0.79. That is a large but not implausible reduction for a building with equal east and west glass; it becomes implausible if the schedule or the envelope is not actually diverse.

Risk of 410 tons (or 425 at 0.4%)

• If the diversity is a modeling artifact - for example, the west zones were given a later occupancy start than the east, or one AHU was left out of the block - the plant is short at the design hour, the chillers run at 100%, supply temperature floats up, VAV boxes hit max, and no capacity is left to recover from a hot weekend or a high-occupancy event.
• There is no standby machine in that proposal, so any chiller outage during a design week is unprotected.

Risk of 520 tons

• 110 tons of excess capacity: roughly $100k-150k of avoidable capital, plus a plant that spends most of its life below 40% load, where centrifugal machines surge and screw machines short-cycle. Part-load (IPLV/NPLV) performance and stability suffer, and the standby machine is even further from a useful operating point.

Recommendation

1. Re-run the block at 0.4% design; carry that figure (approximately 425-430 tons) as the plant load.
2. Select two chillers of about 225 tons each, or a 430-450 ton machine plus a small pony chiller sized near 25-30% of design so the plant can trim without short-cycling.
3. Keep the standby requirement. A single 410-ton machine with no standby is a reliability decision, not a capacity decision, and should be priced and justified separately.
4. Specify VFD centrifugal or screw machines and require the submitted NPLV at the actual part-load profile, not just the full-load rating.

Commissioning confirmations

• Require the block calculation and the hourly load profiles as a deliverable, so the diversity assumption is documented and auditable.
• Perform a functional performance test at a high-load day (or a load bank if a high day is not available) with all AHUs enabled and the OA dampers at design minimum.
• Verify that each zone reaches setpoint with its boxes at or below about 90% open; a box pinned at 100% is the earliest sign that the block was underestimated.
• Trend plant kW/ton at 25%, 50%, 75%, and 100% load and compare against the selection curve; a mismatch here is usually an oversized or mis-sequenced machine, not a load calculation error.