Chilled-Water Plant Staging After a Retrofit Adds a Small High-Efficiency Chiller
A 1,200-ton central plant has three equal 400-ton constant-speed centrifugal chillers (CH-1, CH-2, CH-3) and a new 150-ton variable-speed screw chiller (CH-4) added in parallel. All machines share a common primary loop with a single VFD secondary pump and a differential-pressure sensor at the far end of the distribution. Design flow is 2,400 gpm. The BAS currently stages CH-1/CH-2/CH-3 by a fixed sequence and never calls CH-4 until 90% plant load. Post-retrofit measurements over a mild week show: plant load between 120 and 200 tons for 40% of occupied hours, 200-600 tons for 45%, and above 600 tons for 15%. Meters show the new chiller runs at 0.36 kW/ton at 60% load and 0.30 kW/ton at 80% load; the old chillers run 0.62 kW/ton at 60% load. Operators complain of chiller short-cycling and low delta-T (8°F vs. 12°F design) at part load.
1. Identify the two dominant energy and reliability defects in the current staging strategy and explain the mechanism behind each.
Read the worked answer
Defect 1 - the best machine is locked out of the load range where it dominates. CH-4 is not enabled until 90% plant load, yet the plant spends 40% of hours at 120-200 tons, precisely where a 150-ton variable-speed screw at 0.30-0.36 kW/ton would carry most or all of the load. The staging logic uses a plant-load threshold tied to the old chillers' capacity rather than to marginal kW/ton. The fix is marginal-cost staging: enable CH-4 first and add or drop machines based on the incremental kW/ton of the next machine versus the current operating point, not a fixed percentage.
Defect 2 - low delta-T forces excess flow and short-cycling. At 8°F instead of 12°F, delivering the same tons requires ~50% more secondary flow, so the VFD runs faster and the DP setpoint is met by flow rather than by temperature. Low delta-T also means each chiller reaches its minimum-load point (or its evaporator flow limit) sooner, causing the short-cycling operators report. Root causes are usually coil/valve issues at the terminals (3-way valves left open, fouled coils, or a DP sensor located where it over-satisfies the loop), not the chillers. Diagnose the terminals before re-tuning the plant, because a staging fix alone will not restore delta-T.
2. Propose a revised staging sequence and the control variables you would use to trigger stage changes. Justify each variable.
Read the worked answer
Proposed sequence at low-to-moderate load: CH-4 alone up to its efficient range, then CH-4 + one 400-ton machine, then CH-4 + two, then all four only at peak. Never run the 400-ton machines alone in the 120-200 ton band.
Trigger variables and rationale:
- Marginal kW/ton of the next-on machine vs. current plant kW/ton - the physically correct staging criterion; it captures the real cost of adding capacity rather than a nameplate percentage.
- Per-chiller percent of rated amps (or load) with a deadband and a time delay - prevents hunting; require the lead machine to exceed ~85-90% for 10-15 minutes before adding, and to fall below ~40-50% for 20-30 minutes before dropping.
- Evaporator flow and minimum-load limits - protects the variable-speed screw from operating below its minimum stable load; if CH-4 cannot unload far enough, that argues for a small thermal storage buffer or a minimum-flow bypass rather than cycling.
- Chilled-water supply temperature error and secondary DP - confirms the plant is actually short of capacity and not just short of flow.
Add a soft-start/soft-stop ramp and a maximum starts-per-hour interlock on CH-4 to stop the short-cycling while the terminal delta-T problem is being corrected.