Skip to content

Chilled-water reset and pumping: low delta-T, riding the curve and the VFD minimum

Workbook exercise plant-chw-reset-pumping · PNNL re-tuning chapter 8 · about 40 minutes

Goal

Answer the chilled-water side's three re-tuning questions on a simulated plant: is the supply temperature reset, is the loop delta-T low, and is the loop differential-pressure (DP) setpoint reset or held constant? Then see what a stuck tower bypass does to the chilled water and to the pumps, and how the rules read a constant-flow plant and a pump's own VFD floor.

Learn more

Read these first (PNNL, free):

  • Central Utility Plant Cooling Control, the re-tuning guide: its three questions (a reset on the chilled-water supply temperature, a low loop delta-T, and a constant loop DP setpoint that could be reset at part load) are the three halves of this exercise, plus the pump that serves them.
  • Chapter 8: Central Utility Plant of the re-tuning training.

Datasets and licence

  • lbnl-chiller: LBNL's simulated water-side chiller plant (three chillers, three cooling towers, primary and secondary chilled-water loops), a year at one-minute resolution, as a fault-free run plus labelled faulted runs. Licence CC-BY-4.0 (open; cite it). The download is about 1.6 GB; this exercise needs only the default subset (four runs). The catalog lists the problems found in the published data and how CAMBER handles each, among them a chilled-water reset that spans a different range from the one documented: see its data issues.

Each run becomes PLANT__<run> in the facility ds-lbnl-chiller. The default subset holds PLANT__fault_free, PLANT__coolingtower_fouling_065, PLANT__bypass_stuck_075 and PLANT__chiller_bias_2. The chilled-water temperatures are chiller 1's; the pump is the secondary loop's lead pump.

Setup

In the lab

  1. Run camber lab and open the URL it prints (http://127.0.0.1:8765/lab?token=...).
  2. Tick lbnl-chiller and press Fetch & ingest (the default subset is enough).
  3. When the job is done, the row links to trends (the trend viewer). This exercise uses its own config: use the command line for step 2 of the steps below.

On the command line

camber datasets fetch lbnl-chiller
camber datasets ingest lbnl-chiller --store lab_store
camber datasets config lbnl-chiller --exercise plant-chw-reset-pumping --store lab_store --out chw.json
camber run chw.json --out chw_out

The exercise's config (--exercise plant-chw-reset-pumping) runs two rules. chw_plant_reset fits chiller 1's leaving-water temperature against the outdoor dry-bulb over the chiller's running, occupied hours, checks that the slope goes the way a reset should, and reports the loop delta-T against an 8 °F design minimum (design_deltaT_min_f). It also reads chiller 1's flow: on a constant-flow plant the delta-T is reported but not judged. chw_pump_dp_reset reads the secondary pump's speed and the loop's DP setpoint, and learns the pump's VFD floor from its speeds. The exercise keeps every parameter at its default. The plant's chiller status point is an enable, not a run status, so CAMBER decides when the chiller runs from its power.

Steps

  1. Look before you judge. In the trend viewer, open PLANT__fault_free and plot chiller 1's leaving and return water with the outdoor temperature over a spring month. Then plot the secondary pump speed and the loop DP with its setpoint over the same month.
  2. Run the exercise's config (the commands above) and read the findings.
  3. Reset. In chw_out/findings.json, read chw_plant_reset for the fault-free run: chwst_reset_present, chwst_reset_direction, chwst_slope_per_F and chwst_median_f.
  4. Delta-T. In the same finding, read deltaT_median_f, low_deltaT_pct, flow_mode, flow_cv and the caveats.
  5. Pumping. Read chw_pump_dp_reset for the fault-free run: median_speed_pct, pct_running_near_full, pct_running_near_min, vfd_floor_pct, near_min_band_pct and dp_sp_reset_present. Then find the run where the pump's median speed is lowest.
  6. The stuck bypass. Compare both findings for PLANT__bypass_stuck_075 with the fault-free run's.

Questions

  1. Is the chilled-water supply temperature reset? Which way does it move as the weather warms, and is that the right way?
  2. What delta-T does chw_plant_reset find on the healthy plant, and why is the finding ok anyway? Is a low delta-T here a fault to fix or a property of the plant?
  3. Is the DP setpoint reset? How hard does the secondary pump work, and what does the rule call it?
  4. Where is the pump's VFD minimum, how does the rule find it, and what share of hours does it count near the minimum? On which run does the pump sit at its floor most?
  5. What does the stuck tower bypass do to the chilled water and its delta-T, and to the pump? Which finding points at the cause, and which only at a symptom?

What CAMBER shows

  • Findings. camber run prints one line per finding with its severity and a summary. findings.json holds the metrics: for chw_plant_reset the supply median, its slope on the dry-bulb, whether a reset was found and which way it goes (chwst_reset_direction: expected, reverse or flat), the delta-T median and the share of running hours below 8 °F, the flow mode and the flow's coefficient of variation, and run_source (what decided the chiller was running); for chw_pump_dp_reset the median speed, the shares near full (90 % or more) and near minimum, the learned VFD floor and the near-minimum band it used (near_min_source: learned, or default for 25 % when no floor is found), and whether the DP setpoint is reset.
  • Severity. chw_plant_reset is a fault when half or more of the running hours are below the delta-T minimum, a warn from a fifth, or on a flat or reversed supply temperature; on a constant-flow plant the delta-T does not count. chw_pump_dp_reset is a fault at 60 % or more of the hours near full speed, a warn from 30 %, or at half the hours or more near the minimum.

Chilled-water supply, return and outdoor temperatures over three weeks; the supply stays flat at 44 F and the occupied running hours are shaded

Synthetic illustration, not this exercise's dataset: the chw_plant_reset evidence trend. The supply is held at 44 °F whatever the outdoor temperature, a reset that never happens. The shaded hours are the occupied running hours the rule judged with the supply at or below 46 °F.

Caveats

  • The delta-T is chiller 1's own, on a primary loop whose flow does not follow the load, so it is low at part load by design. The rule recognises this from the flow point; without one it cannot, and judges the delta-T (declare flow_mode then). The 8 °F minimum is generic: set design_deltaT_min_f from the loop's design.
  • The rule expects a reset to lower the supply in hot weather (expected_reset_sign); a plant whose sequence does the opposite on purpose needs that parameter changed.
  • The chilled-water setpoint point is in the data but is not mapped (see the data issues), so CAMBER judges the reset from the supply temperature itself.
  • The learned floor is a plateau in the speeds, not the drive's setting: a pump that rarely reaches its floor falls back to the 25 % band. Set near_min_pct from the VFD parameters when you know them.
  • The data are simulated: one plant, one climate, one control sequence.

Going further

  • In the trend viewer, find the winter months: what does the pump do when there is no cooling load at all? Is that a scheduling opportunity?
  • Run the same config on the full subset and compare the chiller- and pressure-sensor-bias runs with the fault-free one: see the sensor exercise.
  • Compare with the air side's supply-air temperature reset: a warmer chilled-water supply in mild weather only works while the air handlers still reach their discharge temperature.