Chiller staging and condenser-water reset

Workbook exercise plant-chiller-staging · PNNL re-tuning chapter 8 · about 40 minutes

Goal

Find a chiller that short-cycles, a chiller left running through the night, and a cooling tower held at a fixed setpoint instead of reset with the weather. No open dataset in the catalog shows these faults. Then look at a reset that passes the check but saves little, and decide how strict the check should be.

Learn more

Read these first (PNNL, free):

Central Utility Plant Cooling Control, the re-tuning guide: its questions on chiller staging and on resetting the condenser water with the outdoor wet-bulb are the two halves of this exercise.
Chapter 8: Central Utility Plant of the re-tuning training.

Datasets and licence

synthetic-chw-plant-staging: a synthetic chilled-water plant, one chiller and its cooling tower, that CAMBER generates on your computer. Nothing is downloaded and no real building is behind it. It holds the outdoor dry-bulb and wet-bulb, the chiller's power, the chilled-water supply, return and flow, the condenser water leaving and returning to the tower, and the tower fan speed, every 15 minutes for three summer weeks with a heat wave in the third. Licence Apache-2.0 (CAMBER's own; the data may be shared). About 0.7 MB. See synthetic datasets for how CAMBER generates it.

The same weather and the same building drive five plants, each ingested as PLANT__<plant> in the facility ds-synthetic-chw-plant-staging: PLANT__well_staged, PLANT__short_cycling, PLANT__left_on, PLANT__fixed_cw and PLANT__high_floor. Which is which is the exercise; the instructor key describes how each was built.

Setup

In the lab

Run camber lab and open the URL it prints (http://127.0.0.1:8765/lab?token=...).
Tick synthetic-chw-plant-staging (its kind reads synthetic) and press Fetch & ingest. CAMBER generates the file instead of downloading it, then ingests it like any other dataset.
When the job is done, the row links to trends (the trend viewer) and to report, the audit report of the dataset's template at the trend's own 15-minute resolution. The threshold lesson in step 5 below uses the command line.

On the command line

camber datasets fetch synthetic-chw-plant-staging
camber datasets ingest synthetic-chw-plant-staging --store lab_store
camber datasets config synthetic-chw-plant-staging --exercise plant-chiller-staging --store lab_store --out stg.json
camber run stg.json --out stg_out
camber datasets score synthetic-chw-plant-staging --store lab_store --findings stg_out/findings.json

The exercise's config (--exercise plant-chiller-staging) runs two rules at their defaults on the 15-minute trend: chiller_staging (starts per day, and how much of its running time the chiller spends at a small part load) and condenser_water_reset (whether the tower's leaving water follows the wet-bulb). A variant asks condenser_water_reset for a steeper reset:

camber datasets config synthetic-chw-plant-staging --exercise plant-chiller-staging--slope05 --store lab_store --out stg05.json
camber run stg05.json --out stg05_out

Steps

Look before you judge. In the trend viewer, open each of the five plants and plot the chiller power over a week, then the condenser water supply with the wet-bulb over the three weeks. Find the heat wave.
Run the exercise's config (the commands above). For each plant, read chiller_staging in stg_out/findings.json: the severity, starts_per_day, runtime_pct, load_factor_median_pct, low_load_pct, load_spread_pct and any caveat.
The condenser water. Read condenser_water_reset for each plant: cws_slope_per_wetbulb, cws_median_f, cws_above_wb_plus_approach_median_f and reset_present.
Score. Run camber datasets score and read the rates for the two rules.
Ask for more. Run the slope05 variant and note which verdicts change.

Questions

Which plants fail chiller_staging? Do they fail it for the same reason?
Which plant fails condenser_water_reset, and what does its slope say?
One plant passes condenser_water_reset with a slope just above the line. What is its tower doing, and what does its median condenser water say that the verdict does not?
What does the label score say for the two rules?
With a required slope of 0.5, which verdicts change? Is that the right line for this plant, and how would you choose it on a real site?

What CAMBER shows

Findings. camber run prints one line per finding: one per rule and plant.
Staging. chiller_staging reports starts_per_day (the chiller's starts, from its power, divided by the days in the trend), runtime_pct, load_factor_median_pct (the median load while running, as a share of the chiller's own observed peak) and low_load_pct (the share of running samples below 40 % of that peak). It warns at 6 starts a day or 25 % of running time at low load, and faults at 12 starts or 50 %. load_spread_pct is the spread of the running load (its 10th to 90th percentile) as a share of that peak. When it is 5 % or less, every run is at about the same output, and the finding carries a caveat: the load factor is then measured against that one output, so it says nothing about part load.
Condenser water. condenser_water_reset regresses the tower's leaving water on the measured wet-bulb over the samples the tower fan runs and reports the slope (cws_slope_per_wetbulb), the median (cws_median_f) and reset_present. A slope under reset_slope_flat (0.3 °F per °F by default) is a flat setpoint: a warn, an opportunity rather than a broken part. cws_above_wb_plus_approach_median_f is for information only: the median of how far the condenser water sits above the wet-bulb plus the tower's design approach (design_approach_f, 7 °F by default, as in cooling_tower_approach). Near 0 the tower makes the coldest water its approach allows; well above it, a fixed setpoint or a reset's floor holds the water warmer. It never changes the verdict.
Score. camber datasets score prints the true- and false-positive rates for the entry's two declared detectors.

Condenser water supply against the outdoor wet-bulb: most points sit flat at 85 °F, above a shaded band that follows the wet-bulb plus 7 °F

From this exercise's synthetic plant PLANT__fixed_cw: the condenser water leaving the tower against the wet-bulb, with the band a reset to the wet-bulb plus 5 to 9 °F would keep it in (between 65 and 85 °F). Red points are outside the band.

Caveats

The data are synthetic: CAMBER's own model of one plant, built to show these faults. They show what the rules see, not how real plants behave; never cite them as evidence about a building.
A 15-minute trend still hides a run shorter than 15 minutes, so a start count is a floor.
chiller_staging judges one chiller. Whether a plant runs more chillers than its load needs is a question for chiller_staging_fleet, which this single-chiller plant cannot answer.
The lowest condenser water a chiller accepts is the chiller maker's figure. A reset floor above it wastes energy; one below it can trip the chiller.

Going further

Read the cooling tower exercise: its tower holds its leaving water at the wet-bulb plus 8 °F. Run condenser_water_reset there and compare.
Run camber rules params chiller_staging and camber rules params condenser_water_reset to see where each default comes from and how to calibrate it.