← Back to Portfolio

05 — Low Delta-T Syndrome

Deep dive into low-delta-T syndrome using quantile-based thresholds and efficiency classification.

chiller-plant low-delta-t fault-detection efficiency
Pythonpandasnumpymatplotlib

Key Findings

  • 17.3% of operating hours classified as inefficient (plant kW/ton > 4.0)
  • Inefficient hours cluster at low wet-bulb (median 11.9°C) and low load (median 96 tons)
  • Power floor estimated at ~142 kW with incremental cost of 2.87 kW/ton

Systematic investigation of low-delta-T syndrome: configurable threshold analysis, quantile-based fault flagging, efficiency classification, power floor detection, and identification of the worst-performing operating hours.

Notebook