The Challenge
A plant with six compressors saw electricity costs rise 18% over two years with no corresponding increase in production. The energy team had monthly utility bills and nothing else. No one had mapped the actual compressed air demand throughout the day — what was being used, when, and in what quantities.
What Became Visible
Continuous monitoring of compressor load revealed that four of the six units were spending 65% of their runtime in the unloaded state — consuming electricity without producing air. During shift changes, all six compressors were running at full capacity even though pneumatic tools had been shut down 15 minutes earlier. The plant was designed for peak demand (a 12-minute burst during shift startup) but the compressors never throttled back during the remaining 22+ hours of the shift.
What Changed
Real-time load monitoring across all compressors showed, hour by hour, what air was actually needed versus what was being delivered. Load profiles became predictable, not because the production process changed, but because for the first time, the operations team could see what was happening.
How it worked: The fix was a sequencing change: two compressors designated for baseload, one for trim, three held in reserve. Sequencing logic was built from actual demand data, not guesswork or legacy practice. Start/stop times were synchronized with production windows, not operational convenience. Unloaded runtime dropped from 65% to 11% within the first month.
Results
no production change
from sequencing alone
Compressor oversizing is structural waste. When six compressors exist, the tendency is to use all six. When visibility exists, the question changes from 'do we need these compressors' to 'do we need these compressors running right now.' Fixed-speed compressors in the unloaded state don't produce air — they consume power and add heat. Making loading patterns visible is the first step to making them efficient.
Operational Reality
The plant manager initially resisted reducing compressor count, assuming it would compromise air availability. Real-time monitoring proved the opposite: with intelligent sequencing based on actual demand, reliability improved because fewer units meant less wear distribution.