The Challenge
A steel fabrication plant with seven compressors had a reliability problem: two were failing repeatedly — bearings, seals, and valve assemblies on a 14-month replacement cycle. The maintenance team assumed the two failing machines were defective; three sets of replacements later, the failures continued.
The failures were not defects. They were the result of unmanaged load distribution. The two compressors in question were the first in the start sequence — the plant's informal practice was to run them first and only add others if pressure dropped. In a plant with seven compressors and highly variable demand, this meant the two lead units were running loaded for 16–18 hours per day while the others ran 3–4. Uneven loading accelerates wear; the 14-month failure cycle was predictable once load data was visible.
What Changed
Per-compressor runtime monitoring, load/unload cycle tracking, and actual delivery volume per unit. A sequencing system built on actual load profiles rather than informal start order.
The load data made the imbalance unmistakable: the two lead compressors had accumulated 6× the runtime of the lowest-utilised unit. Sequencing logic was restructured to rotate units weekly, with runtime hour targets equalised across the fleet. The first failure-free six-month period in three years followed immediately.
Results
parts, labour, downtime
across all 7 units
“When you can't see how load is distributed across your compressor fleet, wear is hidden until it becomes failure. Runtime equalisation through data-driven sequencing isn't a complex intervention — it's a scheduling change — but it requires the visibility that only monitoring can provide.”