38% increase in throughput

Inside an industrial spray drying tower.

Background

High-volume FMCG production runs on fine margins. Yield, quality and energy use are governed by a set of interacting variables that must be held in balance across long, continuous runs - and those conditions rarely stay still. Small drifts compound into lost product and wasted energy, often without ever announcing themselves.

Challenge

This process combined significant energy demand with operating conditions that changed faster than the controls could respond. Left uncorrected, that variability worked its way through the line as inconsistent product quality, hidden inefficiency and material waste - with the cost absorbed quietly into the cost of production.

The installed control systems could not adapt quickly enough to changing boundary conditions. As the process drifted, product required rework or was lost. Operators did not have the real-time visibility needed to identify the source of variation, yet optimisation had to happen continuously and at line speed.

‍

Solution

PurpleSector deployed a race-derived vision and vibration system, to monitor the line at the critical points where the failure was likely to occur.

Recordings were triggered by events rather than running continuously, concentrating the analysis on the moments that mattered. AI filtration was then used to distinguish normal behaviour from non-normal events and reveal the failure mechanism.