Geicos blog

Predictive maintenance in parts washers: signals, data and interventions

Written by Geicos Group | Feb 17, 2026 7:30:00 AM

Industrial parts washers often operate silently: they do their job cycle after cycle, until one day something changes. Times get longer, results become less consistent, consumption increases, filters clog more often. At that point, maintenance becomes urgent, and machine downtime becomes a problem in the worst possible way: unplanned.

Predictive maintenance was created to avoid this scenario. It doesn't require technological revolutions, but a management approach that starts from already available data (even simple ones) and leads to more timely and targeted decisions. In this article, we'll look at how to apply it to your business.

 

The differences between corrective, preventive and predictive maintenance

Corrective maintenance

Post-failure intervention. This is the most costly in terms of plant downtime and the knock-on effect (delays, rework, urgent spare parts).

Preventive maintenance

Intervention based on timeframe (operating hours, number of cycles, calendar). This reduces breakdowns, but can lead to unnecessary maintenance and doesn't always detect non-standard problems (variable contaminants, different loads, actual usage).

Predictive maintenance

Intervention based on actual machine and process conditions: signals and trends are observed to predict when performance begins to degrade. This is the approach that condition monitoring programs formalize at a general level (ISO 17359 guidelines).

 

The parts washer speaks before the failure: the signs to monitor

A parts washer rarely goes from perfect to down without warning: your goal is to learn to identify the signs. The most useful ones fall into four categories:

Signs of consumption

  • Energie: increase in kWh per cycle or per batch for the same production;
  • Water: higher than normal volumes or replenishment increases;
  • Detergent/Additives: abnormal consumption, more frequent bathroom corrections;

These deviations often anticipate problems with filtration, thermal efficiency, losses, non-optimized cycles.

Cycle time and productivity signals

  • Longer average cycle at the same load;
  • Micro downtime (waits, resets, restarts);
  • Reduction of effective capacity (fewer pieces/hour).

They are simple indicators, useful even without sensors: cycle counters and a time log are enough.

Cleaning performers Signs

  • More rework/rewashes;
  • Recurring residues in specific areas (cavities, holes, shadow zones);
  • Variability between batches.

Here, predictive maintenance is linked to quality: loss of performance is often a symptom of mechanical problems (clogged nozzles, failing pumps), chemical problems (degraded bath) or management problems (incorrect load).

Signs of system clogging and stress

  • Filters that clog more often;
  • Delta pressure increasing (if available);
  • Pump working outside the curve (noise, vibrations, temperature);
  • Repeated alarms on levels, temperatures, pressures.

ISO 17359 cites typical parameters for condition monitoring (vibrations, temperature, flow rates, contamination, power) precisely because they are correlated to the real conditions of the machines.

 

How to identify bath degradation

In many applications, the stability of the washing bath is as important as the mechanics. Some practical signs of suboptimal operation are:

  • Increased foam or abnormal behavior in the tank;
  • Reduced performance without cycle changes;
  • Odor/color abnormal compared to the usual profile;
  • More abundant or rapid sedimentation;
  • Filters saturated more quickly.

Predictive management, here, means avoiding calendar-based replacements and moving to conditional replacements, defined on simple thresholds (hours/cycles + performance indicators and/or internal tests).

 

Alerts and thresholds without disrupting the department

Predictive maintenance works if it's light. The best method is a three-level progression..

1) Operating thresholds based on already available data

Ideal for starting immediately:

  • Cycles/day and total cycles;
  • Average cycle duration;
  • Number of rewashes;
  • Filter replacements;
  • Water and detergent refills.

To give a practical example: if the average cycle duration increases by 10-15% for three days, a quick check is opened: filters, nozzles, parameters, pump status.

2) Alerts on trends, not on single events

The best thresholds look at the trend, not the isolated peak. For example:

  • Three consecutive values ​​out of range;
  • Progressive increase week over week;
  • Increasing standard deviation (variability).

Evaluating processing trends avoids false alarms and makes the alert credible.

3) Integration with third-party technologies

This is where dashboards, telemetry, additional sensors, CMMS, or IIoT platforms come in. The goal remains simple: the same department language with a few well-chosen indicators.

A useful principle, taken from the world of industrial maintenance, is to scale the project only after a well-defined pilot phase (clear KPIs, ownership, routines)..

 

Typical interventions

Once the signals and thresholds have been defined, a basic response table is needed. Here are some examples:

  • Increasing consumption + stable performance: check leaks, thermal efficiency, calibration, insulation, and standby mode;
  • Declining performance + saturated filters: inspect filtration, pre-filters, sediment, and upstream chip/particulate management;
  • Declining performance in specific areas: clean nozzles, check spray angles, loading, and part positioning;
  • Increasing rewashing: check bath (degradation), times and temperatures, and compatibility between contaminant and detergent;
  • Repeated alarms: prioritize sensors, levels, pumps, valves, and safety.

A few checks, in a logical order, with evidence of the results.

 

Benefits of predictive maintenance: fewer downtimes, fewer unexpected costs, more continuity

The most obvious impact is the reduction of unplanned downtime and emergencies. At the industrial level, many analyses report recurring benefits: less downtime and lower maintenance costs thanks to targeted and planned interventions. In practice, for a parts washer, the advantages translate into:

  • Higher machine availability (production continuity);
  • Reduced rewashing and variability;
  • Fewer unnecessary preventative replacements;
  • Better prediction of spare parts and maintenance windows;
  • More stable relationship between the washing process and downstream processes.

 

The contribution of Geicos Group

Effective predictive maintenance starts with a foundation: machines configured to work repeatably and ready to collect useful information.

At Geicos, we design customized industrial parts washers with a focus on robustness, accessibility, and process management. We also offer integrations and monitoring (including through third-party technologies) to simplify signal reading and operating threshold setting.

During the start-up phase, we can also provide specific training on the correct use of the machine, with operational guidance on parameters, controls, inspection routines, and best practices for maintaining consistent performance over time. This is the fastest way to transform data and signals into departmental habits that reduce unexpected events and downtime.