AI System for Prioritising Machine Maintenance Signals

The problem

When sensor series, alarms and maintenance history are held in different sources, fields have to be matched by hand before a draft maintenance review priority can be prepared.

The maintenance engineer must compare every suggestion in the maintenance review priority with its source record; missing or conflicting evidence must not take the place of a final decision.

If signal distortion caused by calibration or a change in production mode is not separated for review, an unnecessary stoppage may occur or a genuine fault may be deferred; the record must not progress without a human decision.

What we set out to improve

• Time for the draft maintenance review priority to reach the maintenance engineer with its supporting evidence
• Outcomes of records separated for human review because calibration or a change in production mode distorted the signal
• Number of suggestions corrected after identifying the risk of an unnecessary stoppage or a deferred genuine fault

How the system works

1. 01 Receive the sensor series, alarm and maintenance history from permitted sources with the record identity
2. 02 Link the source, date and version details to a common work record
3. 03 Validate fields and units
4. 04 Prepare the maintenance review priority with its supporting evidence
5. 05 Check the exception: calibration or a change in production mode may distort the signal
6. 06 Have the maintenance engineer accept, correct or reject the suggestion
7. 07 Write the approved decision and its source to the activity log

Methods we used

• A common, source-identified data schema for the sensor series, alarm and maintenance history
• A decision structure that separates deterministic rules from AI suggestions
• A review screen showing the maintenance review priority beside the raw evidence
• A mandatory human review queue when calibration or a change in production mode distorts the signal
• An event identity that prevents duplicate processing

Where people stay involved

Data and security

Sensor series, alarms and maintenance history should be processed using only the fields required for this decision. Access to the source should be authorised separately from the maintenance engineer role; personal data, commercial information and technical secrets should not be passed into the model context unless needed. Reading, suggestions and human decisions should be logged separately.

Who this suits

Frequently asked questions