Data Collection In Preventive Maintenance
Preventive maintenance is essential for effective running of the industry; it is one of the highlights of industry 4.0. Accurate data needs to circulate round the industry in order to be proactive in preventive maintenance.
What Data Do We Collect For Preventive Maintenance?
Predictive maintenance must be prioritized in those points or systems where a failure can be critical. For failures in parts of production where a failure does not significantly affect production or involve high cost, preventive maintenance will suffice.
In the case of Thermolympic, it was started up and tested on 2 production lines enabled for this purpose and in which the data readings were carried out over time with the following objectives:
- Preventive detection of anomalies and failures,
- Predictive Maintenance,
- Optimization of operating parameters,
In this case, with the data collected and through Deep Learning algorithms, the necessary patterns were found for the purposes of the company, not only for predictive maintenance.
What data do I analyze for predictive maintenance in my company? Each company has its uniqueness, we advise you on the project .
Where Is The Data Collected For Pattern Building And Learning?
Predictions for preventive maintenance will be based on the historical data available to the company and the existing conditions in the plant.
- Information on production parameters and variables related to the injection process : temperature, pressure, injection cycle time, dosing times, temperatures of the injection axes, spindles, flowmeter information, etc.
- Environmental conditions such as temperature and relative humidity in the factory.
- Context data : material used, part being manufactured, information from the different sensors on the machine, etc.
- Information related to inspections entered by operators, quality laboratory, events, etc.
IoT is used to collect data in real time, that is, through sensors that produce large amounts of data in real time and sent to a database that must subsequently be analyzed using Big Data Analytics.