Predictive maintenance is a type of maintenance that is based on predicting when a machine or one of its components/tools could fail, in such a way that is possible to intervene in a preventive manner. Predictive maintenance is carried out following the identification of one or more parameters that are measured and processed using appropriate mathematical models, in order to identify the remaining time before the failure. In order to apply this method, it is needed to connect the system to the machines, real time detect data from the machines equipped with sensors, create the database on which the sophisticated analysis of Machine Learning will be applied. The main techniques are the following ones:
- Sensors & Condition Monitoring: Monitoring of the asset conditions and operational status, by measuring in real time specific parameters (vibrations, temperatures, pressures, oil level etc.) in order to detect deviations that could generate machine failure.
- Predictive Algorithms & Formulas: data collected is processed and analysed by predictive algorithms that allows to predict when an asset could require an intervention of maintenance or replacement.
OPERA PREDICTIVE APP is an application that is perfectly integrated with the machine in order to automatically gather in real time machine critical parameters (vibrations, temperatures, pressures etc.) that could generate machine failures and/or produce scraps or defective products. OPERA PREDICTIVE APP uses the emerging technologies and trends in the manufacturing sector (IoT, Big Data, AI, ML, Cloud Neural Networks) to implement predictive techniques on the device (CNCs, robots or other production equipment). Data collected enables to train the artificial neural network that predicts the future behaviour of the machine.
The application allows to visualise in one single web-based dashboard the device critical parameters and for each variable a dedicated dynamic and intuitive graph is accessible to the user. Marked by different colours, real time and predictive values over time are displayed, including the calculation of the prediction reliability.
- Management of maintenance assets;
- Management of asset critical parameters (vibrations, temperatures, pressures etc.);
- Real time data collection;
- Predictions thanks to algorithms and trained neural networks;
- Visualization of predicted values on machine dashboards & layouts;
- Generation of alerts & notifications for the maintenance team;
- Generation of maintenance requests and orders.
- Minimize unxpected machine failures/downtimes and maximize asset utilization;
- Reduce maintenance costs (do maintenance only when needed);
- Maximize asset productivity and efficiency (OEE)