Do not let it happen: be proactive!
Predictive maintenance is one of the most frequently asked use cases for manufacturing companies. Basically you want to go beyond simple problem analysis: you want to prevent problems.
The list of points that relates to this kind of general use case are:
- you want to prevent process anomalies
- you want to prevent machines breakdowns
- you want to prevent product non conformities
Predictive maintenance is made of some typical ingredients:
Our solution can work on normal and big data, is scalable, modular and portable. We extract relevant features from the time series generated by production machines and use these features to feed an artificial intelligence algorithm. Feature engineering is the creative and most valuable part of the project. Our feature engineering process is based on statistics and mathematical models commonly used in Physics.
We generate alerts with classified risk levels, detailed with the specific part of the process/machine that is at risk.
We do have a methodology that allows us to work also on cases where there are really few historical failures or even none at all. We can also work on different failure modes and identify relationships among them or among many input signals.
Our solution can be deployed both in the cloud and on premise. The product has a client/server architecture: there are a computational server and an interactive graphical user interface, developed with web technologies. You can access the system from any connected device.
The product has more than 100 certified connectors that can be used to acquire data from every type of data source.