Predictive Maintenance

Implementation rests on three things. Data, time and analysis. Especially important is the failure history, meaning the ability to identify past machine failures. Often the information is missing exactly at the intersection between a fault occurring and it being recorded. We address this gap at the start of every project.

Yes, then the sensors become part of the project. Standard IoT sensors are not sensitive or precise enough for some applications, for example in component manufacturing. In such cases, the measurement technology is designed together with the solution.

The foundation is already in place. Condition monitoring shows the current state. Predictive maintenance turns it into a prediction. Your existing sensors deliver the data on which the models are trained.

A standard solution fits when the problem can be limited to a few sensors, is well defined and the data is already available in a standardized form. For more complex plants with many different sensors and complicated workflows, a custom-built solution is the better fit. It is often more expensive at first, but costs less time and money in the long run.

Both are possible. Standard solutions are particularly suitable for companies with cloud infrastructure. If your data must not leave your premises, the solution runs in your own environment.

That depends on the costs that unplanned downtime causes today. The more expensive an hour of downtime and the more critical the asset, the faster the prediction pays for itself. That calculation belongs in the consulting phase at the start of a project, not in a blanket statement.