Data analysis is crucial for obtaining valuable information from available data. The method must take into account not only the limitations of the data source and available expert knowledge, but also the issues of the resources (time, money, commitment, etc.) allocated to the PdM project. Therefore, the RS data analysis method is made up of three levels - from the simplest to the most sophisticated, to increase the chances of success within the known limitations.
In order to forecast when thresholds are exceeded, you can use technological thresholds (for machines and processes) as a basis or anomaly detection to obtain baseline signal thresholds. The next step is a decision tree model that validates thresholds and builds forecasts.
Dedicated models monitor the condition of the machine (e.g. vibrations, temperatures) identifying significant changes indicating a developing failure. This approach is based on domain knowledge, operation physics and advanced statistics.
Advanced machine learning methods are used to predict specific events with a specific time horizon, and multi-dimensional anomaly detection algorithms support the operator by alerting them of an abnormal situation indicating the potential location of the deviation.
The system on a local
network or in the RS cloud
In addition to technical advancement, our solutions are focused on the economic benefits of cutting down costs