The avoidance of unplanned downtimes is realised without elaborate sensor systems and without time-consuming model training and uses expert knowledge about downtime mechanisms. Digitalisation for the use of data.
Large industrial plants in different process industries have to achieve very high levels of availability in order to keep within cost constraints. These plants are in continuous operation. They are only shutdown for large maintenance activities. Every shutdown is potentially damaging. In particular, unplanned shutdown procedures after plant equipment failures are expensive. This is due primarily to the costs of subsequent damage that can occur anywhere in the entire plant, and secondly, due to the lost production capacity during the shutdown. A period of unplanned shutdown can be much longer than planned shutdowns because of possible delays in obtaining spare parts.
Unplanned shutdowns must be avoided. In larger plants that consist of coupled functional units, problems in seemingly insignificant peripheral systems can lead to disturbances in the overall plant via the effect chain, or even make a shutdown necessary. The costs or even the extreme conditions in the plant often prevent equipping it with sensors. Monitoring the plant state must therefore succeed using existing information.
Data is available from plant control, maintenance, and the downtime history that is useful for pattern recognition and trend analysis. The plant failure risks, their indicators, and the damaging conditions are first of all collected for appropriate automated analysis. Based on this, algorithms are developed and calibrated to detect deviations. Uptime EXPERT adds to the indicators the cause diagnostic and the remaining lifetime. A recommender system is then developed together with the service team that supports the control centre and the technicians on-site to resolve issues.
