METHOD AND SYSTEM FOR PREDICTING A PRODUCTION LINE STANDSTILL TIME FOR AN INDUSTRIAL AUTOMATION ARRANGEMENT

摘要

The invention discloses a method and a system for providing a predicted time interval before the standstill of a production line. This task is solved by using a machine learning approach. The approach consists of several steps. In a first step, information on states or change of states of aggregates (Agg.1,..., Agg.N) of a production line is continuously collected. In a second step, the information on the aggregate (RC Agg.) which is the root cause for a bottleneck aggregate (BN Agg.) standstill is determined and the information on the time interval from the start of the root cause aggregate (RC Agg.) fault to the bottleneck aggregate (BN Agg.) standstill is obtained. In a third step, a dataset for the machine learning algorithm is prepared from this information. In a fourth step, the chosen machine learning algorithm is trained with the prepared datasets. Finally, in a fifth step the time to a bottleneck aggregate (BN Agg.) standstill with a trained machine learning algorithm utilizing newly collected data from the first step with a preprocessing described in the third step is predicted.