In automation, the suited scheduled maintenance is one of the keys for plants' operation in order to minimize plants' shutdown. Usually, maintenance is time-interval based or operation-time based to prevent plant shutdowns by means of displacing aged automation devices early. This causes an increasing expenditure due to the fact that the replaced automation devices could be in operation longer and the plant has to be stopped for replacing the device. In this paper we present an approach focusing on combining engineering data and historical process data to extract additional information and applying analysis methods for diagnosis. We introduce how to find cause-effect dependencies of failures during abnormal plant situations to forecast abnormal and critical plant situations.