Automatic Dynamic Synthesis of Context Models for the purpose of Pump Monitoring

  Copyright: PLT Aachen

In the context of an industrial cooperation between the chair of process control technology and the Oil Refinery Emsland (ERE) of the German BP in Lingen, we agreed on the objective of using selected refinery pumps to automatically gather as much information as possible regarding the state of individual pumps. The idea was to identify the required pump parameters from the existing measurements with a model-assisted approach. The problem is that there are no measurements that are specifically assigned to individual pumps. The relevant measurements are the result of different switching states of the plant. As a result, the appropriate models must dynamically and automatically adapt themselves to the relevant state.

The developed monitoring application follows an existing design pattern that is composed of six software layers. The design pattern prepares the existing plant information systematically in order that the information can be uniformly used and visualized by the respective monitoring functions. One must take into account that the higher layers are dependent on the lower layers insofar as they obtain information from those layers, but not vice versa. Hence, the application can be structured on a modular basis and easily extended.