Automated generation of timing models in distributed production plants

• Published in: 2013 IEEE International Conference on Industrial Technology (ICIT) pp. 1086 - 1091
• Main Authors: Maier, A., Niggemann, O., Koester, M., Gatica, C. P.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.02.2013
• Subjects: Computational modeling; Learning automata; Mathematical model; Predictive models; Probability density function; Production; Timing
• ISBN: 1467345679; 9781467345675
• DOI: 10.1109/ICIT.2013.6505823

A key challenge in model-based design approaches is the modeling of time-dependent systems. A manual modeling often proves to be very demanding in terms of time and cost, because exact knowledge of the device or process to be modeled is required. This again requires highly skilled human resources. The development of an exact model can be tedious and requires expert knowledge gathered by human observations. For this reason, in this paper two methods for the automatic generation of models are given: (i) the automatic generation of device models, which comprise the timing behavior of network components and (ii) the automatic learning of the timing process behavior model of the production plant. Further two application scenarios are given: the verification of design goals in the early phase of the design flow (model-based design) where the device models are used to simulate the network behavior and the anomaly detection in the running plant where the prediction of the process behavior models are compared with the current behavior. The learned models can be used in model-based design and for the anomaly detection.