Application of the REx5 Expert System
to Condition Monitoring of the Simple
Liquid Flow Process Control Trainer
Sebastian Rzydzik, Krzysztof Psiuk, Marcin Amarowicz
and Tomasz Rogala
Abstract This work deals with the problems that are related to the condition
monitoring of the selected technical object. For this purpose, specialized software
called REx5 was developed. The REx5 is a modular expert system. Inference
module of this system allows to conduct the inference process for two types of
knowledge representation: the intuitionistic and Bayesian statement networks. All
data required for the inference processes are stored in especially designed OPC UA
Server with archive database of process data. Another module, designed to manage
diagnostic tasks, allows to configure the parameters for condition monitoring of
observed technical object. The simple liquid flow process control trainer is the
studied object. This trainer allows for various control functions, like the control of
the medium level in the tank, control of the flow rate, or control of the medium
temperature. Additionally, the trainer is equipped with a data acquisition subsystem.
The purpose of this subsystem is the acquisition values of the selected process
signals and sharing these data for external systems used by OPC UA Server.
Keywords Diagnostic system
Expert system
Diagnostic inference
Intuitionistic network
Bayesian network
S. Rzydzik (&) K. Psiuk M. Amarowicz T. Rogala
Institute of Fundamentals of Machinery Design, Faculty of Mechanical Engineering,
Silesian University of Technology, 18A Konarskiego Str, 44-100 Gliwice, Poland
e-mail: sebastian.rzydzik@polsl.pl
K. Psiuk
e-mail: krzysztof.psiuk@polsl.pl
M. Amarowicz
e-mail: marcin.amarowicz@gmail.com
T. Rogala
e-mail: tomasz.rogala@polsl.pl
© Springer International Publishing AG 2018
A. Timofiejczuk et al. (eds.), Advances in Technical Diagnostics,
Applied Condition Monitoring 10, DOI 10.1007/978-3-319-62042-8_29
323