(145e) Ontology Based Decision Making for Process Control | AIChE

(145e) Ontology Based Decision Making for Process Control


Dorneanu, B. - Presenter, University of Surrey
Arellano-Garcia, H., Brandenburgische Technische Universität Cottbus-Senftenberg
Heshmat, M., Sohag University
Gao, Y., University of Surrey
In the context of Industry 4.0, engineering systems and manufacturing processes are becoming increasingly complex, combining the physical world of the processing units with the cyber world of the wireless sensing and communication networks, big data analytics, ubiquitous computing and other elements that the Industrial Internet of Things technologies. The fields of ontology, knowledge management and decision-making systems have matured significantly in the recent years and their integration with the cyber-physical system (CPS) facilitates and improves the effectiveness of decision-support systems (DSS) [1]. Yet, this comes with an increase in the system’s complexity and a need for deployment of intelligent systems for process systems engineering (PSE) applications, that should adapt to the continuously new requirements of Industry 4.0. Considering the large number of devices existent in a CPS, distributed methods are required to transfer the computational load from centralised to local (decentralised) controllers.

This has led to the motivation of applying multi-agent systems (MASs) methodologies as a solution to distributed control as a computational paradigm [2]. An agent can be defined as an entity placed in an environment that can sense different parameters used to make a decision based on the goals of the entity. A MAS is a computerised system composed of multiple interacting agents exploited to solve a problem. Their salient features, which include efficiency, low cost, flexibility, and reliability, make it an effective solution for solving tasks [3]. Usually, DSS adopt a rule-based or logic-based representation scheme [4].

For this reason, ontologies have attracted the attention of the PSE community as a convenient means for knowledge representation [1, 5]. An ontology is a formal representation of a set of concepts within a domain and the relationships between those concepts, and it serves as a library of knowledge to efficiently build intelligent systems and as a shared vocabulary for communication between interacting human and/or software agents [6].

In this paper, a multi-agent cooperative-based model predictive control (MPC) system for monitoring and control of a chemical process is proposed. The system uses ontology to formally represent the system knowledge at process, communication and decision-making level. The application of the proposed framework is discussed for a chemical process that produces iso-octane. A cooperative MPC is implemented to achieve the control of the plant. This protocol is defined using a simple algorithm to reach an agreement regarding the state of a number of N agents [7]. The monitoring feature is defined by means of a MAS, consisting of follower agents (FAs), a coordinator agent (CA) and a monitor agent (MoA), that is integrated with the MPC. The MAS has two main tasks: a) decide optimal connectivity between the distributed MPCs for safer and better operation; and b) monitor the system and detect any deviation in the behaviour.

The addition of the MAS makes the cooperative MPC controller more efficient by taking advantage of the communication between the various elements of the CPS. Using the knowledge form the ontology and the agents’ sharing capabilities, the system can detect faster any deviation compared to standard operation. The framework can easily be adapted for other control approaches by very simple modifications in the structure and objectives. A practical demonstration in a pilot plant environment is envisaged for the future.


[1]. Munoz, E., Espuna, A., Puijganer, L., 2010, Towards an ontological infrastructure for chemical batch process management, Computers & Chemical Engineering 34, 668-682

[2]. Xie, J., Liu, C.C., 2017, Multi-agent systems and their applications, Journal of International Council of Electrical Engineering 7 (1), 188-197

[3]. Dorri, A., Kanhere, S.S., Jurdak, R., 2018, Multi-Agents Systems: A survey, IEEE Access 6, 28573-28593

[4]. Nowak-Brzezinska, Wakulicz-Deja, A., 2019, Exploration of rule-based knowledge based: A knowledge engineer’s support, Information Sciences 485, 301-318

[5]. Wiesner, A., Morbach, J., Marquardt, W., 2011, Information integration in chemical process engineering based on semantic technologies, Computer & Chemical Engineering 35, 692-708

[6]. Batres, R., 2017, Ontologies in process systems engineering, Chemie Ingenieur Technik 89 (11), 1421-1431

[7]. Olfati-Saber, R., Fax, J.A., Murray, R.M., 2007, Consensus and cooperation in networked multi-agent systems, Proceedings of the IEEE 95 (1), 215-233