PhD Theses




Andrei ROSU

Title: Contributions to Intelligent Process Control using Programable Logic Controller Networks

Supervisor: Professor Theodor BORANGIU
Final presentation in 2011


Production process control systems have had a significant evolution in recent years due to the technological advance in information technology but also of the increasingly complex requirements regarding overall performance. One of the most relevants aspects is the development of decentralized systems that are able to react rapidly, in real time, to changes in the production environment, unlike the more traditional hierarchical centralized systems that rely on the particularities of the plant in which they work.

Multi-agent architecture tries solving the limitations of traditional control architectures by a distributed approach to manufacturing systems. According to this architecture, the system is a modular composition of components whose interaction is governed by a contractual mechanism that favors the reconfiguration, and whose purpose is to meet operational requirements.

Holonic systems deal with reconfiguration problem by introducing an adaptive control system which dynamically evolves from a hierarchical control architecture into a het-hierarchical based on self-organization and self-learning. The goal of this architecture is to develop highly decentralized systems, designed in standard modular, autonomous, cooperative-intelligent elements, able to cope with the changing manufacturing environment and that, once introduced into the system, behave like "plug and play" devices.

The field of interest in this thesis is the control of holonic manufacturing systems, based on PLC networks, capable of integrating products that come with intrinsic intelligence and use behavioral modeling of the resources involved in the planning and execution of the production process.

This thesis was carried out following two main objectives. The first one is the theoretical foundation of what intelligent control of production using PLC networks really means. To achieve this goal, the characteristics of certain types of management architectures were analyzed. The methods of implementation of control algorithms using specific PLCs programming languages were also analyzed.

The second objective was to implement a fault-tolerant holonic based production system that is perfectly transparent. The control solution, diagnosis and monitoring of this system has as outcome the integration into a modern pilot manufacturing cell with an open structure that allows: distribution of production activities in work areas, adding, deleting and reconfiguring manufacturing resources.