MENU

Automation agents embedded in industrial decision support systems

Dds oriented automation agents will be designed and implemented. These agents are to execute supervisory control actions now belonging to the operator, thus increasing autonomy of industrial units and so decrease dependence of unit's efficiency and safety upon the human factor.

Efficiency of industrial units is indispensably related to the satisfaction of a set of design criteria describing product quality, production time and cost, as well as a set of restrictions related to the safety of the workers and the protection of the environment. Decision Support Systems (DSS) are used in the Industrial Environment, in order to support human decisions regarding assurance of efficient and the safe operation of units. The functionality of DSS provided by modern SCADA (Supervisory Control and Data Acquisition System) systems is limited mainly to data acquisition and data processing while decision-making is left mainly to the operator. Poor exploitation of artificial intelligence techniques in these DSS limits their contribution, keeping the automation level low. The present project intends to cover the aforementioned gap, by developing DSS oriented Automation Agents in the form of software agents designed to be embedded in an abundance of commercial SCADA products. The scope of these Agents is to execute basic supervisory control actions, currently belonging to the operator. These actions are: selection of the type of field controllers, tuning of the controller parameters, provision of external control commands (set points), fault detection and management. The goal is to increase the degree of autonomy of industrial units and so decrease the dependence of unit efficiency and safety on the human factor. As an additional advantage, the operator will be provided with the possibility to deal with higher-level decisions leading to actions supported by artificial intelligence elements. The Automation Agents will upgrade the Industrial DSS efficiency as a direct consequence of superior supervisory control synthesized by generic industrial control techniques handling the bulk of restrictions and difficulties met in industrial environments. One of the Automation Agents will be the Operator Agent, while the others are dedicated to Supervisory Control selecting and tuning the field controllers. The Operator Agent aims at minimizing the human factor by assigning the handling of certain situations to an expert system, which in turn undertakes the role of operator as far as fault detection, selection of set-points, alarm/events analysis, etc. is concerned. The development of the Supervisory Control Agents will be based generally on the following methodologies: Switching Logic Controllers, Simulated Annealing Controllers, Iterative Feedback Tuning Controllers, Finite Automata Controllers. The development of the automation agents will be integrated through the design and implementation of automation frameworks that will appropriately adapt the aforementioned methodologies to each application. In regards to software implementation of Automation Agents, it is mentioned that OPC (OLE (Organisational Learning in Enterprises) for Process Control) will assure bilateral real time data transmission between different makes of equipment (i.e. digital appliances for measurement and process control, etc.). Object-oriented communication tools for heterogeneous software, like COM and ActiveX, will facilitate embedding of the new software agents into the majority of modern commercial SCADA software. Academic or non-academic partners being experienced in industrial and agricultural applications will provide the Automation Agents design specifications. Academic partners will design and test the Automation Agents in a simulation environment. The non-academic partners will undertake the software implementation of the Agents, their integration into selected SCADA software packs, and the validation of their effectiveness through experimentation in a simulation and/or an industrial environment. The following three characteristic cases will be considered to ascertain the efficiency of the Automation Agents: Manufacturing Production Line, Industrial Process, and Greenhouse. The DSS Automation Agents that will be developed constitute generalised tools applicable to numerous industrial applications. Additionally, due to their generic character, their use can be easily extended to other fields of productive activities. Keywords: decision support systems, supervisor control, integrated systems.
Acronym: 
AADSS
Project ID: 
3 219
Start date: 
21-03-2004
Project Duration: 
40months
Project costs: 
2 610 000.00€
Technological Area: 
Data Processing / Data Interchange, Middleware
Market Area: 
Process control equipment and systems

Raising the productivity and competitiveness of European businesses through technology. Boosting national economies on the international market, and strengthening the basis for sustainable prosperity and employment.