Funding entity Ministerio de Ciencia y Tecnología
This project concerns the development of a software tool based on artificial intelligence techniques in order to help carry out the diagnosis of anomalies and to suggest predictive maintenance criteria for machines and industrial processes. This tool will be based on a multiagent distributed achitecture flexible enough in order to carry out diagnosis and local or remote maintenance of machines or industrial processes. It will be based on elements capable of collecting continuous information coming from sensors or from tests done on machines or processes of interest, on efficient modelling instruments used for the detection of abnormal working conditions, on expert knowledge and on procedures of diagnosis of malfunctions and on knowledge of maintenance crteria. All these elements will constitute the multiple agents of the system and they will be distributed, that is to say, they will be located in those computers or in the machinery points or processes where it is most convenient. Some agents will be intelligent systems based on knowledge and yet others will be simple autonomous programs.
All of these agents communicate among themselves using communication based on CORBA technology and so the tool being develpoed will be open in order to be able to incorporate new agents developed with different concepts and forms of technology in a very flexible manner, avoiding new investments in the updating and redesigning of software.
The tool to be developed is planned to be used in two very different industrial fields. The first of them will be directed towards the maintenance of railways material and more concretely towards some componenets that are important on trains. The second is directed towards the diagnosis and maintenance of machinery for measuring energy and the real state of health of the componenets of the first case, and detecting problems regarding the measurement of energy that can cause considerable economic losses in the actual competitive electricity market.
MEPIAD
