Predictive Maintenance Model-Based Approach for Objects Exposed to Extremely High Temperatures Kosta P. Boshnakov University of Chemical Technology and Metallurgy - Sofia 1756 Sofia, Bulgaria kb@uctm.edu Lyubka A. Doukovska Institute of Information and Communication Technologies, Bulgarian Academy of Sciences 1113 Sofia, Bulgaria doukovska@iit.bas.bg Emil G. Mihailov University of Chemical Technology and Metallurgy - Sofia 1756 Sofia, Bulgaria emil@uctm.edu Venko I. Petkov University of Chemical Technology and Metallurgy - Sofia 1756 Sofia, Bulgaria venko@uctm.edu Svetla I. Vassileva Institute of System Engineering and Robotics, Bulgarian Academy of Sciences 1113 Sofia, Bulgaria svassileva@yahoo.com Stefan L. Kojnov Institute of Information and Communication Technologies, Bulgarian Academy of Sciences 1113 Sofia, Bulgaria slk@iinf.bas.bg Abstract—The present paper is an attempt for acceptable balance between theoretical research, computer simulations and experiments on real industrial objects. Several rules and recommendations are retrieved from the obtained results about the exploitation domains of certain classes of industrial metallurgical objects. The practical trend of the presented results is contained in our efforts for a comparative analysis and also to form rules of the presented methods for intelligent predictive diagnostics. An original method is proposed for predictive maintenance of a metallurgical ladle based on simulations of diagnostics and forecasting of the current state of high- temperature objects. Knowing the condition of the insulation of metallurgical ladles especially at the end of the campaign is essential for operational management in a case of local damage. Keywords—Predictive Maintenance, Technical Diagnosis, Infrared Thermography, High-Temperatures Objects, Model-based diagnostics I. INTRODUCTION The model-based technical diagnostics is the well- developed theoretical direction in theoretical diagnostics, [1- 5]. To a large extent this is due to the following circumstances: • The approach developed by specialists in control theory and identification which transfer approved in theory and practice methods in diagnosis with slight modifications. • The scientific community has lots of adequate mathematical models elaborated for control, optimization, simulation and decision making. • A significant positive experience is accumulated in the practical use in areas such as automotive, chemical, biotechnology and other industries. In the present paper an original method for predictive maintenance of a metallurgical ladle based on simulations of diagnostics and forecasting of the current state of high- temperature objects is proposed. Knowing the condition of the insulation of metallurgical ladles especially at the end of the campaign is essential for operational management in a case of local damage. With the thinning insulation and with the more possible emergence of large wear due to deficiencies in masonry, hydrodynamic effect of the liquid metal flow or the chemical interaction, the degree of risk of rupture increases greatly. Hence timely inspection and assessment of the remaining undamaged insulation layer is necessary for the correct residual resource planning of the ladle, production schedule of the remaining ladles and also determining the optimal time to shift the ladle for current support or for a complete masonry. Early stopping the ladle of service leads to undue loss of underutilized labor input and the late masonry of the ladle produces increased heat losses of the liquid metal and also a need for additional heating. In cases of profound failures of local isolation, the unrecognized situation may lead to failure of metal break through the steel casing with significant production losses and a danger for the staff. The surface temperature of masonry units can serve as a reliable source of information to forecast the state of the obtained hot spots in separate regions along the surface of the aggregates. Infrared cameras are the perfect solution for contactless monitoring of high-temperature equipment during its operation, [6, 7]. Temperature measurement and analysis of results makes it possible to prevent accidents and damage from breakthrough and also to avoid consequences associated with additional and extraordinary expenses, metal loss, stays, etc. Modern infrared scanning systems work within a temperature range of -50 0 С up to 2000 0 С and a sensitivity of 0.1 0 С which allows to be found even the slightest temperature variation in the installation, [8, 9]. This enables early detection This work has been partially supported by FP7 grant AComIn № 316087 and partially supported under the Project № DVU-10-0267/10.