Design And Implementation Of Detection System For Furnace Temperature Corrosion In Blast Furnaces

The paper treats the problem of monitoring the wear-line of a melting furnace as the inverse problem of the axi-symmetric steady-state heat equation to monitor and alarm the degree of the high temperature erosion situation. This type of industrial monitoring is important for both economic and safety reasons.Temperature sensors at different locations in the lining of the furnace are utilized for monitoring the state of the lining. By finite element method and perturbation method,the objective function is obvious. It makes the problem easy to calculate. The Tikhonov regularization is used to let the inverse problem well-posed.The numerical results are in good agreement with model data.The direct problem is successfully solved by FEM. As the domain of heat-conduction is always changing in the problem,a program used to create Delaunay mesh automatically is coded. 1-dimension compact storage technique and mesh-node-index optimization are used to decrease the storage-space occupied by the stiffness matrix.Genetic algorithm is used in solving the inverse problem, which helps to avoid local minimum. Some new schemes are also applied on the algorithm so as to adapt it better to our problem and improve the calculation efficiency and accuracy. Numerical test on both artificial and real data verifies the credibility and robust of our algorithm.