| Annealing stove is one of main equipments in the metallurgy industry. It is mainly used in billet rolling heating and heat treatment. Though much exploration and successful experience have been drawn from billet rolling heating and heat treatment techniques, many problems still exist as follows: Proportion of combustible gas and air is adjusted by manual work; fail to realize dynamic regulation according to heating condition. Heating billet temperature cannot be detected in the hearth. Only by detecting temperatures of the stove side or hearth with estimation from mere experience could lead to great error on the measurement of billet temperature and affect the quality of post-processing accordingly; Overheating also exists in the heating process, and the quality of product is sway by overheating oxidation. Though infrared thermoscope is used to detect the temperature of billet when it is taken out of the stove, precision is still affected because of the oxide layer on the surface.This dissertation systematically probes into the basic principles and application of multi- signal integration and amalgamation technology, and puts forward non-linear mapping relationship between input and output signals of roller-type annealing stove steel billet temperature monitoring system combined with its technical features and requirements. On this basis, further study and exploration are made of the basic principles and features of various network signal integration and amalgamation technology, detailed comparisons are made of the technical features, application and scope between typical signal integration and amalgamation and network signal integration and amalgamation, thus provide solutions to several problems appeared in the using signal integration and amalgamation technology.During the reasoning and shaping of the feasibility scheme, further study and exploration are delved into the basic principles, steps and regulations of constructing system mathematical models in BP Neuro-network signal integration and amalgamation technology, comparisons are made among the features between single and multi-layer Neuro-network signal integration and amalgamation technology (in series, parallel connection, series-parallel connection structures of the network), and finally set upon the in series BP Neuro-network signal integration and amalgamation technology based on BP sub-network to complete the research work of roller-type annealing stove steel billet temperature monitoring system.After the system scheme is set, powerful calculation and database management capacity of MATLAB6.x, together with its Neuro-network toolbox, automatic control toolbox, GUI and GUIDE interface programming are applied to work out corresponding application programs. Raw data collecting system is combined to form a complete roller-type annealing stove steel billet temperature monitoring system, and applied to the real-time inspection of the roller-type annealing stove steel billet temperature monitoring in No. 5 Plant of Chongqing Iron and Steel Group. Trustful the reached fruit has reached the research objectives and can be put into application in various plants, if combined with improved hardware system, whilst the network system database has to be improved and strengthened through heating processing. |