Principles of the design of a system for automating blast furnaces

Abstract

Blast-furnace No. 2, with a useful volume of 1033 m 3, has been in operation at the Chusovoi Metallurgical Plant since a class I overhaul in 1983, i.e.. a new overhaul of the same class will be needed in the coming years. A substantial amount of reconstruction is being planned as part of that overhaul. The existing level of automation of the furnace is one reason for its inconsistent performance indices and the uneven quality of the pig iron it produces.Taking into account the capabilities of the plant as a whole and the periods of time allotted to the reconstruction of the furnace, planners have set the goal of creating an automation system which can be efficiently installed and operated. The system should also be designed so as to improve existing smelting practices, while ensuring that all operations are performed safely and that the equipment operates more reliably. In attempting to solve these problems, the planners proceeded on the basis of the following preconditions:I. The indices which characterize blast-furnace smelting depend on the coordinated operation of the hot-blast line, stove block, and stockhouse, on the prescribed changing regime, and on stabilization of the thermal state of the furnace.2. It is best if the automation system as a whole is realized in the form of information-related control subsystems based on relatively simple but reliable programmable controllers (such as the "SEMATIK $5," "VAKS 300," etc.). These devices make it possible to collect and process a large volume of raw data, operate the furnace in accordance with prescribed programs, and -most importantly -transmit information between different objects in the automation system. Duplicates of the controllers and of certain measuring instruments in the system should be provided to enhance system reliability.3. Programmable controllers, including those which contain the so-called intelligent devices, are suited only for stabilizing individual parameters and for programmed control of process operations (such as the furnace changing regime, switching of the stoves, etc.). The more complex problems of diagnosis, prediction, and control of the smelting regime require the expertise of the process engineers and blast-furnace operators. The development of methods which can solve this problem and are based on the use of information technologies and expert systems constitutes a new area of research in control theory -informatics (information technology) [I] and variantics [2, 3].4. New methods developed to solve the given problem were not considered in the design of the automatic process control systems that were installed on certain large blast furnaces. System designers also made several other serious errors of omission in developing the mathematical application software for the upper level of such systems (for example, no provision was made for evaluating and improving the reliability of the initial data, the control algorithms were based on mathematical models that are poorly suited for the given application, no "furnace-operator-compu...