Research & Realization Of Hierarchical Intelligent Control System For Sinter Through Point

Entering 21 Centuries, Our iron and steel Production has risen from 120 million tons to 300 million tons, it plays a important role in Gross Domestic Product, intensives management and social environment protection compelled steel enterprise to rise labor productivity, reduce energy consumption and lengthen the service period of equipment. As shows, it possesses very suggestive meaning to heighten automatic level, research and development of the sinter hierarchical intelligent control system (SHCS) of self-determination knowledge property right.Sintering process is a dynamic system with long production line, multivariable and complex mechanism, it is hard to resolve globe problem with using classical control theory and methods. In recent years, artificial intelligence technology such as expert system, fuzzy control and artificial neural network has been induced to sintering process control, which provides an efficient approach to implement the computer control in sintering process.Burning through point (BTP) control is very important chain in sinter smelting process, it utilizes the propagation between solid, melting state and gas state and the process of regulating water carbon and composition. Due to large delay time, time-varying and strong coupling, it need still man-inspecting and regulating, and restraining heavily the automatic level of whole sintering process. Based on 16 year's researching sinter mechanism and unite-adjusting way of among strand , roll, belt machine, absorbing wind and blower, BTP hierarchical intelligent control system and correlative methed & technology is presented and built by comprehensive utilization of sinter theory, computer technology, control technology, system engineering and artificial intelligent technology.1. According to the layer of a acquiring message, dispatching instruction, flowing operator signal and process variable, The structure of SHICS (Sinter hierararchical intelligent control system) is analysed and compared, it's system structure is divided into universal decision, local select and field control.2. According to the characters of the undetermined, uneven and lower signal/noise ratio of multi-sensors, such as temperature, pressure, flowing, electric value, material level, speed and so on, these signals must be acquired, processed and made fusion, and made a synthesis, made up absorbing advantage and abandoning disadvantage, and constructed whole essence property of the plant. 3. According to the large delay characteristic of the burning through point (BTP), the fusion technology of forward and inverse network, feedback and compound control, analysis predict and control on line has been proposed in the theories and methods. That is, making the mass process message to be divided into subclasses with different feature by using Adaptive Pattern Cluster Map (APCM), then building self-organized feature map (SOFM) to predict the BTP, and making the predicting value to adjust dynamically membership of fuzzy controller (the whole correlative and unique of adjusting technologist have been proved in theory), and those techniques can be used to solve the large-delay problem efficiently.4. Facing to multi-variable coupling characteristic of the predicting and controlling of burning through point, the main composition system is used to orthogonal process for multi-variable, multi-sample sets, and enabled the main composition to exert the influence on predicting and controlling system. Selecting the speed of strand as main-adjusting value, material thick and windbox as assistance value, the main fuzzy controlling system of the strand speed is established, so that the position of burning through point will be realized with effectively and quickly.5. According to the time-varying and multi-mode of state parameter of the predicting and controlling of burning through point, such as the pressure and temperature of 9~#,12~#,14~# ,16~# windbox, the flame status of the field expert observing, a assist inference mechanism of BTP are proposed in order to represent the fuzzy characteristic of knowledge in sintering field completely. The inference mechanism is to realize general goal by hybrid inference of fuzzy diagnosis and backward reasoning, and the factor that give rise to the abnormities and control decision are inferred by forward reasoning. The BTP predicting and controlling function are achieved, which can strengthen people energy to revise and control burning through point, and restrain man-made at will and heighten whole sinter control level. Especially, when an appealing raw material fluctuating and an exterior disturbing of material stack changing, the method for diagnosing the typical abnormities of sinter process is investigated, the fuzzy diagnosis strategy of abnormity of sintering process is put forward, the sintering process would be new balance and stable. Based on sinter field expert operating and regulating experience on long production process, for having the causes and effects, but hard to describe those process parameter in allowance and no way to measure directly, comprehension inference method and non-accurate inference model can be put forward as quick as possible. 6. The hierarchical artificial intelligence control system for adjusting burning through point is developed based on multithreading working model with using VC++ as system program and SQL server as database, one thread is used to drill the parameter of neural network, two is used to build database and call program each other, three is used to run system, and four is used to display on screen. The system has been applied in operating and controlling BTP in 300 m~2 sinter plant of Magang Iron and Steel Company on-line for 6 months, the usage factor of sintering is raised 0.1%, and the working ratio is up 1%.