RU2209837C2 - Method for controlling of processing power unit - blast furnace - Google Patents

Abstract

FIELD: metallurgy. SUBSTANCE: method involves measuring and registering output parameters and controlling actions providing for supporting of output parameters within set norms; preliminarily setting values of output parameters and controlling actions in accordance with normal values; forming "output parameter-controlling actions" pairs of states; setting qualitative dependence values in "normal value", "above normal value" or "below normal value" categories and sequence of variations in controlling actions required for bringing output parameters to normal value; providing stepped controlling, with deviations of output parameters from normal value being fixed at each step with logical feature of "above normal value" or "below normal value"; determining controlling actions and controlling action deviation sign, which are required for bringing output parameters to normal values by sequential analyzing of output parameter deviations upon changing of controlling actions for bringing furnace to subsequent state, provided that all output parameters are brought to normal values. EFFECT: provision for delivery of precise information to operator for controlling blast furnace in real time and reduced cost of testing system. 2 dwg, 3 tbl

Description

 The invention relates to the field of metallurgy, in particular to processes in blast furnaces, and can be used for other complex energy technology units.

 Many energy and energy technology units are complex multidimensional control objects with a large number of state parameters and control actions. Blast furnaces belong to such objects.

 Known methods and expert systems for controlling the progress of a blast furnace and controlling blast furnaces in which quantitative mathematical models and dependencies are used, on the basis of which a forecast is made for monitoring and control [1. Nikadziyama N., Sumigata T., Maki A. et al. Development and application of an expert system for monitoring the progress of a furnace using artificial intelligence. Teuu to Hagane, 1987, v. 73, pp. 2100-2107. 2. Incola P., Karptenen A., Seppyanen M. Experience in operating an expert system in the management of blast furnaces in the company "Rautaruukki". Steel, 1994, 9, pp. 7-12. 3. B.I. Kitaev. Domain process management. Sverdlovsk, ed. UPI, 1984. 94 p.]. At the same time, such a complex control object as a blast furnace is equipped with an expert advisor system (EA), which, based on information about the control object - a blast furnace, presented in the form of this object, produces a forecast-tip that helps the operator to take the right one in each specific situation decision.

 However, the known methods and the expert systems of energy-technology objects, in particular, blast furnaces that implement them, use quantitative mathematical models and dependencies, instantaneous material and thermal balances as the basis for the control forecast, which in the conditions of such a complex object as a blast furnace having multi-dimensional multi-parameter systems management does not provide a sufficiently accurate and timely forecast. At the same time, to solve such a series of problems, it is necessary to install very complex and expensive computing equipment such as supercomputers capable of highly branched calculations, and at the same time in real time, for issuing timely forecast-advice.

 Thus, the forecast method for control and the expert system of the control object - blast furnace are known [2. Incola P., Karptenen A., Seppyanen M. Experience in operating an expert system in the management of blast furnaces at Rautaruukki. Steel, 1994, 9, pp. 7-12], which provides for the use of quantitative physico-chemical mathematical models as the basis for forecasting advice to the operator and which are closest to the proposed technical solution and are therefore selected as a prototype.

 The disadvantage of this method and expert system is the complexity of the quantitative physico-chemical mathematical apparatus used, as a result of which significant errors and significant delays accumulate in the calculation process, and the implementation of real-time forecast-advice issuing requires very complicated and expensive computing equipment.

 An object of the present invention is to provide reliable information to an operator for controlling an object — a blast furnace — in real time and to cheapen the cost of an expert system through the use of accessible computer technology in the form of, for example, personal computers.

 This task is achieved by the fact that the expert system, which implements at the preparatory stage the collection and processing of information about the object - the blast furnace, provides real-time step-by-step forecast - advice to the operator on the necessary actions (changes in control actions) on the output parameters of the blast furnace (performance, content silicon in cast iron, top temperature, pressure drop over the height of the furnace, carbon dioxide content in the top gas, etc.) from some deviated from the normal state (" Orme ") - ie.." more "or" less than "the norm - to normal" rate. " Thus, in this method, ternary logic with the signs of states is used to characterize the state of an object: "norm" (=), "more" (>), "less" (<). In this case, the normal state of the object - the blast furnace - is considered to be the state determined by technological instructions, work experience and mathematical description and modeling of processes that maintain the output parameters of the blast furnace at the required level to ensure the required performance and quality of cast iron with a minimum consumption of coke and other fuel additives, energy carriers and materials and ensuring normal (trouble-free) operation of the blast furnace. Any deviations from the normal state ("norm") to the greater ("greater") or smaller ("less") side, exceeding the limits of the permissible measurement error (1-3%), are considered deviations from the normal state and require a corresponding change in the input control blast furnace parameters. The change in control parameters is also determined within the framework of the ternary logic by the values "norm", "more" or "less" of the norm, and their choice is determined on the basis of technological instructions, work experience and mathematical description and modeling of processes in a blast furnace.

 In this case, on the basis of the technological instruction, work experience and mathematical modeling data, a logical behavioral model of the furnace is formed in the form of a certain set of “state – control” blast furnace pairs, the “state” being determined by the output state parameters, and the “control” by controlling actions in particular, deviations of the output parameters and control actions from the “normal” state.

 Using the indicated set of pairs, a logical computing device, constructed in the form of tables or a composition of blocks, establishes qualitative relationships between the output parameters and the control actions of the furnace in the form of a sign of deviation from the norm of the output parameter (“more” or “less”) and the sign corresponding to this sign ( “more” or “less”) of the control action. In addition, the sequence of changes in the control actions at each set control step is set to bring the output parameters to normal. Based on the actions of this logical calculator, depending on the state of the furnace, an advice is formed to the master technologist on the method and procedure for using control actions. In particular, at each control step, a forecast is formed of the furnace behavior and its transition to the next state, and this process continues until the predicted establishment of all output parameters to the “normal” position.

 A clear reflection of these forecast principles is the logical model of a blast furnace, presented in the form of a table. 1. Similar models can be represented as a system of blocks with corresponding relationships.

 Thus, the construction of this table and the functioning of the expert system on its basis are in accordance with the theory of discrete automata, the basis of which, as you know, is the presence of a discrete set of internal states and the properties of a jump-like transition (within a certain period of time - the interval of a discrete automaton) from one state to another [4. Glushkov V.M. Synthesis of digital machines. M .: Fizmatgiz, 1962. 476 p.].

 Moreover, it is customary for transition tables and outputs to designate the rows of the tables with the input signals of the automaton, and the columns with their states. In our case, the input signals are control actions, and the state parameters of the machine are the output state parameters of the blast furnace.

 This table, in particular, includes the 6 most important output parameters of a blast furnace (productivity, silicon content in cast iron, furnace top temperature, pressure difference over the furnace height, carbon dioxide content in furnace gas, coke humidity) and 10 control actions (blast consumption, blast humidity , temperature of the blast, oxygen content in the blast, natural gas consumption, ore load, kinetic energy of the blast, types of charge loading, level of mash) and establishes, according to the ternary logic, the correspondence to each of the three conditions values of the output parameters (<=>) of the corresponding (for normalization - "normal") value of the control actions: direct, corresponding to a change in the output parameters (<=>), or the reverse, corresponding to the reverse state of the control actions (> = <). In this case, of course, the value "norm" (=) of the output parameter corresponds to the value "norm" of the control action. In the table, direct influences are indicated by a sign (+), reverse ones are indicated by a sign (-).

 In addition, the correspondence table establishes a ranking for each change in the output parameter of the corresponding control actions. In this case, the rank number is determined as the optimal of the combination of two factors - the greatest degree of significance of the influence of a given control action on the corresponding output parameter and the least degree of influence on all other output parameters, except for the considered one, that is, the greatest degree of comunity (selectivity) of this control action.

а условием выбора и оценки номера ранга является соотношение
K_iΣ = K_iΣmax, (2)
где К_i - коэффициент усиления (передачи) параметра i по данному воздействию i, K_γi - коэффициент усиления всех остальных параметров γ по данному воздействию i.In this case, the total optimality criterion for assessing the rank of the i-th impact K _iΣ is determined by the ratio

and the condition for choosing and evaluating the rank number is the ratio
K _iΣ = K _iΣmax , (2)
where K _i is the gain (transmission) of parameter i for a given effect i, K _γi is the gain of all other parameters γ for a given effect i.

 In the table. 1, this rank of control action is reflected by numbers increasing in decreasing importance.

 The presented table reflects the most typical for the domain process correspondence of output parameters and control actions, based on a synthesis of technological instructions, production experience, calculation data and mathematical modeling of blast furnaces in a number of plants and plants. In the case of significant changes in the specifics of the blast furnace process (smelting of special cast irons, ferroalloys, etc.), this table can be somewhat modified accordingly.

In this method, when using the system for collecting and processing information (SDI), the deviation of any of the output parameters presented in the table from the state "norm" (=) to the greater (>) or to the smaller (<) side is provided, the control actions are changed in accordance with with a table. With a sign (+), the change corresponding to the output parameter, with a sign (-) - the opposite. Moreover, at each step, the current sign of the deviation of the jth output parameter from the norm is determined based on the relation
± Δy _j = y _j -y _jn , (3)
where _j is the current value of the j-th output parameter, _jn is the normal value of the j-th output parameter; Δy _j is the algebraic difference of the current and normal values of the j-th output parameter. In the case of obtaining the value Δу _j with the "+" sign, the deviation of the j-th parameter is assigned the index "more than normal"("more" or ">"). In the case of obtaining Δу _j with a "-" sign, the deviation of the j-th parameter is assigned the index "less than the norm"("less" or "<"). When Δy _j ≤δ _j (here δ _j is the metrologically acceptable deviation of parameter j from the norm, set equal to K from the range of parameter changes, here K is the accuracy class of determining the parameter), this index deviation is assigned the index "norm" or " = ".

After this assessment at each step k is selected for the j-th parameter index _j control influences the U-rank 1 in accordance with Table 1. In the case of identification of the j-th parameter in the _j symbol ">" and the matching table for the j-WASTE exposure rank 1 sign "+" je control action U _j is selected with an index of "greater" or ">". In case of identification of the j-th parameter by the index ">" and the correspondence in the table for the j-th action of the sign "-", the j-th control action is selected with the index "<". In case of identification of the j-th parameter of _{j by the} index "<" je, the control action with the sign "+" in the table is selected with the sign "<" and with the sign "-" in the table with the sign ">". The parameter at _j is considered reduced to the “norm” at this step if the corresponding control action required by the table is changed according to the table U _j .

In this case, first of all, at step k the control action, indicated by rank 1, changes. As can be seen from the table, due to the multidimensionality of the object — the blast furnace — the change at step k of one of the control actions U _jk in order to eliminate the deviation from the norm at this step one of the output parameters for _jk may entail, in accordance with the table, a change in other output parameters for _zk . To bring them to normal, at step k + 1, the corresponding control actions U _{jk + 1 are used} , primarily indicated by rank 1. If the control action, changed for the first of the output parameters deviated from the norm, coincides with the rank of this action for another the output parameter is used for this output parameter the influence under rank 2. In further iterations (k + 1) in case of coincidence of the ranks of this control action at step "k" and "k + 1" the rank of the control action is used for one Itsu is more of a controlling influence rank in the previous step.

That is, denoting the rank of the jth action U _jk in the parameter for _{jk by the} index l _jk if, at step k, the rank of the jth action U _jk in the parameter y _{zk + 1} coincides with the rank of this effect U _{jk + 1} at step k + 1 l _{zk + 1}
l _jk = l _{zk + 1} , (4)
for the z-th exposure, according to the parameter y _{zk + l,} at step k + 1, the rank of the impact is chosen to be one greater than the previous one
l _{zk + 1} = l _{jk + 1} . (5)
The iteration data continues until all output parameters of _j are brought to the “normal” position.

 The indicated iterations are implemented in a step-by-step mode in a conditional time, for example, every 30 minutes. At the first step, the control actions are changed in accordance with the required change in the output parameters and the rank of the actions. In the second step, the possible consequences of exiting the normal state of the output parameters due to the application of the first control actions, etc., are eliminated. As indicated, the iterative process continues until all output parameters are brought to the "normal" state.

 The method also provides control of the logical computing device of the blast furnace for controllability. In this case, any non-normative state of the output parameters is set and the “norm” value of all control actions is set. Next, the device is launched. A sign of controllability of a logical computing device is considered to be “normal” as a result of the operation of the device of all output parameters of the blast furnace.

Note that the direct use of logical table 1 in control processes is most convenient when deviating from the "norm" position at the same time, one or two state parameters of the domain process. With a simultaneous deviation from the “norm” state of a larger number of state parameters, the selection of control actions directly using Table 1 can be difficult due to the possible opposite influence of control actions on individual state parameters, multi-stage transition states, etc. Thus, in accordance with the theory of combinatorial analysis [5. Rybnikov K.A. Introduction to combinatorial analysis. M .: Publishing house of Moscow University, 1985.308 p.6. Combinatorial analysis. Tasks and exercises / Menshikov M.V., Revyakin A.M., Kopylova A.N. et al. M.: Science. Fizmatgiz, 1982. 368 pp.] In the presence of q sets - in this case, these are logical signs of the states of the output parameters (within the ternary logic, these are “less”, “equal”, “greater”) and the number of samples n is the number of possible combinations of parameters of the state of the object equally
k = q ⁿ . (6)
For example, with q = 3 and n = 4, the number of combinations of state parameters of an object is already 81, and with a larger number of q and n, it increases according to formula (6).

 Of course, a simultaneous deviation from the “norm” position of several state parameters at once is a rather rare phenomenon, but in principle it can take place.

 For the possible implementation of control in such complex cases using the table. 1, which in this case is a strategic or basic table, auxiliary or tactical state tables of objects are worked out in case of a simultaneous change of n state parameters. When developing these auxiliary or tactical operational tables, the well-known rules of combinatorial analysis and the data of the table are used. 1, which in this case carries out strategic functions.

 In this case, using the state tables at each time step, the required actions are worked out before the object is transferred to the “normal” state.

 Under real control conditions, the accepted number of combinations of object states, which is determined by the power-law dependence (6), depends on the availability of the corresponding memory and the speed of computing devices.

 For modern real control conditions, for example, the value n = 4-5 can be taken at q = 3 and for these conditions auxiliary operational logical tables of state parameters of the object are constructed.

 The simplification of the model of state parameters is also carried out with the decomposition of the table. 1 and dividing it into separate blocks, interconnected.

 So, one of the options for decomposition is the partition of the table. 1 into two interconnected units: a thermal state unit and a differential pressure control unit.

In this case, the thermal state block is represented by the output parameter — the silicon content of Si cast iron, the output parameters are the ore load R _H , the consumption of natural gas GH _d , the moisture content in the blast W _d , and the blast temperature T _d . At the same time, the consumption of natural gas GHG _d and the oxygen content in the blast O _2d for simplicity are considered interrelated. The differential pressure control unit is represented by the output parameter - pressure difference Δp and the control action - blast flow rate V _d .

 An example of auxiliary (tactical) tables for this case of simplification and decomposition of the table. 1 presents table. 2 and 3, constructed in accordance with the theory of combinatorial analysis.

Tab. 2 reflects three state parameters: differential pressure Δp, silicon content in Si cast iron, and current value of blast flow rate V ^T _D. The control actions are the set blast flow rate V _d and the generalized parameter E _ОB , representing in a generalized way the amount of energy introduced into the blast furnace and reflecting the influence of the control actions W _d , T _d , GH _d and R _H blast according to their effect on the thermal state of the hearth ( on the silicon content in cast iron Si) in accordance with table. 1. If there are three signs of the state of the output parameters ("less", "equal", "more") q = 3 and the number of state parameters n = 3 in accordance with formula (6), the dimension of such a table will be k = 3 ³ = 27 .

Tab. 3 works in sequence with table. 2 and establishes the procedure for connecting the indicated energy sources (W _d , T _d , GH _d and R _H ) with a generalized amount of energy E _ABOUT "more"(>) and "less"(<) norms. If there are three signs of states of the output parameters q = 3 and the number of state parameters n = 4, the dimension of the table. 3 is equal to k = 3 ⁴ = 81.

When practicing the reaction of control actions (new state) to the current situation (current values), the following principles were adhered to. Ranks of actions are selected in accordance with the strategic table. 1. Each state and reaction to it represents one step in time, while only one of the control actions is transferred to the required state ("less", "norm", "more"). In the table. 3 shows the reaction of control actions if necessary, increase E _OB in accordance with table. 2. If it is necessary to reduce E _{OB, the} signs of the variable control actions are reversed, although some variations in the ranks of the effects are possible if it is necessary to increase or decrease the value of E _OB .

In FIG. 1 presents a device that implements the proposed method. It contains: a blast furnace 1, including a loading system 2, a supply of blast 3 with the introduction of moisture 4, oxygen 5 and fuel carbon-containing additives 6 (natural gas, liquid fuel, pulverized coal, hot reducing gas, etc.), the production of cast iron and slag 7, the output of blast furnace gas 8; sensors of values - output parameters: productivity (charge descent rate) 9, silicon content in cast iron 10, top temperature 11, total pressure drop 12, CO ₂ 13 content in top gas, coke humidity 14; sensors of values - control actions: blast flow rate 15, blast humidity 16, blast temperature 17, oxygen content in blast 18, flow rate of natural gas or other carbon-containing fuel additives (fuel oil, coal dust, hot reducing gas) 19, kinetic energy of blast 20, ore load 21, loading systems A and B 22 and 23, mound level 24; computing units: systems for collecting and processing information (SDI) 25, database 26, adjustments to database 27, logical model 28, adjustments to logical model 29, signal processing unit 30, time step generator 31, data input when analyzing the system for controllability 32 , control units of the system for controllability 33, as well as an information display unit — a display 34 and, in addition, a unit for setting a frequency of measurement times 35 and an information display unit for an information collection and processing system 36.

 At the same time, the outputs of the value sensors - output parameters 9, 10, 11, 12, 13 and 14 - and the outputs of the value sensors - control actions 15, 16, 17, 18, 19, 20, 21, 22, 23 and 24 - are connected to the input the computing unit of the information collection and processing system 25, the output of the data input computing unit when analyzing the system for controllability 32 is connected to the input to the database computing unit 26, the outputs of the computing units of the information collecting and processing system 25 and the database adjustment 27 are also connected to the inputs to the computing database unit 26, computational output about the correction unit of the logical model 29 is connected to the input of the computing unit of the logical model 28, the outputs of the computing units of the database 26, the logical model 28 and the time step generator 31 are connected to the input of the computing signal processing unit 30, the output of which is connected to the input of the information display unit - display 34. The output of the system control unit for controllability 33 is connected to the input of the signal processing computing unit 30.

 The output of the unit for setting the frequency of the measurement time 35 is connected to the input of the information collection and processing unit 25, and the outputs of the blocks 25 and 26 are connected to the information display unit 36.

 The device operates as follows.

The value sensors - output parameters 9, 10, 11, 12, 13 and 14 - respectively measure productivity (charge descent rate), silicon content in cast iron, top temperature, total pressure drop, CO ₂ content in the top gas and coke humidity.

 The value sensors - control actions 15, 16, 17, 18, 19, 20, 21, 22, 23 and 24 - measure the flow rate of the blast, the humidity of the blast, the temperature of the blast, the oxygen content in the blast, the flow rate of natural gas or other carbon-containing fuel additives ( fuel oil, coal dust, hot reducing gas), ore load, determine the loading system and the level of mound.

 The signals of all these sensors enter the computing unit of the information collection and processing system 25, in which they are normalized and translated into a numerical code. The data on the values of the output parameters and the control actions enter the computing unit of the database 26, which contains data on the values of all these values, representing the "norm", i.e. the value corresponding to the normal course of the technological process in a blast furnace. Data on these normalized values of the values can be adjusted using the computing unit for adjusting the database 27. Data on the values of the values of the output parameters and their nominal values "norm" are transmitted to the computing unit for signal processing 30. Data from the computing unit of the logical model 28 are entered into the same unit , which is a table code in the form of a logical table. 1 or its tactical modifications of the type table. 2 and 3, and establishing a logical relationship between the values of the output parameters and the control actions (in the form of ratios “less”, “norm”, “more”), as well as the ranks of the corresponding control actions in the form of a priority sequence of control actions for this output parameter.

In this case, the computing unit of the step generator 31 forms a sequence of steps in time with a set time step Δt _k .

At each step k of time in the computing unit 30 determines the sign of the deviation of the current value j-th output parameter y _j of a fault in _jn, based on the ratio
± Δy _j = y _j -y _jn .

In the case of obtaining the value Δy _j with the "+" sign, the deviation is assigned the index "more", in the case of obtaining the value of Δy _j with the sign "-" - "less". When Δy _j ≤δ _j , where δ _j is the value of the permissible deviation of parameter j from the "norm", set equal to K from the range of variation of the parameter (here K is the accuracy class for determining this parameter), the norm index is assigned to this deviation of the jth parameter "

After this estimate, at each time step k, the index of control actions of rank 1 U _j is selected for a given j-th parameter in accordance with Table 1. In the case of identification of the j-th parameter for _{j with the} index “greater” (+ Δу _j ) and the correspondence in the table for the j-th control action U _{j the} sign “+” the j-th control action is taken with the sign more (>), and in the case of the sign “-” - with the sign less than (<). In case of identification of the j-th parameter, _j with the “less” sign (-Δy _j ) and, accordingly, in Table 1, for the control action U _j of the “+” sign, the jth control action is taken with the less sign (<), and in the case of “ - "- with a greater than sign (>).

The parameter at _j is considered reduced to the “norm” at this step k in the case of the corresponding selection table and fixing the change in the control action U _j .

With the further operation of the device at the next step k, possible changes in the remaining output parameters at _zk are fixed in case of a change in the control action U _jk . To bring them to normal at step k + 1, the corresponding control actions of the table are used. 1 U _{jk + 1} , primarily designated by rank 1. If the control action U _jk coincides in the rank applied for the previous output parameter deviated from _jk from the norm with the rank of this action for another output parameter y _{jk + 1} to bring it to normal of this output parameter, use the action under rank 2. In further iterations - steps in time, if the ranks of this control action coincide at step k and k + 1, the rank of the control action is used one more than the rank of the control action on previous step.

 The iteration data continues sequentially in time until all output parameters are brought to the "normal" position.

 The indicated iterations in the form of correspondence of the states of the output parameters and the control actions at each step (“more”, “less”, “norm”) are displayed on the display 34 and are the final information — the “adviser” of the technologist. When analyzing the computing device for controllability, the data on output parameters deviated from the norm are artificially entered into the computing unit of the data acquisition and processing system 26 through the data input unit 32. In this case, block 30 by block 33 switches to the controllability control mode. The process of bringing the output parameters to normal in the same sequence as described above. A sign of controllability is the reduction to the norm of all output parameters for any given variants of deviation of the output parameters from the norm. If the controllability condition is not fulfilled, the mathematical model can be adjusted using block 29 or the normalized values of the output parameters using block 27.

 When the device is operating, the current values of the output parameters and control actions and their deviations from the norm are fixed on the display 36, and the time of the frequency of fixing and averaging information is set by block 35.

The block diagram of the device for a simplified presentation of the table. 1 in the form of a table. 2 and 3 are shown in figure 2. Figure 2: 1 - blast furnace; 2 - block automatic and manual control actions (the latter is necessary in the "simulator"mode); 3 - database block (state table) for changing control actions 7-11; 4 - a database block (state tables) for changing the control action 12; 5 - site for measuring the silicon content in cast iron; 6 - node measuring differential pressure; 7 - ore load; 8 - temperature of the blast; 9 - humidity of the blast; 10 - consumption of natural gas; 11 - oxygen consumption; 12 - blast consumption; 13 - change in the energy state (index) E _{ABOUT a} blast furnace; 14 - the current value of the flow rate of the blast. In this case, the database 3 includes table. 3. Database 4 includes table. 2. In the table. 2 and 3, the value of E _OB determines the need for introducing energy impact on the top and bottom of the blast furnace when changing the silicon content in cast iron. The signs and ranks of specific impacts (W _d , T _d , PG _d and R _H ) are determined, as indicated, in accordance with table. 1.

The specified device (figure 2) works as follows. When changing the silicon content in cast iron Si - 5 and the pressure drop Δp - 6, the corresponding signals enter the database unit 4, where, in accordance with table. 2, the necessary impacts are determined in the form of changes in the flow rate of blast V _d - 7 and the E _OB index - 13. Then, if necessary, changes in the E _OB index in accordance with the database (Table 3), block 3 selects the change in the impact on the top of the blast furnace in the form of an ore load 7 and the bottom of the blast furnace in the form of effects 8-11. Block 2 implements the operation of the device in automatic mode or in the simulator mode. In specific conditions, when the device is operated, certain control actions may be subject to restrictions in accordance with technological instructions. For example, it is often required that the blast flow rate V _{D be} maintained at a constant and maximum level (“normal” value) and can only be reduced with the aim of reducing the pressure drop in the case of, for example, increased permeability of the charge.

 The advantage of the proposed method for controlling an energy-technological unit - a blast furnace - is to increase the efficiency of controlling a blast furnace, providing advice to the operator in real time and, as a result, reducing coke consumption and preventing disturbances and emergency conditions of blast furnace operation.

Claims (1)

 A method of controlling an energy-technological unit - a blast furnace, in which the output parameters and control actions are measured and recorded, which make it possible to maintain the output parameters within the established norm, characterized in that the values corresponding to the norm are set for the output parameters and control actions, and form pairs of states " output parameter - control actions ", and also establish qualitative dependencies in the categories" norm "," more than norm "or" earlier than the norm "between them and the sequence of changes in the control actions required to bring the output parameters to normal, then carry out step-by-step control, while at each control step the deviations of the output parameters from the norm are recorded with the logical sign" more "or" less than the norm ", determine control actions and the sign of deviation of control actions required to bring output parameters to normal by sequential analysis of changes in output parameters when changing control actions to transfer the furnace to the next state, provided that all output parameters are brought to normal.

Images