JP2009084636A - Method and apparatus for controlling combustion in hot blast stove - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and an apparatus for controlling the combustion in a hot blast stove with which the input heat quantity into each stove can be optimized under restricting conditions on the facility and on the operation. <P>SOLUTION: In the method for controlling the combustion in the hot blast stove, in which the hot blasting having the desirable temperature and the flowing rate, is supplied to a blast furnace with the plurality of hot blast stove sets 12, this operation of the hot blast stove is performed as the following processes, that a process, in which the input heat quantity in each stove, is made to the input power and the heat level in the blasting period in each stove is made to the output power, and the predicting value of the heat level transition from the next blasting period to the future second section to the time-series of the input heat quantity from the present to the future first section, by using a model showing a moving characteristic between the input power and the output power; a process, in which the time-series of the input heat quantity in each stove in the above first section, is obtained so as to optimize the pre-decided evaluation function in the restricting range of pre-giving on the operation/or on the facility based on this predicted values; and a process, in which the input heat quantity is adjusted in each stove, based on the input heat quantity corresponding to the present time in the time-series of the input heat quantity in each stove. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a combustion control method and a combustion control apparatus for a hot stove for supplying hot air to a blast furnace and the like, and in particular, the input heat amount in the combustion phase of the hot stove is optimized under equipment and operational constraints. It relates to control.

  In the hot air furnace, the heat storage part in the furnace constructed from bricks and the like is heated by the combustion gas during the combustion period to accumulate heat energy, and in the subsequent air blowing period, cold air is passed through the furnace to exchange heat with the heat storage part. It is a facility that obtains hot air and supplies it to the blast furnace.

  FIG. 8 is a schematic diagram showing a structure in the furnace of the hot stove. As shown in FIG. 8, the hot stove 1 includes a combustion chamber 12 and a heat storage chamber 13. A heat storage brick 19 is stacked inside the heat storage chamber 13, a portion 14 is called a dome, and a portion 15 is called a lower part of a quartz brick. In the combustion period, in the combustion chamber 12, fuel gas is supplied from the supply port 16 and combustion air is supplied from the supply port 17 to be combusted, and the combustion exhaust gas is passed through the heat storage chamber 13 so that the internal heat storage brick 19 is removed. Heat. In the subsequent blowing period, cool air is passed from the supply port 18 to the heat storage chamber 13, and hot air is obtained by heat exchange with the heat storage brick 19. By installing a plurality of such hot air furnaces 1 and repeating the cycle of the combustion period and the air blowing period while shifting the phase for each furnace, hot air having a temperature and a flow rate necessary for blast furnace operation is supplied without interruption.

  FIG. 9 shows a schematic diagram of a blowing temperature control system in the blowing period. There are cases where only one hot-blast furnace is ventilated and two are ventilated simultaneously. In the former, by adjusting the opening of the mixed cooling butterfly valve MB and mixing the cold air, in the latter, the cold air of one hot-blast furnace The temperature of the hot air sent to the blast furnace is controlled by fully opening the butterfly valve and adjusting the opening degree of the cold air butterfly valve of the other hot air furnace.

  In the combustion period, it is necessary to input the amount of heat necessary to ensure the hot air temperature during the subsequent blowing period, but it is necessary to suppress the input heat amount as much as possible in order to save energy and reduce carbon dioxide emissions. However, when the temperature falls below 573 ° C., which is the transformation point temperature, the silica bricks cause rapid expansion and may collapse. Therefore, it is necessary to keep the temperature of the quartz brick lower portion 15 at the transformation point temperature or higher even at the end of the blowing period. Further, from the viewpoint of equipment protection, the dome 14 cannot be heated above a certain temperature. It is desirable to minimize the input heat amount under such constraints, and the determination of the input heat amount is an important issue in the combustion control of the hot stove.

  For the operation of the hot stove, a single operation that mixes hot air and cold air from one furnace to adjust the temperature and supply hot air to the blast furnace, and at least two of the three or more hot air furnaces are shifted in time. There is a staggered parallel operation that ventilates at the same time, mixes the hot air obtained, and supplies it to the blast furnace.

  FIG. 10 is a conceptual diagram of staggered parallel operation. In FIG. 10, k is a discretized time and coincides with the furnace switching time. Hereinafter, the furnace switching interval is referred to as a period. The hot stove 1 (1HS in FIG. 10) finishes the blowing period at time k = 0 and enters the combustion period. The hot stove 2 (2HS in FIG. 10) ends the blowing period and enters the combustion period at time k = 1 with a delay of one period. The hot stove 3 (3HS in FIG. 10) ends the combustion period and enters the blowing period at time k = 0. The hot stove 4 (4HS in FIG. 10) ends the combustion period and enters the blowing period at time k = 1. Therefore, between k = 1 and 2, there are two air blowing states of 3HS and 4HS. Between k = 0 and 1, 3HS is a trailing furnace, which has a higher heat level than 2HS preceding 3HS and plays a role of supplying hot air at a higher temperature. Therefore, at k = 1, the hot air temperature from 3HS needs to be in a state higher than the set value of the blowing temperature to the blast furnace. Therefore, at k = 1, the hot air from 3HS and 4HS is higher than the blowing temperature set value. Therefore, 4HS waits until the time when the hot air temperature from 3HS decreases to the air temperature setting value (indicated by A in FIG. 10), and during that time, 3HS is used as a single air fan, and cold air (mixed cold air) is used as the hot air from 3HS. The blast temperature is controlled by mixing.

  In the case of single operation, the hot air temperature from each furnace must always be higher than the air temperature setting value, but in the staggered parallel operation, the hot air from the preceding furnace lower than the air temperature setting value and the air temperature setting value Since the hot air from the succeeding furnace higher than the former is mixed and used, the level of the heat storage amount of the furnace (hereinafter referred to as the heat level) can be lowered compared to the former, and the thermal efficiency can be increased. Therefore, in the case of a large blast furnace, a staggered parallel operation is generally performed.

  Conventionally, regarding the combustion control method of a hot stove, a method of determining the heat input of the hot stove after one cycle consisting of a combustion period and a blowing period and determining the next input heat amount based on the heat level is generally performed. Yes. For example, in Japanese Patent Publication No. 6-37651 (Patent Document 1) and Japanese Patent Publication No. 7-100806 (Patent Document 2), the fuzzy theory describes the heat level of a hot stove at the end of blowing, the change in the heat level, and the brick temperature. Is converted to a decision function based on the above, and a method of selecting an operation amount and determining an input heat amount based on fuzzy inference based on each actual measurement value is disclosed. As indices representing the heat level, dome temperature, quartz brick joint temperature, and cold air butterfly valve opening are used.

Japanese Patent Laid-Open No. 10-226809 (Patent Document 3) uses a hot stove simulator to determine the heat level after the end of air blowing several cycles ahead, and indicates a thermal tendency indicating whether the heat of the several cycles ahead is excessive or insufficient. A method is disclosed in which an index is obtained and the input heat quantity is corrected by the current heat level and the heat trend index. As an index representing the heat level, the deviation between the hot air temperature from each furnace and the set air temperature, the quartz brick lower temperature, and the amount of mixed cold air are used.
JP-A-1-361919 (Patent Document 4) monitors the deviation between a heat level at a certain point in the air blowing period and its target value, and sets fuzzy inference when this deviation is larger than a set value. There is disclosed a method of calculating an exhaust gas temperature final value in the combustion period by optimal feedback control when the value is less than the value, and controlling the input heat amount based on the exhaust gas temperature final value.

Japanese Patent Publication No. 6-37651 Japanese Patent Publication No. 7-100806 Japanese Patent Laid-Open No. 10-226809 Japanese Patent Laid-Open No. 1-319619

  In staggered parallel operation, two units are ventilated simultaneously, so the amount of thermal energy that the subsequent furnace should give to the cold air is affected by the heat level of the preceding furnace. For example, if the heat level of the preceding furnace is high, the timing at which ventilation to the succeeding furnace is started (point A in the case of 4HS in FIG. 10) is delayed, and the substantial blowing period is shortened. You don't have to use heat energy, and you get a state of excess heat. It also propagates to the next furnace heat level. On the other hand, if a certain furnace becomes short of heat, it propagates to the subsequent furnaces and the tendency of lack of heat continues for a long time. In staggered parallel operation, due to such inter-furnace interference, it takes a few days for the furnace heat level to reach a steady state when the input heat quantity is changed. In comparison, the overall response of the hot stove is very slow. Since the hot-blast furnace under staggered parallel operation is a system with a very large time constant in this way, in a control system that does not consider dynamic characteristics, the input heat quantity operation becomes inappropriate, and the heat level of the furnace fluctuates greatly. There is a possibility that.

  Moreover, in the prior art including the above example, since the constraint condition on the brick temperature cannot be explicitly considered, there is no guarantee that the constraint condition is observed when the calculated input heat amount operation is performed. That is, the brick temperature may overshoot during the control process, and may be out of the restriction range. Furthermore, in order to ensure the air temperature and flow rate to the blast furnace, the heat level of the furnace must be maintained above a certain level. However, in the prior art, such constraints imposed on the heat level are explicitly considered. In addition, since the large time constant of the controlled object is not taken into consideration, the heat level may fall below the lower limit in the transient state, and hot air with the conditions required by the blast furnace may not be supplied. There is sex.

Looking at the prior art from the above viewpoint, Japanese Patent Publication No. 6-37651 (Patent Document 1) and Japanese Patent Publication No. 7-100806 (Patent Document 2) are based on fuzzy inference, and the dynamic characteristics of the entire hot stove Is not considered. Japanese Patent Laid-Open No. 10-226809 (Patent Document 3) uses a predicted value of several cycles ahead by a hot stove simulator, and some consideration is given to dynamic characteristics, but it is not handled explicitly. In the control method using the optimum regulator disclosed in Japanese Patent Laid-Open No. 1-319619 (Patent Document 4), the method itself can take into consideration the dynamic characteristics, but in Japanese Patent Laid-Open No. 1-319619 (Patent Document 4). ) Has a control system for each furnace, so it does not take into account interference between furnaces, and does not consider the dynamic characteristics of the entire hot stove.
Furthermore, although the characteristics of the hot stove differ from furnace to furnace, and the heat level varies depending on the furnace, the prior art does not take these into consideration, and the input heat quantity cannot be set for each furnace.

  Thus, in the conventional technology, in the combustion control of the hot stove, the consideration of the dynamic characteristics of the controlled object is insufficient, and the constraints on the equipment and operation cannot be considered explicitly. The amount of heat input to the furnace becomes too large or too small, and there is a risk that the hot air required for the blast furnace cannot be supplied, the equipment is damaged, and the thermal efficiency is lowered. Therefore, in the prior art, the input heat amount setting with a considerable margin has to be performed, and the thermal efficiency is deteriorated.

The problems of the above-described conventional hot stove combustion control are summarized in the following four points.
(1) Interference between furnaces in staggered parallel operation and the resulting large time constant of the entire hot stove, that is, the dynamic characteristics of the entire hot stove are not considered.
(2) Since constraints on the operation such as the upper and lower limits of the brick temperature are not explicitly taken into account when calculating the operation amount, there is no guarantee that it will be satisfied.
(3) Since the lower limit of the heat level itself cannot be explicitly considered, and the dynamic characteristics of (1) cannot be considered, the lower limit value of the heat level cannot be maintained transiently, and the temperature and flow rate required by the blast furnace cannot be maintained. Hot air may not be supplied.
(4) The characteristics of the hot stove differ from furnace to furnace, and the heat level varies depending on the furnace. However, the conventional technology does not take these into consideration, and the amount of input heat cannot be set for each furnace.

  The present invention has been made in order to solve the above-described problems, and a combustion control method for a hot stove capable of optimizing the amount of heat input to each furnace under the constraints of facilities and operation, and It aims at providing the combustion control apparatus.

  The combustion control method for a hot stove according to the present invention includes a desired temperature and flow rate for a blast furnace by repeatedly performing a cycle of a combustion period and a blowing period for each of a plurality of hot stoves. In the combustion control method of a hot stove that supplies hot air, a vector whose input is the amount of heat input for each furnace is an input, and a vector whose element is the heat level in the blowing period for each furnace is an output between the input and the output. Calculating a predicted value of the transition of the heat level in the second section in the future from the next blowing period to the time series of the input heat amount in the first section from the present to the future using the model representing the dynamic characteristics of A time system of the amount of heat input to each furnace in the first section that optimizes a predetermined evaluation function within a predetermined range of operation and / or facility restrictions based on the predicted value. And a step of obtaining a and the step of adjusting the amount of heat put into the furnace based on the heat input corresponding to the current time of the time series of heat input to each furnace.

  The combustion control method for a hot stove according to the present invention includes a desired temperature and flow rate for a blast furnace by repeatedly performing a cycle of a combustion period and a blowing period for each of a plurality of hot stoves. In the method for controlling the combustion of a hot-blast furnace for supplying hot air, a vector whose input is the amount of heat input for each furnace is used as an input, and a vector whose element is the heat level in the blowing period of each furnace is used as an output. Using the model that represents the dynamic characteristics between the input and the disturbance and output as the disturbance, the set values of the temperature and the flow rate are used as disturbances. Calculating a predicted value of the transition of the heat level in the second section and, based on the predicted value, optimizing a predetermined evaluation function within a predetermined operational and facility restriction range The amount of heat to be input to each furnace based on the input heat amount corresponding to the current time in the time series of the input heat amount to each furnace in the time series of the input heat amount to each furnace in the first section Adjusting the process.

The combustion control method for a hot stove furnace according to the present invention provides a predicted value of the transition of the heat level in the second period from the next blowing period to the time series of the input heat amount in the first period from the present to the present time. The time series of the input heat quantity is obtained using the corrected value as the predicted value.
In the hot stove combustion control method according to the present invention, the operational and / or facility restrictions include at least one of the hot air temperature at the air blowing point, the brick temperature at the time of the air blowing, and the amount of mixed cold air during the air blowing period.
The combustion control method for a hot stove furnace according to the present invention is obtained by adding, as the evaluation function, the square or absolute value of the deviation between the predicted value of the transition of the heat level of each furnace and its target value for the future second section. And the weight of the input heat amount or the square of the change amount or the absolute value added for the future second interval.

  A combustion control apparatus for a hot stove according to the present invention is configured to perform a desired temperature and flow rate for a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period for each of a plurality of hot stoves. In a combustion control device for a hot stove that supplies hot air, a vector whose input is the amount of heat input for each furnace is an input, and a vector whose element is a heat level in the blowing period for each furnace is an output, and the hot air sent to the blast furnace is Using the model that represents the dynamic characteristics between the input and the disturbance and output as the disturbance, the set values of the temperature and the flow rate are used as disturbances. Predicted temperature calculating means for calculating a predicted value of the transition of the heat level in the second section, and based on the predicted value, predetermined within operational and / or facility restrictions An input heat amount calculating means for obtaining a time series of input heat amount to each furnace in the first section so as to optimize a valence function, and an input heat amount corresponding to the current time in the time series of input heat amount to each furnace Input heat amount adjusting means for adjusting the amount of heat input to each furnace based on the above.

  A combustion control apparatus for a hot stove according to the present invention is configured to perform a desired temperature and flow rate for a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period for each of a plurality of hot stoves. In a combustion control device for a hot stove that supplies hot air, a vector whose input is the amount of heat input for each furnace is an input, and a vector whose element is a heat level in the blowing period for each furnace is an output, and the hot air sent to the blast furnace is Using the model that represents the dynamic characteristics between the input and the disturbance and output as the disturbance, the set values of the temperature and the flow rate are used as disturbances. A predicted temperature calculating means for calculating a predicted value of the transition of the heat level in the second section, and an evaluation function determined in advance within a predetermined operational and facility restriction range based on the predicted value. Input heat amount calculating means for obtaining a time series of heat input to each furnace in the first section so as to optimize, and based on the input heat amount corresponding to the current time in the time series of heat input to each furnace Input heat amount adjusting means for adjusting the amount of heat input to each furnace.

The combustion control apparatus for a hot stove furnace according to the present invention calculates the predicted value of the transition of the heat level in the second period from the next blowing period to the time series of the input heat amount in the first period from the present to the current time. And a heat level predicted value correcting means for correcting the corrected value as the predicted value to the input heat amount calculating means.
In the combustion control apparatus for a hot stove furnace according to the present invention, the predicted value calculation means includes, as constraints on the operation and / or equipment, the hot air temperature at the air blowing point, the brick temperature at the time of air blowing, and the amount of cold air mixed during the air blowing period. At least one or more are used.
In the combustion control apparatus for a hot stove furnace according to the present invention, the input heat amount calculation means calculates the square or absolute value of the deviation between the predicted value of the transition of the heat level of each furnace and its target value as the evaluation function in the future. The weighting is applied to the sum of the two sections and the sum of the input heat amount or the square or absolute value of the change amount for the future second section.

  According to the present invention, the problems (1) to (3) described above are based on the assumption that the time series of the input heat amount in a future fixed section (first section) is assumed, and that the hot air considering the interference between the furnaces is used. Input into a model that represents the dynamic characteristics of the entire furnace and predict the transition of the heat level of each furnace in a certain future section (second section), and the predicted value is within the operational and / or facility constraints. Solve the problem by selecting the one that optimizes the evaluation function set in advance from the time series of the input heat amount that can be accommodated, and then operating the input heat amount based on the selected function. is doing. For this reason, there is no case where the input heat amount operation becomes inappropriate and the heat level does not exceed the restriction range. Also, in the time series of the input heat quantity manipulated variables that satisfy the constraints, the one that optimizes the predetermined evaluation function is selected and the input heat quantity is manipulated based on that, so Thus, the theoretically optimal control performance can be obtained, and the control response can be freely changed by changing the configuration and parameters of the evaluation function.

Furthermore, as a model for predicting the transition of the heat level, a vector having the input heat amount for each furnace as an element is input, and a vector having the heat level in the blowing period for each furnace as an output is an output. The problem (4) is solved by using a model representing dynamic characteristics. That is, since a model representing the dynamic characteristics between the input and the output is used, the differences in dynamic characteristics and heat levels for each furnace can be represented, so that they can be reflected in the input heat amount setting.
In addition, since the input heat amounts of all furnaces are set and calculated at the same time, compared to the case where the input heat amounts are set and calculated only for the furnace that has reached the timing of input heat amount setting, it is optimal to consider interference between furnaces. The amount of input heat can be set.
Furthermore, since the above-mentioned manipulated variable calculation process is performed at each control cycle, if there is an error between the actual hot stove dynamic characteristics and the model, or the setting value of the blast furnace air temperature and air flow rate is changed. Even if there is, control can be performed while making appropriate corrections.
In addition, if the set values of the temperature and flow rate of hot air blown to the blast furnace in the future fixed section (first section) are known, the transition of the heat level of each furnace should be predicted using a model that takes this into consideration as a disturbance. Thus, the prediction accuracy of the heat level can be improved, and the input heat amount can be made more appropriate.
In addition, as an evaluation function, for example, a value obtained by adding the square or absolute value of the deviation between the predicted value of the heat level of each furnace and its target value for a certain future interval (second interval), and the input heat amount or its change It is possible to use a weighted value obtained by adding the absolute value or square of the quantity to the future fixed interval (first interval). By changing the weight, it is possible to design a control system that takes into account the tradeoff between the amount of input heat and the obtained control performance, and a control system that balances both can be obtained.

  FIG. 1 is a block diagram showing the configuration of a hot stove control device and related equipment according to an embodiment of the present invention. This combustion control apparatus 20 corrects the predicted value of the transition of the heat level obtained by the predicted temperature calculation unit 21 that predicts the transition of the heat level using the heat level prediction model, based on the actual value. The amount of heat input to each furnace is calculated based on the predicted heat level transition value corrected by the predicted heat level correction unit 22, the storage unit 23 storing the constraint condition data, and the predicted heat level correction unit 22. The input heat amount calculation unit 24 and the input heat amount adjustment unit 25 that adjusts (sets) the input heat amount to each furnace based on the input heat amount calculated by the input heat amount calculation unit 24. Although this combustion control device 20 is specified as having such a function, its specific configuration is realized by a computer and a program that defines its arithmetic processing. A thermometer 31 for measuring the brick temperature is attached to the lower part of the quartz brick of the hot stove 1, and a thermometer 32 for measuring the temperature of the hot air is attached to the outlet side of the hot stove 1. Outputs 31 and 32 are supplied to the heat level predicted value correction unit 22. In FIG. 1, the other hot air ovens 2 to 4 are omitted, but in this embodiment, four hot air ovens are connected as shown in FIG. 9, and the hot air ovens 2 to 4 are the same as in FIG. Thermometers 31 and 32 are provided, and their outputs are supplied to the heat level predicted value correction unit 22 of FIG. A gas valve 16a is attached to the upstream side of the fuel gas supply port 16 of the hot stove 1, and the input heat amount adjustment unit 25 adjusts the opening of the gas valve 16a based on the set heat amount. Adjust gas flow or gas calorie. The operation of the combustion control device 20 will be described with reference to FIG. 1 when describing the method for optimizing the input heat amount (described later).

Next, prior to the description of the operation of the combustion control device 20 described above, the staggered parallel operation and the heat level prediction model, which are the premise of the control method of the present embodiment, will be described.
As described above, the control object of the combustion control device 20 is obtained by connecting the hot stove of FIG. 1 as shown in FIG. As shown in FIG. 2, the staggered parallel operation is carried out while repeating the combustion / air blowing cycle, and the input heat amount is set at the discrete time k in FIG. 2 prior to the combustion period of each furnace. = 1, 2,...

  At time k = 0, the hot stove 1 (1HS in FIG. 2) enters the combustion period, and its input heat amount is set. The hot stove 1 finishes the combustion period at k = 2 and enters the blowing period. The hot stove 2 (2HS in FIG. 2) is delayed by one period, finishes the combustion period at time k = 3, enters the blowing period, and enters the two-air blowing state of 1HS and 2HS. Between k = 2 and 3, 1HS is a trailing furnace, which has a higher heat level than 4HS preceding 1HS and plays a role of supplying hot air at a higher temperature. Therefore, at k = 3, the temperature of hot air from 1HS needs to be higher than the setting value of the blowing temperature to the blast furnace. Therefore, at k = 3, the hot air from 1HS and 2HS is higher than the air temperature setting value. Therefore, 2HS waits until the time when the hot air temperature from 1HS decreases to the blow temperature setting value (indicated by B in FIG. 2), and during that time, 1HS is set as a single air blow and cold air (mixed cold air) is used as hot air from 1HS. The air temperature is controlled by mixing.

  Next, an index representing the heat level of the hot stove used in the present embodiment and a control timing (timing for calculating the input heat amount) will be described with reference to FIG. In order to improve thermal efficiency, the smaller the mixed cold air, the better. However, when the hot air temperature from 1HS at k = 3 matches the blast furnace air temperature setting value, immediately enter into the two-air blowing state of 1HS and 2HS. The amount of mixed cold air is zero. That is, this is an ideal state from the viewpoint of minimizing the input heat amount and maximizing the thermal efficiency. Similarly, it is an ideal state to match the hot air temperature from each furnace with the setting value of the blowing temperature to the blast furnace at the midpoint of the blowing period of each furnace. Therefore, the hot air temperature from each furnace at the intermediate point in the blowing period of each furnace (hereinafter referred to as the blowing intermediate point hot air temperature) is used as an index representing the heat level of each furnace, and this is controlled as a control value to a target value. The target value of the hot air temperature at the air blowing point is ideally the air temperature setting value for the blast furnace, but the air temperature at the hot air at the air blowing point is set by the air temperature required by the blast furnace due to changes in the air temperature and flow rate. It must be avoided to fall below the value. Therefore, with a margin, the target value of the hot air temperature at the air blowing point is set slightly higher than the air temperature setting value for the blast furnace.

  In the present embodiment, the setting calculation of the input heat amount is performed at the timing indicated by the white squares in FIG. 2 prior to the combustion period of each furnace. Perform every period. Hereinafter, this period is called a cycle. The length of one cycle is equal to the length of 4 periods. In order to distinguish from the discrete time k in units of periods, the discrete time in units of cycles is denoted as t. Operation data necessary for setting calculation must be collected prior to that. Therefore, it is assumed that the hot air temperature at the air blowing point is collected at the end of each period, and for convenience, the data is assumed to be collected at a time one period before the discretized time when the setting calculation is performed. For example, the hot air temperature of 3HS is measured immediately before k = 1, and this is set as the blowing intermediate point hot air temperature at k = 0. In FIG. 2, the hot air temperature at the air blowing intermediate point is indicated by a black circle. To show that they are collected at the end of each period, they are plotted slightly before the vertical line representing each time.

  Next, the operational and operational constraints that should be considered when setting the input heat quantity will be described. In order to protect the quartz brick, it is necessary to keep the temperature of the quartz brick lower portion 15 at the transformation point temperature or higher even at the end of the blowing. Therefore, the temperature at the bottom of the silica brick at the end of blowing (hereinafter, the brick temperature at the end of blowing) is also used as an index of the heat level, and control is performed so that it does not interrupt the lower limit value. The brick temperature at the end of blowing is also collected at a time one period before the discretized time when the setting calculation is performed. For example, the 2HS brick temperature is measured by the thermometer 31 immediately before k = 1, and is set as the brick temperature at the end of blowing at k = 0. In FIG. 2, the brick temperature at the end of the blowing is indicated by a white circle.

In order to carry out the present invention, a model that represents the relationship between the input heat amount and the heat level is required. In this embodiment, a multi-input multi-output that represents a dynamic characteristic between the inputs and outputs a vector whose element is the heat level in the blowing period of each furnace as an input, and a vector whose element is the amount of heat input for each furnace. Use the model. First, the input vector will be described. In FIG. 2, since 1HS to 4HS enter the combustion period at time k = 0 to 3, respectively, it is necessary to set the input heat amount at those timings, but in this embodiment, setting calculation for that is performed for all furnaces. (4 units in this case) are collectively performed and the input heat amounts of 1HS to 4HS are simultaneously optimized. Therefore, after the setting calculation of heat input at t = 0, heat input u _g1 (0) of _{1HS~4HS, u g2 (0),} u g3 (0), and shall determine the u _g4 (0), the following This is expressed as a vector.

  Note that the right shoulder t on the right side of equation (1) represents transposition of the vector. Similarly, regarding the output vector, the hot air temperature of the hot air furnace 1 to 4 observed at k = 0 to 3 is regarded as the value of t = 0,

It expresses. Similarly, an input vector u (t) and an output vector y (t) at time t can be defined. The model may be anything as long as it represents the relationship between the input vector and the output vector. For example, the model can be configured as follows.

  For example, in Japanese Patent Laid-Open No. 10-226809, a heat exchange model between a combustion gas in the combustion period and a cold air passing through the furnace in the blowing period and a brick in the heat storage chamber is created and solved by numerical calculation. Discloses a hot stove simulator that performs combustion and blowing of a hot stove. In this simulator, the combustion phase is simulated using calorie and flow rate of fuel gas, air-fuel ratio, gas and brick physical property data, and the temperature change of each part of the brick in the heat storage chamber is calculated using a heat exchange model. To do. Next, simulation of the blowing period was performed using the set values of the blowing temperature and flow rate to the blast furnace, the temperature of the cold air passing through the heat storage chamber, and the physical properties of bricks and air, and the brick temperature, hot air temperature from each furnace, Calculate the time variation of the air flow to the furnace and the mixed cold air flow. By giving the initial value of the temperature of each part of the brick and repeating the simulation cycle of the combustion period and the blowing period, the time change of the index of the heat level such as the temperature of each part of the brick, the hot air temperature, and the amount of mixed cold air is calculated It is possible to obtain, as time-series data, fluctuations in the heat level with respect to the operation of the input heat amount and the change in the set values of the blowing temperature and blowing flow rate to the blast furnace.

  In order to obtain a model used in the present embodiment using this hot stove simulator, for example, the following may be performed. The amount of heat input to the hot stove 1 is changed in steps, and the data corresponding to the hot air temperature at the air blowing midpoint among the time series data of the hot air temperatures from the hot stove 1 to 4 after that, and the temperature of each part of the brick The data corresponding to the brick temperature at the end of blowing is extracted from the time series data of each, and arranged in order of time for each furnace. A step response model to can be obtained for each furnace. Similarly, a step response model is obtained for each furnace from the input heat amount of the hot blast furnaces 2 to 4 to the blowing intermediate point hot air temperature of each furnace and the brick temperature at the end of blowing.

Next, the model variables used in this embodiment will be described in detail.
FIG. 3 shows the step response model (the behavior of t _blast and t _brick when the input heat quantity is changed to a step shape). The figure (a) is the unit step operation for the hot stove, the figure (b) is the response of the hot air temperature at the air blow point at time t to the unit step operation for the amount of heat input to the hot stove, and the figure (c) is the hot step furnace. FIG. 4 conceptually shows the response of the brick temperature at the end of blowing at time t to the unit step operation with respect to the input heat amount.

As shown in FIG. 3A, the amount of heat input to the hot stove j is increased by a unit after t = 0. Then, as shown in FIGS. 3B and 3C, the hot air temperature at the air blow point at each furnace and the brick temperature at the time of the air blow at t = 0, 1, 2,. {T _blastij0 , t _blastij1 , t _blastij2 , ...} and {t _brickij0 , t _brickij1 , t _brickij2 , ...}. Here, t _blastijm and t _brickijm respectively represent the response of the hot air temperature at the hot air furnace i at the time t = m and the brick temperature at the end of air blowing at the time t = m with respect to the unit step operation with respect to the amount of heat input to the hot air furnace j. Here, after the time t = n, if the response is settled and the series of step responses is to be discontinued without any change, the hot-air furnace i blowing midpoint hot air temperature T _blastij and the air blowing with respect to the operation of heat input to the hot-blast furnace j The response of the end brick temperature T _brickij , respectively

It can be expressed as. Here, u _gj (t) and Δu _gj (t) represent the amount of heat input to the hot stove at time t and the amount of change, respectively,

It is. Since the response of the hot air temperature at the air blowing point of each furnace and the brick temperature at the end of air blowing is a superposition of responses from the input heat amount operations of all furnaces,

Y (t) is obtained from the equations (5), (6), (8), (9), and (10) using the time series of (5), and the input vector and the output vector are associated with each other.

  Next, a method for optimizing the input heat amount will be described with reference to the configuration of the combustion control device 20 of FIG.

  In the present embodiment, the predicted temperature calculation unit 21 assumes a time series of the input heat amount in a future fixed section (first section), and inputs the time series into the model, and in the future fixed section (second section). Predict changes in the heat level of each furnace. Then, the heat level predicted value correction unit 22 performs a predetermined correction on the predicted value, and the input heat amount calculation unit 24 inputs the input heat amount so that the corrected predicted value falls within the operational and / or facility restriction range. In the operation amount time series, a value that optimizes a predetermined evaluation function is selected, and the input heat amount adjustment unit 25 operates the input heat amount based on the selected input heat amount operation amount.

The operation of the combustion control device 20 will be described in more detail. As shown in the flowchart of FIG. 4, the predicted temperature calculation unit 21 sets the current time to t and covers m steps from time t to t + m−1. Assuming time series Δu (t), Δu (t + 1),..., Δ (t + m-1) of the input heat quantity operation, at the time of hot air temperature at the midpoint of air blowing over p steps from time t + 1 to t + p Series T _blast (t + 1), T _blast (t + 2), ..., T _blast (t + p) and time series of brick temperature at the end of blowing T _brick (t + 1), T _brick (t + 2), ..., T _brick (t + p) and y (t + 1), y (t + 2), ..., y (t + p) 6), (8), (9), and (10) are used to determine (S11). Here, m ≦ p. If the models of (5) and (6) completely represent the dynamic characteristics of the actual hot stove, the blast midpoint hot air temperature and the brick temperature at the end of blast obtained by these models are used as they are for control. Also good. However, in actuality, due to disturbances and modeling errors that are not taken into account, there is a discrepancy between the actual hot air temperature at the air blowing point and the brick temperature at the time of air blowing and the value calculated by the model. In order to correct this, the heat level predicted value correction unit 25 corrects the future predicted value based on the model as follows based on the difference between the predicted value based on the model and the actual value at the current time t (S12).

Here, y (t + i | t) is the predicted value after the correction of the output vector at time t + i, and y _actual (t) is the hot air temperature at the air blowing point observed at time t and the brick temperature at the time of air blowing end. The actual value of the output vector consisting of By performing the same correction, the corrected output vector y (t + 1 | t), y (t + 2 | t), ..., y (t + p | t) can be obtained. Is used to calculate the input heat quantity operation.

_Assuming that the target value of the hot air temperature at the air blowing point is T _blastref and the lower limit of the brick temperature at the time of air blowing is T _brickmin , the time series Δu (t), Δu (t + 1), ... , Δu (t + m−1) can be obtained by satisfying the constraints of the following equations (13) and (14) and minimizing the evaluation function of equation (15) (S13).

  Here, a and b are weights related to the control error of the hot air temperature at the air blowing point and the magnitude of the input heat amount, respectively, and the responsiveness of the control system can be adjusted by adjusting this. The above problem is a quadratic programming problem that minimizes the evaluation function of the equation (15) under the constraints of the equations (13) and (14), and can be solved by a known method. The time series Δu (t), Δu (t + 1),..., Δu (t + m−1) of the input heat amount operation can be obtained. The input heat amount calculation unit 25 outputs only the input heat amount operation Δu (t) at the current time t to the input heat amount adjustment unit 25 as the input heat amount. The input heat amount adjustment unit 25 sets the input heat amount from the input heat amount calculation unit 24 as the input heat amount at time t, and adjusts the gas flow rate, for example, the opening of the gas valve 16a based on the set input heat amount. To control the input heat amount (S14). Instead of adjusting the gas flow rate, the gas calorie may be adjusted. The operation of the input heat amount is similarly performed for the other three hot air ovens 2 to 4. Then, at time t + 1, which is the next timing for setting the input heat amount, the above optimization problem is solved again to obtain a time series of the input heat amount operation, and the operation is similarly performed according to the input heat amount operation corresponding to the time. . Thereafter, the input heat amount is adjusted by repeating this similarly for each control cycle.

  In the above embodiment, the input vector and the output vector are associated with each other using the step response models of the expressions (5) and (6) as they are, but the model used in the present invention is not limited to this. Hereinafter, a method using a state equation will be described as a second embodiment.

First, step response time series {t _blastij0 , t _blastij1 , t _blastij2 , ...}, {t _brickij0 , t _brickij1 , t _brickij2 , ...} is approximated by a first-order lag system, and a transfer function model is obtained from the amount of heat input to each furnace to the hot air temperature at the air blowing midpoint and the brick temperature at the end of air blowing. Since the number of elements in the input vector is 4 and the number of elements in the output vector is 8, the transfer function matrix has 4 inputs and 8 outputs. Next, this transfer function matrix is converted into the following equation of state.

  Here, A and B are parameters of this model, and are matrices with constant elements. Expression (16) is an expression for obtaining an output vector at time t + 1 from an input vector and output vector at time t. By giving the input vector and output vector at initial time and repeatedly calculating expression (16) A future output vector y can be determined.

  In order to show the effect of this embodiment, the result of having performed computer simulation is shown in FIG.6 and FIG.7. 6 and 7, the horizontal axis is time, and the unit is a period shown in FIG. Moreover, although the vertical axis | shaft is an input calorie | heat amount and a ventilation midpoint hot air temperature, it is each made dimensionless. At time k = 0, both the hot air temperature at the air blowing point and the brick temperature at the time of air blowing are both higher than the lower limit, and the input heat amount is reduced without breaking the lower limit value, that is, either is set to the lower limit value by the input heat amount operation. Reaching is the goal of control.

  FIG. 6 shows a case where the optimum regulator, which is a conventional technology, is applied, and the brick temperature at the end of the blowing finally reaches the lower limit of the target value, but there is a period during which the lower limit is broken in the process. . This indicates that the heat level of the furnace is equal to or lower than the target value during this period, and there is a possibility that the blowing temperature and blowing flow rate for the blast furnace cannot be maintained. Here, it is assumed that the brick temperature at the time of the end of blowing reaches the lower limit first, and that it reaches the lower limit that is the target value.

  FIG. 7 shows a case where the present embodiment is applied. By appropriately manipulating the input heat amount, a response such that the brick temperature at the end of blowing as shown in FIG. 6 overshoots and falls below the lower limit is observed. Therefore, good control can be performed.

  In this embodiment, the hot air temperature at the air blowing point and the brick temperature at the end of air blowing are used as indicators of the heat level of the furnace, but there is a sufficient margin in the brick temperature at the end of air blowing, and it does not fall below the lower limit value. If it is clear, it is not necessary to consider the constraint condition of the equation (14), and it is not necessary to predict the brick temperature at the end of blowing. In that case, the element of the output vector is only the hot air temperature at the air blowing midpoint. It is also possible to use the amount of mixed cold air as an index of the heat level and incorporate it into the output vector. In that case, as shown by the gray circles in FIG. 2, the amount of mixed cold air at the timing when the blowing of the preceding furnace is completed and the subsequent furnace becomes a single blowing may be used, or the mixed cold air in each blowing period. You may obtain | require the average value of quantity and use it. If there is a constraint on the amount of mixed cold air, it may be expressed by an inequality like the equations (13) and (14).

  Further, in the present embodiment, a plurality of heat level indicators can be taken into the output vector, and if there are constraints on them, they are input by expressing them as inequalities in the same manner as in equations (13) and (14). It can be taken into account explicitly in the calorific value setting.

  In addition, if you know the set values of the temperature and flow rate of hot air blown to the blast furnace in the future fixed section (first section) above, by predicting the transition of the heat level of each furnace with a model that takes that into consideration, The prediction accuracy of the heat level can be improved, and the input heat amount can be made more appropriate. The responses of the hot air furnace i at the air blow intermediate point hot air temperature and the air temperature at the end of air blowing to the unit step operation of the air blowing temperature are respectively expressed by the following equations.

Similarly to the equations (5) and (6), the response D _blasti of the blowing intermediate point hot air temperature to the blowing temperature operation and the brick temperature D _bricki at the end of blowing are represented by the following equations, respectively.

  Here, BT (t) and ΔBT (t) are the air blowing temperature and the amount of change at time t. Using this, formulas (8) and (9)

By doing, it can obtain | require the ventilation intermediate point hot-air temperature and the brick temperature at the time of completion | finish of ventilation which considered the future change of the ventilation temperature setting value. Thereafter, using the equations (19) and (20) instead of the equations (8) and (9), the input heat amount can be optimized in the same manner as in the above embodiment. Also in the case of using another heat level index, for example, the amount of mixed cold air, a predicted value that takes into account future changes in the blow temperature setting value can be similarly obtained. Further, for the change of the blast flow rate setting value, the step response from the blast flow rate to the heat level index is similarly obtained, and the equations (8) and (9) may be corrected in the same manner as the equations (19) and (20). .
In the present embodiment, the model is derived using the hot stove simulator.However, the present invention is not limited thereto, and actual values of time series data of the input heat amount and the heat level are collected in the operation of the hot stove. The model may be derived by applying a known system identification method and used.
The present invention is characterized in that the amount of heat input to all four hot blast furnaces is obtained simultaneously, and this makes it possible to set the amount of input heat utilizing a total of four degrees of freedom corresponding to each furnace. In the above embodiment, the example of performing the input heat amount setting calculation at the time t = 0, 1, 2,... (1 cycle = 4 period time interval) shown in FIG. 2 has been described, but the time k shown in FIG. It is also possible to perform setting calculation of the amount of heat input to the four hot stoves at = 0, 1, 2,... (Time interval of one period). In this case, at the time k = 0, 1, 2,..., Among the input heat amounts of the four hot stove furnaces obtained by performing the setting calculation in one period period, when k = 0, only the one for 1HS is used, and k It is sufficient to use only for 2HS when = 1 and only for 3HS when k = 2. That is, the setting calculation for four hot blast furnaces is performed at intervals of one period to obtain the input heat amount of the four units, and among the obtained four units, the furnace that first enters the combustion period (at the timing of setting the input heat amount) It is also possible to set the hot stove using only the amount of heat input to the furnace. As a result, compared to the case where the setting calculation of the input heat amount is performed at time t = 0, 1, 2,..., The setting calculation cycle becomes ¼, so the setting value of the blast furnace air temperature and the heat amount is changed. The effect of quick response to is obtained.

It is a block diagram which shows the structure of the control apparatus of a hot stove furnace and related equipment which concern on one Embodiment of this invention. It is a diagram showing the control timing in this embodiment, and the timing of data collection. It is the diagram which showed the step response model. It is the flowchart which showed the process in this embodiment. It is a diagram showing the simulation result at the time of applying the optimal regulator which is a prior art. It is a diagram showing the simulation result at the time of applying the embodiment of the present invention. It is a schematic diagram which shows the structure of a hot stove. It is a schematic diagram showing the ventilation control system of a hot stove. It is a conceptual diagram of staggered parallel operation.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 20 Combustion control apparatus, 21 Predicted temperature calculation part, 22 Heat level predicted value correction | amendment part, 23 Memory | storage part, 24 Input heat amount calculation part, 25 Input heat amount adjustment part, 31 Thermometer, 32 Thermometer

Claims (10)

In a combustion control method of a hot stove for supplying hot air of a desired temperature and flow rate to a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period at predetermined control periods for a plurality of hot stoves ,
The input is a vector whose element is the amount of heat input for each furnace, and the output is a vector whose element is the heat level in the ventilation period for each furnace. Calculating a predicted value of the transition of the heat level in the second section in the future from the next blowing period with respect to the time series of the input heat amount in the first section;
A time series of the amount of heat input to each furnace in the first section that optimizes a predetermined evaluation function within a predetermined range of operation and / or facility restrictions based on the predicted value. The process of seeking
And a step of adjusting the amount of heat input to each furnace based on the amount of input heat corresponding to the current time in the time series of the input heat amount to each furnace. In a combustion control method of a hot stove for supplying hot air of a desired temperature and flow rate to a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period at predetermined control periods for a plurality of hot stoves ,
The input is a vector whose element is the amount of heat input for each furnace, the output is a vector whose element is the heat level in the blowing period for each furnace, and the disturbance is the set value of the temperature and flow rate of hot air sent to the blast furnace. Calculates the predicted value of the transition of the heat level in the second period in the future from the next blowing period to the time series of the input heat amount in the first period from the present to the future using a model representing the dynamic characteristics between the disturbance and the output And a process of
Based on the predicted value, a time series of the amount of heat input to each furnace in the first section is determined so as to optimize a predetermined evaluation function within a predetermined range of operation and equipment. Process,
And a step of adjusting the amount of heat input to each furnace based on the amount of input heat corresponding to the current time in the time series of the input heat amount to each furnace.   The predicted value of the transition of the heat level in the second section in the future from the next blowing period to the time series of the input heat amount in the first section from the present to the future is corrected based on the actual value of the heat level at the current time, 3. A method for controlling combustion of a hot stove furnace according to claim 1 or 2, wherein a time series of the input heat quantity is obtained using the correction value as the predicted value.   The operation and / or facility restriction includes at least one of a hot air temperature at the air blow intermediate point, a brick temperature at the time of the air blow end, and a mixed cold air flow amount during the air blow period. The hot-blast furnace combustion control method described.   As the evaluation function, a value obtained by adding the square or absolute value of the deviation between the predicted value of the heat level of each furnace and its target value for the future second interval, and the square or absolute value of the input heat amount or its change amount The method for controlling combustion in a hot stove according to any one of claims 1 to 4, wherein a value obtained by weighting a value added for a future second interval is used. In a combustion control device for a hot stove that supplies hot air at a desired temperature and flow rate to a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period at predetermined control periods for a plurality of hot stoves ,
The input is a vector whose element is the amount of heat input for each furnace, the output is a vector whose element is the heat level in the blowing period for each furnace, and the disturbance is the set value of the temperature and flow rate of hot air blown into the blast furnace. Calculate the predicted value of the transition of the heat level in the second interval in the future from the next blowing period to the time series of the input heat amount in the first interval from the present to the future using the model representing the dynamic characteristics between the disturbance and the output Predicted temperature calculation means for
A time series of the amount of heat input to each furnace in the first section that optimizes a predetermined evaluation function within a predetermined range of operation and / or facility restrictions based on the predicted value. Input heat amount calculating means for obtaining
A combustion control device for a hot stove, comprising: input heat amount adjusting means for adjusting a heat amount input to each furnace based on an input heat amount corresponding to a current time in a time series of input heat amounts to each furnace. In a combustion control device for a hot stove that supplies hot air at a desired temperature and flow rate to a blast furnace by repeatedly performing a cycle of a combustion period and an air blowing period at predetermined control periods for a plurality of hot stoves ,
The input is a vector whose element is the amount of heat input for each furnace, the output is a vector whose element is the heat level in the blowing period for each furnace, and the disturbance is the set value of the temperature and flow rate of hot air blown into the blast furnace. Calculate the predicted value of the transition of the heat level in the second interval in the future from the next blowing period to the time series of the input heat amount in the first interval from the present to the future using the model representing the dynamic characteristics between the disturbance and the output Predicted temperature calculation means for
Based on the predicted value, a time series of the amount of heat input to each furnace in the first section is determined so as to optimize a predetermined evaluation function within a predetermined range of operation and equipment. Input heat amount calculating means;
Combustion control device for a hot stove furnace, comprising: a heat input heat amount adjusting means for adjusting a heat amount input to each furnace based on an input heat amount corresponding to a current time in a time series of input heat amounts to each furnace .   Correct the predicted value of the transition of the heat level in the second period in the future from the next ventilation period to the time series of the input heat amount in the first period from the present to the future, based on the actual value at the current time, and the correction value The combustion control device for a hot stove furnace according to claim 6 or 7, further comprising a heat level predicted value correcting means for outputting the predicted value to the input heat quantity calculating means.   The predicted value calculation means uses at least one or more of a hot air temperature at an air blowing point, a brick temperature at the time of air blowing, and a mixed cold air amount at the time of air blowing as the operational and / or facility restrictions. The combustion control apparatus for a hot stove according to any one of 6 to 8.   The input heat amount calculation means includes, as the evaluation function, a value obtained by adding the square or absolute value of the deviation between the predicted value of the heat level of each furnace and its target value for the future second section, the input heat amount or The combustion control device for a hot stove according to any one of claims 6 to 9, wherein a weighting is applied to the square of the amount of change or the absolute value added to the future second section.

Images