Computer system based on a blast furnace - hot blast stove safety temperature control calculation model

Abstract

The application discloses a kind of based on blast furnace-hot blast furnace safety temperature control calculation model computer system, the system includes: feed identification module, for the smell, shape and color data of ore after ore enters blast furnace are acquired;Processing information database contains two sub-databases;Wherein, the first sub-database pre-stores the information word of ore, for realizing the matching of ore processing time, vault temperature;Second sub-database pre-stores the process word of ore, for realizing the matching of initial processing parameter of ore;Real-time monitoring module is used to collect real-time data in the process of ore processing;Model prediction module is based on ore processing information and real-time data, utilizes based on blast furnace-hot blast furnace safety temperature control calculation model, predicts and outputs the predicted parameter of blast furnace-hot blast furnace next time, and carries out furnace-hot blast furnace regulation and control according to these parameters, to realize temperature control in processing process.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of model-based control computer systems, in particular to a computer system based on a blast furnace-hot blast stove safety temperature control calculation model. BACKGROUND

[0002] In the steel industry, temperature is an important environmental factor affecting tapping rate and safety, and too high or too low temperature will adversely affect the quality of molten iron and molten steel. However, there are a large number of temperature control devices and process links in the steel industry, such as blast furnaces, hot blast stoves, electric arc furnaces, etc.

[0003] The traditional ironmaking and steelmaking temperature control method is to use valve control or manual adjustment, which has the problems of inaccurate temperature control, slow response, high energy consumption, etc. The more advanced control method is PID control, but it lacks real-time monitoring and dynamic adjustment capability for temperature distribution, and the parameters are difficult to adjust, sensitive to interference, and have performance limitations. In addition, the PID control system may be sensitive to noise and external disturbances, causing the system output to become unstable when facing high-frequency noise. SUMMARY

[0004] The purpose of the present application is to provide a computer system based on a blast furnace-hot blast stove safety temperature control calculation model, which solves the problems of inaccurate temperature control, slow response, high energy consumption, etc. in ironmaking and steelmaking processing temperature control, while taking into account the safety and service life of the blast furnace and hot blast stove.

[0005] In order to achieve the above-mentioned task, the present application adopts the following technical solutions:

[0006] The computer system based on a blast furnace-hot blast stove safety temperature control calculation model comprises:

[0007] A feedstock identification module is used to obtain the odor, shape and color data of the ore after the ore enters the blast furnace;

[0008] A processing information database comprises two sub-databases; wherein the first sub-database pre-stores information entries of the ore, which is used to realize the matching of the processing time and the dome temperature of the ore; and the second sub-database pre-stores process entries of the ore, which is used to realize the matching of the initial processing parameters of the ore;

[0009] A real-time monitoring module is used to collect real-time data in the ore processing process;

[0010] The model prediction module predicts and outputs the blast furnace-hot blast furnace next time's cold blast main pipe flow, blast furnace gas branch pipe flow, net gas main pipe flow, combustion air main pipe flow, and steam injection amount based on the ore processing information and real-time data by using a blast furnace-hot blast furnace safety temperature control calculation model, and the blast furnace-hot blast furnace adjusts and controls according to the parameters to realize temperature control in the processing process.

[0011] Further, the ore identification module includes an electronic nose for identifying the smell of the ore, and a high-definition camera for identifying the shape and color of the ore; when the ore enters the blast furnace, the smell strength and smell category of the ore are identified by the smell sampler in the electronic nose and the gas sensor array, and are converted into digital signals of the smell, which are transmitted to the processing information database together with the shape and color obtained and identified by the high-definition camera.

[0012] Further, the first sub-database pre-stores information entries of the ore, and each information entry includes the ore category, ore name, shape, color, smell, processing time, dome temperature, etc.; taking the shape, smell and color collected by the ore identification module as the search condition, the first sub-database performs information matching on the stored information entries to find the ore category matching the search condition, and obtains the processing time t and the dome temperature target value T from the corresponding entry information. g0 ;

[0013] The second sub-database pre-stores process entries of the ore, and each process entry includes the ore category and the initial processing parameters matching the ore category, including the cold blast main pipe flow initial value Q 01 , the blast furnace gas branch pipe flow initial value Q 02 , the net gas main pipe flow initial value Q 03 , the combustion air main pipe flow initial value Q 04 , and the steam injection amount initial value Q 05 ; the second sub-database selects the initial processing parameters in the corresponding process entry according to the ore category matched by the first sub-database, and transmits the initial processing parameters, the processing time t g , and the dome temperature target value T g0 to the real-time monitoring module as the ore processing information.

[0014] Further, the monitoring module includes two sub-monitoring modules, and the first sub-monitoring module is used to detect the dome temperature T g and the waste gas temperature T f in the blast furnace., optimal air-fuel ratio A / F, air humidity RH; the second sub-monitoring module is used for detecting cold air main pressure P1, cold air main temperature T1, blast furnace gas branch pipe pressure P2, blast furnace gas branch pipe temperature T2, net gas main pipe pressure P3, net gas main pipe temperature T3, combustion air main pipe pressure P4, combustion air main pipe temperature T4, atmospheric humidity rh; cold air main pipe flow Q1, blast furnace gas branch pipe flow Q2, net gas main pipe flow Q3, combustion air main pipe flow Q4, and steam injection amount Q5, and the real-time data and ore processing information are transmitted to the model prediction module.

[0015] Further, the construction process of the blast furnace-hot blast stove safety temperature control calculation model is as follows:

[0016] Based on the ore processing information and real-time data in the ore processing process, a blast furnace-hot blast stove safety temperature control calculation model is constructed; the calculation model is a nonlinear model, and the nonlinear model is converted and processed to construct a state space equation;

[0017] For the established state space equation, the Taylor series expansion method is applied to determine the coefficient matrix of the state space equation, so as to linearize the state space equation;

[0018] The state space equation obtained by linearization processing is discretized by using a zero-order holder, and a discretization system is constructed;

[0019] For the discretization system, the future system state for a period of time is predicted by using a rolling optimization method;

[0020] A quadratic control law is designed, the weight coefficients of the system state and the weight coefficients of the input variables in the quadratic control law are obtained by using an entropy weight method; the future system state for a period of time obtained by prediction is brought in, and a constraint condition is set, the system input variable sequence is solved, the first input variable in the solved sequence is extracted, and the parameter in the input variable is used by a computer system to regulate and control the processing process of the blast furnace-hot blast stove.

[0021] Further, the blast furnace-hot blast stove safety temperature control calculation model is constructed based on the ore processing information and real-time data in the ore processing process; the calculation model is a nonlinear model, and the nonlinear model is converted and processed to construct a state space equation, including:

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] wherein, t is time parameter, P1 is cold air main pipe pressure, T1 is cold air main pipe temperature, P2 is blast furnace gas branch pipe pressure, T2 is blast furnace gas branch pipe temperature, P3 is net gas main pipe pressure, T3 is net gas main pipe temperature, P4 is combustion air main pipe pressure, T4 is combustion air main pipe temperature, rh is atmospheric humidity; Q1 is cold air main pipe flow, Q2 is blast furnace gas branch pipe flow, Q3 is net gas main pipe flow, Q4 is combustion air main pipe flow, Q5 is steam injection amount; T g is dome temperature, T f is exhaust gas temperature, A / F is optimal air-fuel ratio, RH is supply air humidity, and the parameter superscript represents first-order derivative thereof;

[0031] Let state variable wherein ; input variable wherein ; output variable wherein , and superscript T represents transposition;

[0032] Then, the state space equation is constructed from the nonlinear model as follows:

[0033] ;

[0034] ;

[0035] wherein, is first-order derivative of state variable, A is system matrix, B is control matrix, C is output matrix or observation matrix, and D is direct transfer matrix.

[0036] Further, the method of applying Taylor series expansion is used to determine the coefficient matrix of the state space equation, and is specifically expressed as:

[0037] ;

[0038] ;

[0039] ;

[0040] ;

[0041] wherein, is the partial derivative, denotes the coefficient of the corresponding element after integrating the derivative, and t is the time parameter.

[0042] Further, the state space equation obtained by the linearization processing is discretized by using a zero-order holder, and a discretized system is constructed, denoted as:

[0043] ;

[0044] wherein, k is the current time, is the value of the state variable x at the k time, is the value of the input variable u at the k time, denotes the predicted system state at the k+1 time at the k time, , is a coefficient matrix of the discrete system, and has:

[0045] ;

[0046] ;

[0047] wherein, I is a unit matrix, A is a system matrix, B is a control matrix, and T is a sampling period.

[0048] Further, for the discretized system, a rolling optimization method is used to predict the system state in the future period of time, including:

[0049] the system state in the future period of time to is expressed as follows:

[0050] ; ; ;... ;

[0051] wherein, is the initial system state at the k time, and the value is the state variable collected at the k time ; is the predicted system state at the k+N time at the k time; is the input variable at the k time ; is the predicted input variable at the k+N-1 time at the k time, and N is the length of the prediction interval.

[0052] Further, the quadratic control law is designed, including:

[0053] The general form of the quadratic control law J is designed as:

[0054]

[0055] wherein Q is a weight coefficient of the system state at the moment, R is a weight coefficient of the input variable, F is a weight coefficient of the system state at the moment k+i predicted at the moment k, and is the system state at the moment k+i predicted at the moment k, i=1,2,…,N-1. is the input variable at the moment k+i predicted at the moment k, i=1,2,…,N-1.

[0056] The constraint condition of the quadratic control law J is:

[0057]

[0058]

[0059] Further, the weight coefficient of the system state and the weight coefficient of the input variable in the quadratic control law are obtained by using the entropy weight method, including:

[0060] The historical data of the currently processed ore are acquired, and fourteen parameters including the cold air main pressure P1, the cold air main temperature T1, the blast furnace gas branch pipe pressure P2, the blast furnace gas branch pipe temperature T2, the net gas main pipe pressure P3, the net gas main pipe temperature T3, the combustion air main pipe pressure P4, the combustion air main pipe temperature T4, the atmospheric humidity rh, the cold air main flow Q1, the blast furnace gas branch pipe flow Q2, the net gas main pipe flow Q3, the combustion air main pipe flow Q4, and the steam injection amount Q5 in the same prediction interval length from the moment k+1 to the moment k+N are extracted, and each data value of each parameter is taken as a sample.

[0061] A forward index of each sample is established :

[0062]

[0063] wherein is the forward index of the sample a ij , min and max are minimum and maximum operation, and a ij represents the data value of the jth historical data at the moment k+i.

[0064] Then, the forward index of each sample is calculated in sequence The proportion of the forward index in all moments , and the entropy value of the forward index corresponding to each kind of historical data :

[0065] ;​​​​​​

[0066] ;

[0067] wherein, ;

[0068] Finally, the weight of the positive index of each historical data is obtained :

[0069] ;

[0070] wherein, ; The weight coefficient Q of the system state at the moment is a positive definite matrix, and each element on the diagonal is composed of ; the weight coefficient R is a unit matrix, and the diagonal elements are ;

[0071] Then The weight coefficient F of the system state at the moment is:

[0072] ;

[0073] wherein, is the coefficient matrix of the discrete system, ;

[0074] In the above formula, I is a unit matrix, A is a system matrix, and T is a sampling period.

[0075] Further, at the initial moment k, the computer system controls the ore processing machine to start working with the initial processing parameters in the ore processing information, i.e. the initial cold air main flow Q 01 , the initial blast furnace gas branch flow Q 02 , the initial net gas main flow Q 03 , the initial combustion air main flow Q 04 , the initial steam injection amount Q 05 , and sets the processing time t g , the dome temperature target value T g0 , and starts the first cycle. The process of each cycle is as follows:

[0076] The real-time monitoring module obtains the state variable at the moment k , takes the state variable x as the initial state of the system at the moment k in the rolling optimization method , predicts and obtains the system state at the future moment to , and then takes to as into the quadratic form control law J, obtains the input variable sequence by deriving J and setting the derivative equal to zero , take the first value Controlling the blast furnace-hot blast furnace at the next moment k+1;

[0077] At the next moment k+1, start the second cycle; and repeat the above cycle until the ore processing is completed or the processing time t is reached g And the vault temperature target value T g0 .

[0078] Compared with the prior art, the present application has the following technical features:

[0079] 1. The present application integrates feed identification, real-time monitoring, model prediction and other modules, maps the historical data and real-time data of the flow control and temperature control system, and realizes the coupling control of the computer system on the flow and temperature through the construction of a nonlinear computer model; the model predictive control method is adopted, the robustness and stability of the system are improved, and the dynamic accurate control of the temperature is realized.

[0080] 2. The present application applies large model technology to compare the collected iron ore property data with the historical data in the database, uses electronic nose and high-definition camera instead of manual identification and manual selection, reduces labor cost and working time, and realizes the automation of iron ore and its batching identification work.

[0081] 3. The present application dynamically updates the weight coefficients of the linear quadratic control law, applies entropy weight to obtain the weight coefficients of the system state and system input; the real-time data in the database are processed in the same way, and the weight coefficients are updated constantly, so that the prediction result is closer to the expected trajectory, and accurate and rapid temperature control is realized. BRIEF DESCRIPTION OF DRAWINGS

[0082] Figure 1 The structural diagram of the system of the present application;

[0083] Figure 2 The model establishment and data analysis flowchart in one embodiment of the method of the present application;

[0084] Figure 3 The data processing flowchart in one embodiment of the method of the present application. DETAILED DESCRIPTION

[0085] Referring to the accompanying Figure 1 , the present application provides a computer system based on a blast furnace-hot blast furnace safety temperature control calculation model, comprising:

[0086] The feedstock identification module includes an electronic nose for identifying the smell of the ore, and a high-definition camera for identifying the shape and color of the ore; when the ore enters the blast furnace, the smell sampling operator in the electronic nose and the gas sensor array identify the smell intensity and smell category of the ore, and convert them into digital signals of the smell, which are transmitted to the processing information database together with the shape and color obtained and identified by the high-definition camera.

[0087] The processing information database contains two sub-databases; the first sub-database pre-stores information entries of the ore, each information entry including the ore type, ore name, shape, color, smell, processing time, and dome temperature, etc.; using the shape, smell, and color collected by the ore identification module as the search conditions, the first sub-database performs information matching on the stored information entries to find the ore type matching the search conditions, and obtains the processing time t g and the dome temperature target value T g0 from the corresponding entry information.

[0088] The second sub-database pre-stores process entries of the ore, each process entry including the ore type and the initial processing parameters matching the ore type, including the initial cold blast main pipe flow Q 01 , initial blast furnace gas branch pipe flow Q 02 , initial net gas main pipe flow Q 03 , initial combustion air main pipe flow Q 04 , and initial steam injection amount Q 05 ; the second sub-database selects the initial processing parameters in the corresponding process entry according to the ore type matched by the first sub-database, and transmits the initial processing parameters, processing time t g , and dome temperature target value T g0 to the real-time monitoring module as the ore processing information.

[0089] The real-time monitoring module contains two sub-monitoring modules for collecting real-time data in the ore processing process; the first sub-monitoring module is used to detect the dome temperature T g , exhaust gas temperature T f , optimal air-fuel ratio A / F, and blast humidity RH in the blast furnace; the second sub-monitoring module is used to detect the cold blast main pipe pressure P1, cold blast main pipe temperature T1, blast furnace gas branch pipe pressure P2, blast furnace gas branch pipe temperature T2, net gas main pipe pressure P3, net gas main pipe temperature T3, combustion air main pipe pressure P4, combustion air main pipe temperature T4, atmospheric humidity rh, cold blast main pipe flow Q1, blast furnace gas branch pipe flow Q2, net gas main pipe flow Q3, combustion air main pipe flow Q4, and steam injection amount Q5, and transmits these real-time data and the ore processing information to the model prediction module.

[0090] The model prediction module predicts and outputs the blast furnace-hot blast furnace next time cold blast main pipe flow, blast furnace gas branch pipe flow, net gas main pipe flow, combustion air main pipe flow, steam injection amount based on the ore processing information and real-time data by using the blast furnace-hot blast furnace safety temperature control calculation model, and the blast furnace-hot blast furnace adjusts and controls according to the parameters to realize temperature control in the processing process.

[0091] The specific construction process of the blast furnace-hot blast furnace safety temperature control calculation model of the model prediction module in the scheme is as follows:

[0092] Step 1, based on the ore processing information and real-time data in the ore processing process, a blast furnace-hot blast furnace safety temperature control calculation model is constructed; the calculation model is a nonlinear model, and the nonlinear model is converted and processed to construct a state space equation.

[0093] This step uses the heating and heat conduction mechanism to identify the system and construct a nonlinear calculation model according to the relationship between the dome temperature and the processing time, and the calculation model is converted and processed to obtain a state space equation; first, system identification is needed; the nonlinear model of the blast furnace-hot blast furnace temperature control is established as follows:

[0094] ;

[0095] ;

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] Wherein, t is a time parameter, P1 is a cold blast main pipe pressure, T1 is a cold blast main pipe temperature, P2 is a blast furnace gas branch pipe pressure, T2 is a blast furnace gas branch pipe temperature, P3 is a net gas main pipe pressure, T3 is a net gas main pipe temperature, P4 is a combustion air main pipe pressure, T4 is a combustion air main pipe temperature, rh is atmospheric humidity; Q1 is a cold blast main pipe flow, Q2 is a blast furnace gas branch pipe flow, Q3 is a net gas main pipe flow, Q4 is a combustion air main pipe flow, Q5 is a steam injection amount; T g is a dome temperature, T f is an exhaust gas temperature, A / F is an optimal air-fuel ratio, and RH is an air supply humidity.

[0103] Parameters in this solution denotes the first derivative, for example is the first derivative of the cold air main pipe pressure; , , , , is a composite function.

[0104] Let the state variable , ; the input variable , ; the output variable , , the superscript T represents the transpose.

[0105] Then the state space equation is constructed from the nonlinear model as follows:

[0106] ;

[0107] ;

[0108] wherein is the first derivative of the state variable, A is the system matrix, B is the control matrix, C is the output matrix or observation matrix, and D is the direct transfer matrix.

[0109] Step 2, for the established state space equation, the method of Taylor series expansion is applied to determine the coefficient matrix of the state space equation, thereby linearizing the state space equation.

[0110] ;

[0111] ;

[0112] ;

[0113] ;

[0114] wherein is the partial derivative, denotes the coefficient of the corresponding element after integrating the derivative, and t is the time parameter.

[0115] Step 3, the state space equation obtained by linearization is discretized using a zero-order holder to construct a discrete system.

[0116] Since step 2 determines the coefficient matrix to realize the continuous of state space equation, in order to simplify the model, improve the stability of the model, reduce the risk of overfitting and improve the calculation efficiency, the linearized state space equation is discretized by using zero-order holder to obtain the discrete system:

[0117] ;

[0118] wherein k is the current time, is the value of state variable x at k time, is the value of input variable u at k time, represents the predicted system state at k+1 time at k time, 、 is the coefficient matrix of discrete system, and has:

[0119] ;

[0120] ;

[0121] wherein I is a unit matrix, is a time constant, is a natural constant, A is a system matrix, B is a control matrix, and T is a sampling period.

[0122] Step 4, for the discrete system, the system state in the future time period is predicted by using a rolling optimization method.

[0123] System state in the future time period to The expression is as follows:

[0124] ; ; ;... ;

[0125] wherein, is the initial system state at k time, and the value is the state variable collected at k time ; is the predicted system state at k+N time at k time; is the input variable at k time ; is the predicted input variable at k+N-1 time at k time, and N is the length of the prediction interval.

[0126] Step 5: Design a quadratic control law. Use the entropy weight method to obtain the weight coefficients of the system state and the input variables in the quadratic control law. Substitute the predicted system state for a future period into the law and set constraints. Solve the sequence of system input variables and extract the first input variable from the solved sequence. The computer system uses the parameters in this input variable to regulate the processing of the blast furnace-hot blast stove.

[0127] First, design the general form of the quadratic control law J:

[0128] ;

[0129] Where Q is The weighting coefficients of the system state at time step R are the weighting coefficients of the input variables, and F is the weighting coefficient of the system state at time step step F. Weighting coefficients of the system state at any given time; The system state at time k+i is predicted after step 4. Let be the input variable for time k+i predicted at time k, where i = 1, 2, ..., N-1, and the superscript T indicates transpose.

[0130] The temperature fluctuation range of the hot blast stove is typically designed to be within ±10℃ to ensure the stability and accuracy of the ironmaking and steelmaking processes. For the above quadratic control law J, considering the safety and service life of the hot blast stove and blast furnace, its constraints are as follows:

[0131] ;

[0132] ;

[0133] For the above quadratic control law J, its The weighting coefficient Q of the system state at time t, The weighting coefficients F of the system state at time step and R of the input variables are obtained using the entropy weighting method, as follows:

[0134] Historical data of the currently processed ore is obtained, and fourteen parameters are extracted from the time intervals k+1 to k+N: cold blast main pipe pressure P1, cold blast main pipe temperature T1, blast furnace gas branch pipe pressure P2, blast furnace gas branch pipe temperature T2, clean gas main pipe pressure P3, clean gas main pipe temperature T3, combustion air main pipe pressure P4, combustion air main pipe temperature T4, atmospheric humidity rh, cold blast main pipe flow rate Q1, blast furnace gas branch pipe flow rate Q2, clean gas main pipe flow rate Q3, combustion air main pipe flow rate Q4, and injected steam quantity Q5. Each data value of each parameter is taken as a sample; the data is then presented in a tabular format with N rows and 14 columns.

[0135]

[0136] wherein the sample a ij is the data value of the (k+i)th moment of the jth parameter historical data; a forward index is established for each sample

[0137]

[0138] wherein, is the forward index of the sample a ij , min and max are minimum and maximum operation.

[0139] Then, the forward index of each sample is calculated in turn The proportion in the forward index at all moments , the entropy value of the forward index corresponding to each kind of historical data

[0140]

[0141]

[0142] wherein, .

[0143] Finally, the weight of the forward index of each kind of historical data is obtained

[0144]

[0145] wherein, The weight coefficient Q of the system state at the moment is a positive definite matrix, and each element on the diagonal is composed of The weight coefficient R is an identity matrix, and the diagonal elements are .

[0146] Then The weight coefficient F of the system state at the moment is:

[0147]

[0148] The first N-1 values in the system state in the future period obtained in step 4, i.e. to are taken as , and the Nth value is taken as , which is substituted into the quadratic form control law J. The derivative of the quadratic form control law J is taken, and the derivative is set to 0. The input variable sequence of the system , i=1…N-1. ​​​​​​​​

[0149] The obtained system input variable is a combined matrix of the input variables u of the system from the kth moment to the k+N-1th moment, and the first input variable in the combined matrix is taken as the input variable of the system at the kth moment The predicted total cold air pipe flow, the blast furnace gas branch pipe flow, the net gas total pipe flow, the combustion air total pipe flow and the steam injection amount are used for controlling the cold air total pipe, the blast furnace gas branch pipe, the net gas total pipe, the combustion air total pipe and the steam injection machine respectively, so as to adjust the ore processing process.

[0150] In the specific application, the working process is as follows:

[0151] At the initial moment k, the computer system controls the medicine calcining machine to start working with the initial processing parameters in the ore processing information, that is, the initial cold air total pipe flow Q 01 , the initial blast furnace gas branch pipe flow Q 02 , the initial net gas total pipe flow Q 03 , the initial combustion air total pipe flow Q 04 , the initial steam injection amount Q 05 , and sets the processing time t g and the dome temperature target value T g0 , and starts the first cycle, and the process of each cycle is as follows:

[0152] The real-time monitoring module obtains the state variable at the kth moment , takes the state variable x as the initial state of the system at the kth moment in the rolling optimization method in step 4 , predicts and obtains the system state at the future moment , then takes as into the quadratic form control law J of step 5, obtains the input variable sequence by deriving J and making the derivative equal to zero , takes the first value of the input variable sequence as the input variable of the system at the next moment k+1 , and controls the blast furnace-heat furnace at the next moment k+1;

[0153] At the next moment k+1, the second cycle starts, and the above cycle process is repeated until the ore processing is completed or the processing time t g and the dome temperature target value T g0 are reached.

[0154] ​​The above examples are only used to illustrate the technical solutions of the present application, but not limit the same; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A computer system based on a blast - hot blast stove safety temperature control calculation model, characterized in that, The application relates to a blast furnace-heat stove safety temperature control system. The application comprises: A feedstock identification module for obtaining smell, shape and color data of the ore after the ore enters the blast furnace; A processing information database comprising two sub-databases; wherein the first sub-database pre-stores information entries of the ore, for realizing matching of processing time and dome temperature of the ore; and the second sub-database pre-stores process entries of the ore, for realizing matching of initial processing parameters of the ore; A real-time monitoring module for collecting real-time data in the ore processing process; A model prediction module for predicting and outputting blast furnace-heat stove next-time cold blast main pipe flow, blast furnace gas branch pipe flow, net blast furnace gas main pipe flow, combustion air main pipe flow and steam injection amount based on the ore processing information and the real-time data by using a blast furnace-heat stove safety temperature control calculation model, so that the blast furnace-heat stove is regulated according to the parameters to realize temperature control in the processing process; The first sub-database pre-stores information entries for ores. Each information entry includes: ore type, ore name, shape, color, odor, processing time, and dome temperature. Using the shape, odor, and color collected by the ore identification module as search criteria, the first sub-database matches the stored information entries to find ore types that match the search criteria and retrieves the processing time from the corresponding entry information. t and the target value of the dome temperature T g0 ; The second sub-database pre-stores process terms of the ore, each of which includes: ore type and initial processing parameters matched with the ore type, including initial value of cold blast main pipe flow Q 01 , initial value of blast furnace gas branch pipe flow Q 02 , initial value of net gas main pipe flow Q 03 , initial value of combustion air main pipe flow Q 04 , initial value of steam injection amount Q 05 ; the second sub-database selects the initial processing parameters in the corresponding process term according to the ore type matched by the first sub-database, and transmits the initial processing parameters, processing time t g , target value of dome temperature T g0 to the real-time monitoring module as ore processing information.

2. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 1, wherein, The monitoring module comprises two sub-monitoring modules, a first sub-monitoring module for detecting the dome temperature in the blast furnace T g , exhaust gas temperature T f , optimal air-fuel ratio The ore identification module comprises an electronic nose for identifying the smell of the ore and a high-definition camera for identifying the shape and color of the ore; when the ore enters the blast furnace, the smell intensity and smell category of the ore are identified by a smell sampler in the electronic nose and a gas sensor array, and are converted into digital signals of the smell, which are transmitted to the processing information database together with the shape and color obtained and identified by the high-definition camera; , blast humidity A / F ; a second sub-monitoring module for detecting the cold blast main pressure P 1. Cold blast main temperature T 1. Blast furnace gas branch pipe pressure P 2. Blast furnace gas branch pipe temperature T 2. Net gas main pressure P 3. Net gas main temperature T 3. Combustion air main pressure P 4. Combustion air main temperature T 4. Atmospheric humidity rh ; cold blast main flow rate Q 1. Blast furnace gas branch pipe flow rate Q 2. Net gas main flow rate Q 3. Combustion air main flow rate Q 4. Steam injection amount Q 5, and transmitting these real-time data and ore processing information to the model prediction module.

3. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 1, wherein, RH The construction process of the blast furnace-heat stove safety temperature control calculation model is as follows: A blast furnace-heat stove safety temperature control calculation model is constructed based on ore processing information and real-time data in the ore processing process; the calculation model is a nonlinear model, and state space equations are constructed by conversion processing of the nonlinear model; A coefficient matrix of the state space equations is determined by applying a Taylor series expansion method, so that the state space equations are linearized; The state space equations obtained by linearization processing are discretized by using a zero-order holder, and a discretized system is constructed; The system state in a future period of time is predicted by using a rolling optimization method for the discretized system; 4. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 3, wherein, A quadratic control law is designed, the weight coefficients of the system state and the weight coefficients of the input variables in the quadratic control law are obtained by using an entropy weight method, the system state in the future period of time obtained by prediction is brought in, and constraint conditions are set, so that the system input variable sequence is solved, the first input variable in the solved sequence is extracted, and the parameter in the input variable is used by a computer system to regulate the processing process of the blast furnace-heat stove. ； ； ； ； ； ； ； ； wherein t t is a time parameter, P 1 is the cold blast main pressure, T 1 is the cold blast main temperature, P 2 is the blast furnace gas branch pipe pressure, T 2 is the blast furnace gas branch pipe temperature, P 3 is the net gas main pressure, T 3 is the net gas main temperature, P 4 is the combustion air main pressure, T 4 is the combustion air main temperature, rh is the atmospheric humidity; Q 1 is the cold blast main flow, Q 2 is the blast furnace gas branch pipe flow, Q 3 is the net gas main flow, Q 4 is the combustion air main flow, Q 5 is the amount of steam injected; T g is the dome temperature, T f is the exhaust gas temperature, A blast furnace-heat stove safety temperature control calculation model is constructed based on ore processing information and real-time data in the ore processing process; the calculation model is a nonlinear model, and state space equations are constructed by conversion processing of the nonlinear model, including: is the optimum air-fuel ratio, A / F is the supply air humidity, the upper index of the parameter denotes the first derivative thereof; Let the state variable where ; input variable where ; output variable where the superscript T denotes the transpose; RH ； ； wherein, is a first derivative of the state variable, A is a system matrix, B is a control matrix, C is an output matrix or observation matrix, D is a direct transfer matrix.

5. The blast - hot blast stove safety temperature control calculation model based computer system as claimed in claim 4 wherein, The state space equations are constructed from the nonlinear model as follows: ； ； ； ； wherein, is the partial derivative, denotes the coefficient of the corresponding element after integrating the derivative, t is the time parameter.

6. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 4, wherein, The Taylor series expansion method is applied to determine the coefficient matrix of the state space equations, and the method is specifically expressed as: The quadratic control law is designed, including: ； wherein Q is the weight coefficient of the system state at time R is the weight coefficient of the input variable at time F is the weight coefficient of the system state at time is k the predicted k + i the system state at time is k the predicted k + i the input variable at time i = 1, 2, …, N -1; Quadratic form control law J The constraint condition is: ； 。 7. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 6, wherein, The general form of the quadratic control law J is designed as follows: The weight coefficients of the system state and the weight coefficients of the input variables in the quadratic control law are obtained by using an entropy weight method, including: Acquire the historical data of the current processed ore, extract the same prediction interval length k +1 to k + N Instant cold air main pressure P 1. Cold air main temperature T 1. Blast furnace gas branch pipe pressure P 2. Blast furnace gas branch pipe temperature T 2. Net gas main pressure P 3. Net gas main temperature T 3. Combustion air main pressure P 4. Combustion air main temperature T 4. Atmospheric humidity rh , Cold air main flow Q 1. Blast furnace gas branch pipe flow Q 2. Net gas main flow Q 3. Combustion air main flow Q 4. Steam injection amount Q 5. These fourteen parameters, each data value of each parameter as a sample; Establishing a forward indicator for each sample : ； wherein, is a sample a ij forward indicator, min, max are min, max operations, a ij denotes the j data value of the k + i time instant of the i-th historical data. Then, the positive indicator of each sample is calculated in turn The proportion in the positive indicator at all times The entropy value of the positive indicator corresponding to each kind of historical data : ； ； wherein ; Finally, the weight of the positive indicator of each historical data is obtained : ； wherein ; the weight coefficient of the system state at the moment Q is a positive definite matrix, each element on its diagonal is composed of ; the weight coefficient R is an identity matrix, whose diagonal elements are ; Then Weighting coefficients of system state at time instant F is: ； wherein is the discrete system coefficient matrix: In the above formula, I is the identity matrix, A is the system matrix, T is the sampling period.

8. The blast furnace - hot blast stove safety temperature control calculation model based computer system according to claim 7, wherein, At the initial time k , the computer system controls the medicine machine to first cool the total pipe flow initial value of the ore processing information Q 01 , the initial value of the branch flow of blast furnace gas Q 02 , the initial value of the total pipe flow of net gas Q 03 , the initial value of the total pipe flow of combustion air Q 04 , the initial value of the amount of steam injected Q 05 Start working and set the processing time t g , the target value of the dome temperature T g0 , start the first cycle, and the process of each cycle is: Real-time monitoring module acquisition k State variables at time 1 , will state variables x As a rolling optimization method k Initial state of the system at time Predict and obtain the system state at future moments. to Then to As Substituting into the quadratic control law J In the middle, through the J Find the derivative and set it to zero to obtain the sequence of input variables. Take its first value For the blast furnace-hot blast stove at the next moment k +1 is used for control; At the next instant k +1, start the second cycle; and repeat the above cycle until the ore processing is completed or at the same time the processing time is reached t g and vault temperature target value T g0 .

Images