A blast furnace gas utilization rate multi-time scale adjusting method and system

Abstract

The present application provides a blast furnace gas utilization rate multi-time scale adjusting method and system, comprising the following steps: S1: obtaining actual blast furnace data, preprocessing the actual blast furnace data, obtaining blast furnace operation parameters and blast furnace gas utilization rate; S2: obtaining the time sequence A of the blast furnace operation parameters and the time sequence B of the blast furnace gas utilization rate, and calculating the dynamic time warping distance matrix of the time sequence A and the time sequence B; S3: obtaining the optimal distance path in the dynamic time warping distance matrix, and improving the prediction and control of the blast furnace index by solving the optimal distance path. The present application can effectively analyze the relationship between the blast furnace operation parameters and the blast furnace gas utilization rate through the dynamic time warping distance matrix, and effectively guide the prediction and control of the blast furnace index through the optimal distance path in the dynamic time warping distance matrix.

Description

Technical Field

[0001] This invention relates to the field of gas utilization technology, and in particular to a method and system for adjusting the utilization rate of blast furnace gas over multiple time scales. Background Technology

[0002] A blast furnace is a large, vertical, and complex modern metallurgical vessel used to produce pig iron. The process involves a continuous production flow where iron ore and other iron-containing compounds react with coke under high temperature and pressure to produce molten iron and blast furnace gas. Blast furnace gas mainly consists of carbon monoxide, carbon dioxide, nitrogen, and hydrogen. The ratio of carbon dioxide content to the total content of carbon monoxide and carbon dioxide in the top blast furnace gas stream is defined as the blast furnace gas utilization rate (GUR). The blast furnace gas utilization rate is a crucial indicator reflecting the overall condition of the blast furnace, and analyzing its relationship with operating parameters is of great significance for reducing energy consumption and increasing blast furnace output.

[0003] Traditional methods for analyzing blast furnace gas utilization mainly focus on the impact of key blast furnace operating parameters, such as charging and blasting operations, on gas utilization at a specific time scale. However, these methods do not consider the multi-time scale characteristics of the blast furnace production process. This leads to inaccurate analysis of the relationship between blast furnace operation and blast furnace gas utilization, making it difficult to formulate precise blast furnace production operation strategies to effectively control blast furnace gas utilization.

[0004] Existing methods for analyzing blast furnace gas utilization rates often rely on H2O, as the blast furnace is a typical "black box" system with numerous complex interferences during actual operation. ∞ Noise is removed using filters; then, correlation analysis is used on a single time scale to analyze the relationship between blast furnace operating parameters and gas utilization. However, due to the nonlinear, large time-delay, and multi-time-scale characteristics of the blast furnace ironmaking process, there is strong coupling between blast furnace operating parameters. Therefore, these methods yield poor correlation results, making it difficult to effectively determine the specific factors affecting gas utilization.

[0005] Many studies have been conducted by scholars both domestically and internationally on issues such as modeling of blast furnace charging and blasting processes, modeling and control of blast furnace gas utilization, and multi-timescale analysis, achieving certain results. However, the multi-timescale characteristics of blast furnace production processes have not been considered, and effective blast furnace system analysis methods based on these characteristics have not been developed. Current research on the relationship between blast furnace operation and blast furnace gas utilization (GUR) mostly focuses on establishing separate models of the relationship between charging and blasting operations and the blast furnace GUR state. These analyses are essentially performed on the same timescale, neglecting the different timescale characteristics of the effects of charging and blasting operations on blast furnace GUR. Therefore, an effective multi-timescale analysis method for the impact of charging and blasting operations on blast furnace GUR is lacking. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a method for adjusting blast furnace gas utilization rate across multiple time scales, comprising the following steps:

[0007] S1: Obtain actual blast furnace data, preprocess the actual blast furnace data, and obtain blast furnace operating parameters and blast furnace gas utilization rate;

[0008] S2: Obtain the time series A of blast furnace operating parameters and the time series B of blast furnace gas utilization rate, and calculate the dynamic time warp distance matrix between time series A and time series B;

[0009] S3: Obtain the optimal distance path in the dynamic time-warped distance matrix, and improve the prediction and control of blast furnace indicators by solving for the optimal distance path.

[0010] Preferred:

[0011] The time series A of blast furnace operating parameters and the time series B of blast furnace gas utilization rate are respectively expressed as: A={A1,A2,...,A... i ,...,A p} and B = {B1, B2, ..., B j ,...,B q};

[0012] Among them, A i Let B be the i-th sample point of the blast furnace operating parameters, p be the total number of sample points of the blast furnace operating parameters, and B be the ith sample point. j Let A be the j-th sample point of blast furnace gas utilization rate, q be the total number of blast furnace gas utilization rate sample points, and A be the total number of sample points. i and B j All are m-dimensional variables.

[0013] Preferably, the calculation process of the dynamic time-warped distance matrix is ​​as follows:

[0014] S21: Obtain a P×q Euclidean distance matrix by calculating the Euclidean distance between time series A and time series B. The distance of the current point for element (i,j) in the Euclidean distance matrix is ​​d. E (A i B j );

[0015] S22: Define the cumulative distance of element (i,j) as d DTW (i,j), let d DTW (1,1)=d E (A1,B1);

[0016] S23: Let i = 2, 3, ..., p, and calculate p-1 cumulative distances d in sequence. DTW (i,1);

[0017] S24: Let j=2, 3,..., q, sequentially calculate q-1 cumulative distances d DTW (i, j);

[0018] S25: Let i=2, 3,..., p, j=2, 3,..., q, sequentially calculate (p-1)*(q-1) cumulative distances d DTW (i, j);

[0019] S26: Construct a dynamic time warping distance matrix of blast furnace operation parameters and blast furnace gas utilization rate by p*q cumulative distances, the dynamic time warping distance matrix is a large matrix composed of p*q small rectangles, each small rectangle corresponds to a cumulative distance, wherein d DTW (1, 1) is located at the upper left corner, d DTW (p, q) is located at the lower right corner.

[0020] Preferably, the cumulative distance d DTW (i, 1) of the first sample point of time series A to time series B is calculated according to the formula: d DTW (i, 1) = d DTW (i-1, 1) + d E (A i , B1).

[0021] Preferably, the cumulative distance d DTW (1, j) of the first sample point of time series A to time series B is calculated according to the formula: d DTW (1, j) = d DTW (1, j-1) + d E (A1, B j ).

[0022] Preferably, the cumulative distance d DTW (i, j) of the (p-1)*(q-1) sample points of time series A to time series B is calculated according to the formula:

[0023] d DTW (i, j) = d E (A i , B j ) + min{d DTW (i, j-1), d DTW (i-1, j), d DTW (i-1, j-1)}.

[0024] Preferably, the process of obtaining the optimal distance path is as follows:

[0025] Starting from d DTW (1, 1) in the dynamic time warping distance matrix, sequentially connect to d DTW(p,q), the small rectangle d DTW (i,j) right, lower edge and the right lower edge of the three small rectangles corresponding to the minimum cumulative distance as d DTW (i,j) connection point, the optimal distance path of blast furnace operating parameters and blast furnace gas utilization rate is obtained by solving.

[0026] A blast furnace gas utilization rate multi-time scale adjustment system, comprising the following modules:

[0027] A data acquisition module is used for acquiring actual blast furnace data, pre-processing the actual blast furnace data, obtaining blast furnace operating parameters and blast furnace gas utilization rate;

[0028] A dynamic time warping distance matrix acquisition module is used for acquiring the time sequence A of blast furnace operating parameters and the time sequence B of blast furnace gas utilization rate, and calculating the dynamic time warping distance matrix of the time sequence A and the time sequence B.

[0029] An optimal distance path acquisition module is used for acquiring the optimal distance path in the dynamic time warping distance matrix, and improving the prediction and control of blast furnace indexes by solving the optimal distance path.

[0030] The present application has the following beneficial effects:

[0031] The present application constructs a dynamic time warping distance matrix by performing multi-time scale analysis on the time sequence of blast furnace operating parameters and the time sequence of blast furnace gas utilization rate, can effectively analyze the relationship between blast furnace operating parameters and blast furnace gas utilization rate through the dynamic time warping distance matrix, and effectively guides the prediction and control of blast furnace indexes through the optimal distance path in the dynamic time warping distance matrix, has important application value in stabilizing blast furnace production, improving molten iron quality and yield, prolonging blast furnace life and energy saving and emission reduction, etc. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 The present application is an embodiment method flow chart;

[0033] Figure 2 The present application is a blast furnace production process schematic diagram;

[0034] Figure 3 The present application is a blast furnace operating parameter and blast furnace gas utilization rate (GUR) time sequence;

[0035] Figure 4 The present application is a center focal ratio and GUR time sequence;

[0036] Figure 5 The present application is a center focal ratio and GUR dynamic time warping distance;

[0037] Figure 6The optimal distance path of the current time center focal ratio and GUR;

[0038] Figure 7 The optimal distance path of the center focal ratio and GUR delayed for 6 hours;

[0039] Figure 8 The time series of the air volume and GUR;

[0040] Figure 9 The dynamic time warping distance of the air volume and GUR;

[0041] Figure 10 The optimal distance path of the current time air volume and GUR;

[0042] Figure 11 The influence of cloth operation parameters on GUR;

[0043] Figure 12 The influence of air supply operation parameters on GUR;

[0044] The implementation of the object of the present application, functional features and advantages will be further illustrated with reference to the accompanying drawings. DETAILED DESCRIPTION

[0045] It should be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application.

[0046] Referring to Figure 1 , the present application provides a blast furnace gas utilization rate multi-time scale adjustment method, comprising the following steps:

[0047] S1: obtaining actual blast furnace data, pre-processing the actual blast furnace data to obtain blast furnace operation parameters and blast furnace gas utilization rate;

[0048] S2: obtaining the time series A of the blast furnace operation parameters and the time series B of the blast furnace gas utilization rate, and calculating the dynamic time warping distance matrix of the time series A and the time series B;

[0049] S3: obtaining the optimal distance path in the dynamic time warping distance matrix, and improving the prediction and control of blast furnace indexes by solving the optimal distance path.

[0050] Further, the blast furnace production process flow is as shown in Figure 2 .

[0051] The blast furnace production process is a continuous production process in which iron ore and other iron-containing compounds and coke are subjected to oxidation-reduction reactions at high temperature and high pressure to produce molten iron and blast furnace gas. In the production process, the ore and coke are intermittently distributed into the interior of the furnace body in a set proportion through the top distribution system, forming a solid material layer in which the coke and ore are longitudinally alternately distributed; hot air containing oxygen is blown into the blast furnace from the tuyere at the bottom of the blast furnace, and oil, coal powder or natural gas and other fuels are injected. A series of oxidation-reduction reactions occur between the descending charge in the blast furnace and the ascending gas stream, and finally the blast furnace gas is discharged from the top of the furnace and the iron oxide in the iron ore is reduced to molten iron, which flows out of the iron runner, and the slag floating on the molten iron is discharged through the slag runner.

[0052] The blast furnace operation is changed by the distribution operation and the blast operation, which changes the charge distribution, the size of the combustion zone and the blast kinetic energy, and affects the blast furnace gas utilization rate (GUR). The upper distribution mainly adjusts the ore and coke sequence, the material weight, the distribution position, the material line depth and other parameters, changes the state of the blast column, the center coke ratio and the edge coke ratio, the raw material distribution structure in the blast furnace, thereby adjusting the distribution of the charge layer and the gas flow in the blast furnace, affecting the development of the gas flow, and changing the degree of chemical reaction in the blast furnace, thereby controlling the blast furnace gas utilization rate and operation. According to the actual industrial data of a certain steel plant, it takes 6-7 hours for the charge to be completely reduced to liquid pig iron from the top to the bottom of the blast furnace, and it takes longer to replace the charge layer, so the distribution operation is a long-term operation. The blast operation sends the heated gas from the hot blast stove into the blast furnace through different tuyeres, forming an approximately elliptical tuyere rotation zone in the lower part of the blast furnace. The shape and size of the tuyere are determined by the blast kinetic energy of the blast furnace, which reflects the inlet state of the tuyere, and affects the distribution of the gas flow and temperature, and also affects the degree of activity of the hearth. The blast kinetic energy refers to the energy of the blast at a tuyere of the blast furnace per unit time, which indicates the resistance of the blast to the material layer above the tuyere. The calculation formula is shown in formula (1):

[0053]

[0054] where E is the blast kinetic energy; R is a constant; is the oxygen enrichment rate; S is the tuyere area; V b is the air volume.

[0055] The blast operation changes the air volume of the blast port, increases the blast kinetic energy, changes the distribution of the air flow, reduces the edge air flow, enhances the center air flow, and relatively increases the smelting intensity. With the increase of the smelting intensity, the reaction speed in the hearth is accelerated, and the air flow is rapidly developed, thereby affecting the equilibrium coefficient of the chemical reaction and the gas utilization rate. According to the actual industrial data, the air flow in the blast furnace only needs a few seconds from the generation of the air flow through the charge gap to the final discharge from the top of the blast furnace, so the blast operation is a short-time operation. Through the analysis of the smelting mechanism of the blast furnace production process, it is qualitatively proposed that the influence of the blast furnace operation parameters on the GUR has the characteristics of multiple time scales.

[0056] In view of the change characteristics of the operation parameters and the GUR time series in the blast furnace production process, the similarity of the blast furnace operation parameters and the GUR in two different time series is measured by combining the dynamic time warping (DTW), the optimal distance of the blast furnace operation parameters and the GUR sequence is calculated, the most matched sample point in the blast furnace operation parameter sequence corresponding to each sample point in the GUR sequence is found, and the multiple time scale division of the influence of the blast furnace operation parameters on the GUR is realized in different time scales.

[0057] The time series A of the blast furnace operation parameters and the time series B of the blast furnace gas utilization rate are respectively represented as: A = {A1, A2,..., A i ,...,A p} and B = {B1, B2,..., B j ,...,B q};

[0058] Wherein, A i is the i-th sample point of the blast furnace operation parameter, p is the total number of sample points of the blast furnace operation parameter, B j is the j-th sample point of the blast furnace gas utilization rate, q is the total number of sample points of the blast furnace gas utilization rate, A i and B j are m-dimensional variables.

[0059] Further, the traditional Euclidean distance adopts a "one-to-one" time series alignment method, but cannot flexibly capture the similarity of two similar trajectories. The dynamic time warping distance can successfully match the wave peaks and troughs compared with the Euclidean distance. The dynamic time warping distance is an effective method for calculating the similarity between time series by lengthening or shortening the time series, and is suitable for calculating the similarity degree and optimal distance of time series with different lengths of blast furnace unordered data. Given two time series of blast furnace data A = {A1, A2,..., A i ,...,A p} and B = {B1, B2,..., B j,...,B q Dynamic time warping distance aligns two time series to minimize their differences, and can be used as a method to calculate the delay time of the impact of blast furnace operating parameters on GUR.

[0060] The calculation process of the dynamic time-warped distance matrix is ​​as follows:

[0061] S21: Obtain a P×q Euclidean distance matrix by calculating the Euclidean distance between time series A and time series B. The distance of the current point for element (i,j) in the Euclidean distance matrix is ​​d. E (A i B j );

[0062] Specifically, for two blast furnace data time series A = {A1, A2, ..., A...} n} and B = {B1, B2, ..., B n}, A i ={a i1 ,a i2 ,...,a im}, B i ={b i1 ,b i2 ,...,b im Their Euclidean distance is:

[0063]

[0064] in

[0065] like Figure 3 As shown, (a) is the Euclidean distance between the two time series, and (b) is the dynamic time warping distance between the two time series.

[0066] S22: Define the cumulative distance of element (i,j) as d DTW (i,j), let d DTW (1,1)=d E (A1,B1);

[0067] S23: Let i = 2, 3, ..., p, and calculate p-1 cumulative distances d in sequence. DTW (i,1);

[0068] S24: Let j = 2, 3, ..., q, and calculate q-1 cumulative distances d in sequence. DTW (1,j);

[0069] S25: Let i = 2, 3, ..., p, j = 2, 3, ..., q, and calculate (p-1)*(q-1) cumulative distances d in sequence.DTW (i,j);

[0070] S26: Construct a dynamic time warping distance matrix of the blast furnace operation parameters and the blast furnace gas utilization rate by the p*q cumulative distances, the dynamic time warping distance matrix is a large matrix composed of p*q small rectangles, each small rectangle corresponds to a cumulative distance, wherein d DTW (1,1) is located at the upper left corner, d DTW (p,q) is located at the lower right corner.

[0071] Further, the cumulative distance d DTW (i,1) of the time series A to the first sample point of the time series B is calculated according to the formula: d DTW (i,1) = d DTW (i-1,1) + d E (A i ,B1).

[0072] Further, the cumulative distance d DTW (1,j) of the first sample point of the time series A to the time series B is calculated according to the formula: d DTW (1,j) = d DTW (1,j-1) + d E (A1,B j ).

[0073] Further, the cumulative distance d DTW (i,j) of the (p-1)*(q-1) sample points of the time series A and the time series B is calculated according to the formula:

[0074] d DTW (i,j) = d E (A i ,B j ) + min{d DTW (i,j-1), d DTW (i-1,j), d DTW (i-1,j-1)}.

[0075] Further, the process of obtaining the optimal distance path is as follows:

[0076] Starting from d DTW (1,1) in the dynamic time warping distance matrix, sequentially connecting to d DTW (p,q), taking the small rectangle with the minimum corresponding cumulative distance in the three small rectangles right to, below and right below d DTW (i,j) as the connection point of d DTW (i,j), and solving to obtain the optimal distance path of the relationship between the blast furnace operation parameters and the blast furnace gas utilization rate.

[0077] Experimental results verification and analysis:

[0078] The time lag characteristics experiment between the blast furnace operating parameters and GUR was carried out by using the field data of a 2800m3 blast furnace. 100 sample points were selected, the center coke ratio sampling interval was 20min, the air volume sampling interval was 1min, and the center coke ratio and GUR curves after denoising were as shown in Figure 4 The field noise was removed by filtering.

[0079] As shown in Figure 4 , it was the GUR sequence and the center coke ratio sequence sample data. The dynamic time warping distance of the two time series was calculated to measure the time corresponding to the most similar sample points in the two time series, and the time lag of the influence of the blast furnace burden operation on GUR was calculated.

[0080] As shown in Figure 5 , the dynamic time warping distance of the blast furnace center coke ratio and GUR was calculated, and the distance of the two most similar sample points in the time series of the two blast furnace variables was calculated.

[0081] As shown in Figure 6 , it was the dynamic time warping distance heat map of the current time GUR and the center coke ratio. The most similar center coke ratio sequence of the GUR sequence lagged behind about 20 sample points, and the sampling interval was 20min, so the delay was about 6 hours. It was explained that it took about 6 hours to reflect the change of burden operation through GUR, which had a long time scale characteristic, which was consistent with the mechanism qualitative analysis result.

[0082] The dynamic time warping distance of the GUR after 6 hours and the current center coke ratio was calculated.

[0083] As shown in Figure 7 , it was the dynamic time warping distance heat map of the GUR after 6 hours delay and the current center coke ratio. The dynamic time warping distance of the GUR after 6 hours delay and the current center coke ratio time series was calculated, and the optimal solution presented approximately along the diagonal line, which proved that the dynamic time warping distance of the GUR after 6 hours delay and the current center coke ratio was the closest, and further verified that the influence of burden operation on GUR was a long time scale.

[0084] The curve of the blast furnace air volume and GUR after data denoising was as shown in Figure 8 The field noise was removed by filtering.

[0085] As shown in Figure 8As shown, the dynamic time warping distance of the two time series at the current time is calculated for 100 sample data of GUR sequence and air volume sequence, the most similar sample points in the two time series of GUR and air volume are measured, and the time difference corresponding to each sample point in the two time series is the delay time. After the change of air volume, the delay time of GUR change is calculated, and the time lag of the influence of blast furnace air supply operation on GUR is calculated.

[0086] As shown in Figure 9 , the dynamic time warping distance of blast furnace air volume and GUR is calculated, and the distance of the two most similar sample points in the time series of the two blast furnace variables is calculated.

[0087] As shown in Figure 10 , the dynamic time warping distance of GUR and air volume after 1 minute delay is shown in the heat map. The optimal dynamic time warping distance of GUR and air volume presents a diagonal line, which shows that the air supply operation has a quick effect on GUR. The most similar air volume sequence of GUR sequence lags about 1 sample point, so the delay is about 1 min, which means that GUR will change after about 1 min or even a very short time after the air supply operation changes, which has a short time scale characteristic, which is consistent with the mechanism qualitative analysis result.

[0088] Based on the dynamic time warping distance method, the similarity distance between the blast furnace operation burden air supply and GUR is calculated, and the time characteristic relationship between the blast furnace operation parameters and GUR is established. The optimal distance of the blast furnace operation parameters affecting GUR is calculated to divide the time delay scale interval. Combined with mechanism analysis, it is proved that the influence of blast furnace operation parameters on GUR has multi-time scale characteristics, and the correctness of the proposed method is verified.

[0089] Using the data of a 2800m3 blast furnace steelmaking plant, the multi-time scale characteristics of different blast furnace operations affecting GUR are verified by case matching method through real blast furnace production process data.

[0090] As shown in Figure 11 , the center focus ratio sequence in the burden operation is matched with the sequence of GUR delayed for 6 hours. As can be seen in the figure, the GUR delayed for 6 hours has the highest similarity with the center focus ratio at the current time, so it is further verified that the center focus ratio changes will affect the development trend of GUR after 6 hours, which corresponds to the results of mechanism analysis and DTW, proving the effectiveness of dynamic time warping distance calculation; the relationship between air volume sequence in air supply operation and GUR is shown in Figure 12 , the change of GUR delayed for 1 minute is similar to the change of air volume, so the change of air volume will affect the development trend of GUR after 1 minute, which corresponds to the results of mechanism analysis and DTW, and the air supply operation affects GUR in a short time.

[0091] The present application combines mechanism analysis and dynamic time warping distance analysis, and obtains different time scales and lag time of the influence of blast furnace operation on GUR through real blast furnace production process data of a steel plant. On this basis, the multi-time scale characteristics of the influence of the blast furnace operation and GUR time series case matching probability are verified.

[0092] The present application provides a blast furnace gas utilization rate multi-time scale adjustment system, comprising the following modules:

[0093] A data acquisition module is used to acquire actual blast furnace data, pre-process the actual blast furnace data, obtain blast furnace operation parameters and blast furnace gas utilization rate;

[0094] A dynamic time warping distance matrix acquisition module is used to acquire time series A of blast furnace operation parameters and time series B of blast furnace gas utilization rate, and calculate the dynamic time warping distance matrix of time series A and time series B;

[0095] An optimal distance path acquisition module is used to acquire the optimal distance path in the dynamic time warping distance matrix, and improve the prediction and control of blast furnace indexes by solving the optimal distance path.

[0096] It should be noted that in this paper, the term "including", "containing" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or system. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or system including the element.

[0097] The above-mentioned embodiment serial numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments. In the unit claims of several devices, several of these devices can be embodied by the same hardware item. The use of the words first, second, and third does not represent any order, and these words can be interpreted as identifiers.

[0098] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation using the content of the present application specification and drawings, or direct or indirect application in other related technical fields, is also included in the patent protection scope of the present application.

Claims

1. A blast furnace gas utilization rate multi-time scale adjustment method, characterized in that, The method comprises the following steps: S1: obtaining actual blast furnace data, preprocessing the actual blast furnace data, and obtaining blast furnace operation parameters and blast furnace gas utilization rate; S2: Obtain time series of blast furnace operating parameters A and blast furnace gas utilization B , calculate the dynamic time warping distance matrix of the time series A and the time series B of blast furnace gas utilization S3: obtaining the optimal distance path in the dynamic time warping distance matrix, and improving the prediction of the blast furnace index and the time series of the blast furnace operation parameters by solving the optimal distance path A and the time series of the blast furnace gas utilization rate B respectively represented as: and ; wherein, A i is the number of sample points for the blast furnace operating parameter, i p is the total number of sample points for the blast furnace operating parameter, B j is the number of sample points for the blast furnace gas utilization rate, j q is the total number of sample points for the blast furnace gas utilization rate, A i and B j are each m dimensional variables;​​ The calculation process of the dynamic time warping distance matrix is as follows: S21: By calculating time series A With time series B The Euclidean distance is obtained as The Euclidean distance matrix, the elements in the Euclidean distance matrix The current point distance is ; S22: define the accumulated distance of elements as , let ; S23: Let , sequentially calculate the obtained p -1 cumulative distance ; S24: Let , sequentially calculate the obtained q -1 cumulative distance ; S25: Let , , sequentially calculate the obtained p -1)*( q -1) cumulative distances ; S26: passing through p q a dynamic time warping distance matrix of the blast furnace operation parameters and the blast furnace gas utilization rate is constructed by using the cumulative distance, the dynamic time warping distance matrix is p q a large matrix composed of a plurality of small rectangles, each small rectangle corresponds to a cumulative distance, wherein is located at the upper left corner, is located at the lower right corner;​​ The obtaining process of the optimal distance path is as follows: From the dynamic time warping distance matrix , sequentially connect to , the small rectangle The small rectangle with the smallest cumulative distance in the right, lower and right lower edges is connected to The optimal distance path of the relationship between the blast furnace operation parameters and the blast furnace gas utilization rate is obtained.

2. The blast furnace gas utilization multi-time scale regulation method according to claim 1, characterized in that, time series A For the first sample point of the time series B The cumulative distance The formula for calculating the cumulative distance is 3. The blast furnace gas utilization multi-time scale regulation method according to claim 1, characterized in that, first sample point of the time series A cumulative distance of the time series B The calculation formula of the cumulative distance of the time series is as follows: .

4. The blast furnace gas utilization multi-time scale regulation method according to claim 1, characterized in that, time series A with time series B of( p -1)*( q -1) sample points The formula for calculating the cumulative distance of the time series 。 5. A blast furnace gas utilization rate multi-time scale regulating system, characterized in that, The method comprises the following modules: A data acquisition module is configured to obtain actual blast furnace data, preprocess the actual blast furnace data, and obtain blast furnace operation parameters and blast furnace gas utilization rate; a dynamic time warping distance matrix obtaining module, configured to obtain the time sequence of the blast furnace operation parameters A and the time sequence of the blast furnace gas utilization rate B , and calculate the dynamic time warping distance matrix of the time sequence A and the time sequence B ​ An optimal distance path acquisition module is configured to obtain an optimal distance path in the dynamic time warping distance matrix, and improve blast furnace index prediction and control by solving the optimal distance path; Time series of blast furnace operating parameters A and blast furnace gas utilization B are represented as: and ; wherein, A i is the number of sample points for the blast furnace operating parameter, i p is the total number of sample points for the blast furnace operating parameter, B j is the number of sample points for the blast furnace gas utilization rate, j q is the total number of sample points for the blast furnace gas utilization rate, A i and B j are each m dimensional variables;​​ The calculation process of the dynamic time warping distance matrix is as follows: S21: By calculating time series A With time series B The Euclidean distance is obtained as The Euclidean distance matrix, the elements in the Euclidean distance matrix The current point distance is ; S22: define the accumulated distance of elements as , let ; S23: Let , sequentially calculate the obtained p -1 cumulative distance ; S24: Let , sequentially calculate the obtained q -1 cumulative distance ; S25: Let , , sequentially calculate the obtained p -1)*( q -1) cumulative distances ; S26: constructing a dynamic time warping distance matrix of the blast furnace operation parameters and the blast furnace gas utilization rate by p q accumulative distances, the dynamic time warping distance matrix being p q a large matrix composed of a plurality of small rectangles, each small rectangle corresponding to an accumulative distance, wherein located at the upper left corner, located at the lower right corner;​​ The obtaining process of the optimal distance path is as follows: from the dynamic time warping distance matrix , sequentially connected to , small rectangle The smallest small rectangle with the minimum cumulative distance in the three small rectangles on the right, the lower edge and the right lower edge is connected as The optimal distance path of the relationship between the blast furnace operation parameters and the blast furnace gas utilization rate is obtained.

Images