TW201319260A - Method for predicting channeling phenomenon of blast furnace - Google Patents

Abstract

A method for predicting a channeling phenomenon of a blast furnace is disclosed. The method for predicting a channeling phenomenon of the blast furnace includes a data analysis phase and a predicting phase. In the pre-analyzing phase, at first, a plurality of historical gate-position data of an entirely opened dust-cleaning device disposed in the top of the blast furnace is provided. Thereafter, a plurality of historical pressure data of the blast furnace is provided. The historical pressure data correspond to the historical gate-position data in a one to one manner and each of the historical pressure data includes a plurality of historical pressure values of a plurality of layers of the blast furnace. Then, indicators for the channeling phenomenon are calculated in accordance with the historical pressure values and an indicator equation. Thereafter, a warning layer of the blast furnace is determined in accordance with the historical pressure values and the indicators for the channeling phenomenon. In the predicting phase, the channeling phenomenon is predicted in accordance with the pressure variation value of the warning layer of the blast furnace and the indicator corresponding to the waning layer.

Description

Prediction method of blast furnace pipeline flow phenomenon

The present disclosure relates to a method for predicting a blast furnace flow phenomenon.

In a general blast furnace ironmaking operation, raw materials such as iron ore, coke or lime are deposited in layers in a blast furnace to cause it to pass through a redox reaction to form molten iron. In order to allow the hot air of the tuyere portion to stably flow into the center portion of the furnace, the material in the center portion of the blast furnace is smaller than the periphery of the center portion, and the raw material in the furnace can be inclined at a certain angle toward the center portion of the furnace.

The shape, distribution and particle size of the raw materials will affect the gas flow in the furnace, which will affect the furnace condition of the blast furnace and the production of molten iron. However, when the blast furnace is subjected to a redox reaction at a high temperature, the arrangement of the internal raw materials is often unable to uniformly flow the gas in the furnace as expected, and may even cause a channeling phenomenon (Channeling Phenomenon). The generation of the pipe flow will cause the reaction in the blast furnace to be uneven and damage the blast furnace wall, which will result in a decrease in the output of the molten iron and a decrease in the life of the blast furnace.

In order to avoid the negative impact of pipeline flow, blast furnace operators have developed various techniques to predict the occurrence of pipeline flow, in order to carry out some preventive measures in advance, such as reducing the amount of air blasting. However, current pipeline flow prediction technology is still unable to make immediate predictions on-line, and the accuracy rate is only 40-60%. Therefore, a new prediction method for blast furnace pipeline flow phenomenon is needed.

One aspect of the present invention is a method of predicting the flow of a blast furnace pipeline for the blast furnace operator to predict the occurrence of pipeline flow and to take appropriate measures to avoid damage to the pipeline flow.

According to an embodiment of the invention, the blast furnace comprises a plurality of pressure detectors, the pressure detectors being located in a plurality of blast furnace layers having different heights. The prediction method of the blast furnace pipeline flow phenomenon includes a data analysis phase and a pipeline flow prediction phase.

In the data analysis stage, firstly, the full history time of the historical valve of the top of the stove is provided, wherein the historical time of the full opening of the valve corresponds to the full opening time of the plurality of historical valves, and the blast furnace is at the full opening time of the historical valves. Pipe flow phenomenon occurs at the point. Next, a plurality of historical blast furnace pressure data are provided, wherein the historical blast furnace pressure data is one-to-one corresponding to the historical valve full-open time point data, and each historical blast furnace pressure data includes a plurality of historical blast furnace pressure values of the blast furnace layer. Then, according to the historical blast furnace pressure value of each historical blast furnace pressure data and the pipeline flow index equation, a plurality of pipeline flow indexes (RMSD) corresponding to each historical valve full opening time point are calculated, wherein the pipeline flow indexes are one-to-one correspondence To the blast furnace floor, and the pipe flow index equation is as follows:

Where x _i is the i-th historical blast furnace pressure value, x _i +1 is the i+1th historical blast furnace pressure value, and the pressure change value of the at least one high pressure change layer is greater than the pressure change threshold. Then, according to the historical blast furnace pressure value and the pipeline flow index of each historical blast furnace pressure data, at least one blast furnace warning layer is determined.

Then, the pipeline flow prediction phase is performed. In the pipeline flow prediction stage, the current blast furnace pressure data is first provided, wherein the current blast furnace pressure data includes the pressure change value of the blast furnace warning layer. Then, according to the pressure change value of the blast furnace warning layer and the corresponding pipeline flow index, it is predicted whether the pipeline flow phenomenon occurs. When it is judged that the pipe flow phenomenon is about to occur, a warning message is issued to warn the blast furnace operator that the pipe flow may occur.

It can be seen from the above description that the method for predicting the flow phenomenon of the blast furnace pipeline in the embodiment of the present invention first uses the pipeline flow index and the historical pressure data of the blast furnace to find the warning layer of the blast furnace, and in the subsequent prediction stage, the pressure of the warning layer is utilized. Change values and pipe flow indicators to predict whether a pipe flow has occurred. It has been experimentally proved that the prediction method of the blast furnace pipeline flow phenomenon in the embodiment of the present invention has a more accurate prediction rate than the prior art, and the generation of the pipeline flow can be predicted 3 to 5 minutes earlier.

Please refer to FIG. 1 , which is a schematic structural view of a blast furnace smelting system 100 according to an embodiment of the present invention. The blast furnace smelting system 100 includes a blast furnace body 110, a pressure detecting device 120, and a dust washer 130. The blast furnace body 110 has a plurality of blast furnace layers 112 to be monitored, and the pressure detecting devices 120 are disposed one-to-one in the blast furnace layer 112, such that each blast furnace layer 112 has a pressure detecting device 120 to detect The pressure of the blast furnace layer 112 is measured. In the present embodiment, each blast furnace layer 112 has only one pressure detecting device 120 for detecting the pressure of the blast furnace layer 112, and these pressure devices 120 are disposed in the same orientation, such as the north side of the blast furnace body 110. However, in other embodiments of the present invention, each blast furnace layer 112 may be provided with a plurality of pressure detecting devices 120, for example, a pressure detecting device is disposed on the north, south, east, and west sides of the furnace body of each blast furnace layer 112. 120, so each blast furnace layer 112 will have four pressure detecting devices 120 to provide a complete pressure sensing mechanism.

The dust washer 130 is connected to the top of the blast furnace body 110 to receive the exhaust gas discharged during the smelting of the blast furnace body 110 and to purify the exhaust gas. The dust washer 130 has a control valve 132 for regulating the flow of exhaust gas into the dust washer 130. This control valve 132 may also be referred to as an annular slot element (Annular Gap Element; AGE). In general, the degree of opening of the control valve 132 (hereinafter simply referred to as the opening degree) is automatically adjusted according to the amount of blast furnace exhaust gas. For example, when the amount of exhaust gas discharged from the blast furnace increases, these exhaust gases may impact the control valve 132 to increase the opening of the control valve 132. For another example, when the amount of exhaust gas discharged from the blast furnace is reduced, the control valve 132 automatically reduces the opening to match the current blast furnace displacement.

Please refer to FIG. 2, which is a flow chart showing a method 200 for predicting a blast furnace pipeline flow phenomenon according to an embodiment of the present invention. The method for predicting the phenomenon of blast furnace pipeline flow in the embodiment of the present invention can be applied to a computer system to automatically monitor the condition of the blast furnace by using a computer system, and issue a warning signal to notify the operator of the blast furnace before the pipeline flow occurs, so that the blast furnace operator Take appropriate action early.

The prediction method 200 of an embodiment of the present invention includes an analysis phase 210 and a prediction phase 220. The analysis stage 210 is for analyzing a plurality of blast furnace historical data to find a warning layer of the pipeline flow from the blast furnace layer 210, and the prediction stage 220 uses the warning layer and the pipeline flow index provided by the embodiment of the present invention to predict the pipeline. Whether the flow is about to happen.

In the analysis phase 210, a valve data providing step 212 is first performed to provide a full history time point data for the plurality of historical valves of the dust cleaner 130. As described above, the control valve 132 of the dust washer 130 adjusts the opening according to the amount of exhaust gas discharged from the blast furnace. Therefore, when the pipe flow phenomenon occurs, the opening degree of the control valve 132 should be 100%, and the historical valve full opening time point data corresponds to the occurrence of the pipe flow phenomenon.

Next, a historical blast furnace pressure data providing step 214 is performed to provide a plurality of historical blast furnace pressure data. These historical blast furnace pressure data are one-to-one correspondence to the historical valve full-open time point data, which means that each blast furnace pressure data represents the blast furnace pressure state when the blast furnace pipeline flow occurs. Each historical blast furnace pressure data contains a plurality of blast furnace pressure values and a plurality of historical blast furnace pressure changes.

The historical blast furnace pressure value represents the pressure value of each blast furnace layer 112 when the blast furnace piping flow occurs, and the historical blast furnace pressure change value represents the pressure change value of each blast furnace layer 112 when the blast furnace piping flow occurs. For example, the blast furnace body 110 of the present embodiment includes eight blast furnace layers 112, and each blast furnace layer 112 has a pressure detecting device 120. Therefore, in the present embodiment, each blast furnace pressure data includes eight blast furnace pressures. Value and 8 pressure change values. However, in other embodiments of the present invention, if the number of blast furnace layers 112 is maintained at 8 but each layer has 4 pressure detecting devices 120, each blast furnace pressure data contains 8 x 4 blast furnace pressure values and 8 x 4 pressure changes. value.

Then, a pipeline flow index (RMSD) calculation step 216 is performed to calculate a pipeline flow index corresponding to each historical valve full-open time point data according to the historical blast furnace pressure value of each historical blast furnace pressure data and the pipeline flow index equation. The pipe flow index equations are expressed as follows:

Where x _i is the i-th historical blast furnace pressure value, x _i +1 is the i+1th historical blast furnace pressure value, and the height pressure change referred to herein means that the pressure change value is greater than the preset pressure change threshold. In the present embodiment, the pressure change threshold value is 0.15 kg/cm 2 , but the embodiment of the present invention is not limited thereto.

The Pipeline Flow Index (RMSD) calculation step 216 of the present embodiment provides an abnormal threshold value for the change in the pressure of the furnace body. Since the blast furnace body 110 of the present embodiment is divided into eight blast furnace layers 112, and the pressure detecting device 120 is disposed in only one direction in each layer, the pipe flow index (RMSD) calculating step 216 of the present embodiment is The blast furnace layer 112 provides only one pipeline flow index, and the pipeline flow index corresponding to each historical valve full opening time point data is eight. However, in other embodiments of the invention, if the number of blast furnace layers 112 is maintained at eight but each layer has four pressure detecting devices 120, a pipe flow index (RMSD) calculation step 216 is provided for each blast furnace layer 112. There are 4 pipeline flow indicators, and the pipeline flow index corresponding to each historical valve full-open time point data is 32.

Next, a blast furnace alert layer decision step 218 is performed to determine the blast furnace alert layer based on historical blast furnace pressure data and its corresponding pipe flow index. Please refer to FIG. 3, which is a flow chart showing a blast furnace alert layer determining step 218 according to an embodiment of the present invention. In this embodiment, each historical valve full-open time point data corresponds to eight pipe flow indicators, and the eight pipe flow indexes respectively represent the critical pressure abnormality threshold values of the eight blast furnace layers 112 of the blast furnace 110. Therefore, in the blast furnace warning layer determining step 218, the abnormal blast furnace layer determining step 218a is first performed to determine whether the blast furnace pressure change value of each blast furnace layer 112 is greater than the corresponding pipeline flow index for each historical valve full opening time point, and At least one abnormal blast furnace layer is selected from the blast furnace layers, wherein the abnormal blast furnace layer is a blast furnace layer having a pressure change value greater than a corresponding pipeline flow index. Next, a candidate blast furnace layer determining step 218b is performed to select a candidate alert layer from the abnormal blast furnace layer based on the position of the abnormal blast furnace layer. This candidate warning layer is the highest blast furnace layer in the abnormal blast furnace layer. Then, a blast furnace alert layer selection step 218c is performed to select a blast furnace alert layer from a plurality of candidate alert layers corresponding to all historical valve full open time points for use in the subsequent prediction phase 220.

In this embodiment, the selection of the blast furnace alert layer can be determined based on the proportion of the blast furnace layer in the candidate alert layer. For example, in the candidate alert layer, the 7th blast furnace layer appears the most frequently, and its proportion is 60%, so the 7th blast furnace layer can be selected as the blast furnace warning layer. In other embodiments of the present invention, the time point data of the history valve may not be fully opened for cross-comparison, so as to select a blast furnace layer having a low correlation with the historical valve not fully open time point from the candidate warning layer. Blasting alert layer.

Please return to Figure 2. After the blast furnace alert layer is determined, a prediction phase 220 is then performed to utilize the blast furnace alert layer to predict whether the current blast furnace will experience pipe flow. In the prediction phase 220, a current pressure data providing step 222 is first performed to provide pressure data for the current blast furnace, such as current pressure values and pressure changes for each blast furnace layer 112. Next, a prediction step 224 is performed to predict whether a pipe flow phenomenon has occurred based on the pressure change value of the blast furnace alert layer and the corresponding pipe flow index.

Please refer to FIG. 4, which is a flow chart showing a prediction step 224 according to an embodiment of the present invention. In the prediction step 224 of the present embodiment, the pipe flow index calculation step 224a is first performed to calculate the pipe flow index corresponding to the blast furnace warning layer according to the pressure data provided in step 222. Then, a determining step 224b is performed to determine whether the pressure change value of the blast furnace alert layer is greater than the corresponding pipe flow index. When the pressure change value of the blast furnace warning layer is greater than the corresponding pipeline flow index, it is judged that the pipeline flow phenomenon is about to occur.

In other embodiments of the invention, the predicting step 224 may further include an AGE opening comparison step to utilize the opening of the control valve of the scrubber 130 as a cofactor for predicting the flow of the conduit. For example, the predicting step 224 can compare the current opening of the blast furnace control valve to a predetermined degree of valve opening. When the opening degree of the blast furnace control valve is greater than the preset valve opening degree (60% in this embodiment), and the pressure change value of the blast furnace warning layer is greater than the corresponding pipe flow index, it is judged that the pipe flow phenomenon is about to occur.

Referring back to FIG. 2, when it is predicted that the pipeline flow phenomenon is about to occur, the prediction method 200 of this embodiment performs a warning step 226 to issue a warning message through a computer system (for example, a blast furnace program computer) to warn the use of the blast furnace. Pipeline flow is about to happen. In this embodiment, the warning message may be a text message, a light message or a voice message, but embodiments of the invention are not limited thereto.

In addition, in the embodiment of the present invention, the neural network module can be used to interpolate the pressure values of the blast furnace layers 112 to obtain the pressure values at various points on the blast furnace wall (including no pressure is set). Detecting the point of the device 120, so that even if the blast furnace layer 112 is not layered according to the pressure detecting device 120, the blast furnace pipe flow phenomenon of the embodiment of the present invention can be performed according to the pressure value and the pressure change value obtained by the interpolation method. Prediction method 200.

While the invention has been described above in terms of several embodiments, it is not intended to limit the scope of the invention, and the invention may be practiced in various embodiments without departing from the spirit and scope of the invention. The scope of protection of the present invention is defined by the scope of the appended claims.

100. . . Blast furnace smelting system

110. . . Blast furnace body

112. . . Blast floor

120. . . Pressure detecting device

130. . . Dust washer

132. . . Control valve

200. . . Pipeline flow phenomenon prediction method

210. . . Analysis phase

212. . . Valve data supply steps

214. . . Historical pressure data providing steps

216. . . Pipe flow indicator calculation steps

218. . . Blast furnace alert layer decision step

218a. . . Abnormal blast furnace layer decision step

218b. . . Candidate blast furnace layer decision step

218c. . . Blast furnace warning layer selection step

220. . . Prediction stage

222. . . Current pressure data supply steps

224. . . Prediction step

224a. . . Pipe flow indicator calculation steps

224b. . . Judgment step

226. . . Warning step

The above and other objects, features, and advantages of the present invention will become more apparent and understood.

1 is a schematic view showing the structure of a blast furnace smelting system according to an embodiment of the present invention.

2 is a flow chart showing a method for predicting a phenomenon of a blast furnace pipe flow according to an embodiment of the present invention.

Figure 3 is a flow chart showing the steps of determining the blast furnace alert layer according to an embodiment of the present invention.

Figure 4 is a flow chart showing the prediction steps according to an embodiment of the present invention.

200. . . Pipeline flow phenomenon prediction method

210. . . Analysis phase

212. . . Valve data supply steps

214. . . Historical pressure data providing steps

216. . . Pipe flow indicator calculation steps

218. . . Blast furnace alert layer decision step

220. . . Prediction stage

222. . . Current pressure data supply steps

224. . . Prediction step

226. . . Warning step

Claims (10)

A method for predicting a blast furnace pipeline flow phenomenon, wherein the blast furnace comprises a plurality of pressure detectors, wherein the pressure detectors are located in a plurality of blast furnace layers having different heights, and the method for predicting the blast furnace pipeline flow phenomenon comprises: performing a data The analysis stage comprises: providing a plurality of history valve full-opening time points of a top-top dust washer, wherein the historical valve full-opening time point data corresponds to a plurality of historical valve full-opening time points, and the blast furnace is attached to the historical valves A pipeline flow phenomenon occurs at a full opening time; a plurality of historical blast furnace pressure data are provided, wherein the historical blast furnace pressure data are one-to-one corresponding to the historical valve full opening time points data, and each of the historical blast furnace pressure data includes the blast furnaces a plurality of historical blast furnace pressure values and a plurality of historical blast furnace pressure change values; calculating each of the historical valve full opening time points according to the historical blast furnace pressure values of each of the historical blast furnace pressure data and a pipeline flow index equation Corresponding multiple pipe flow indicators (RMSD), wherein the pipe flow indicators are a pair Layer corresponds to the plurality of the blast furnace, and the pipeline flow equation based index is expressed as follows: Where x _i is the i-th historical blast furnace pressure value, x _i +1 is the i+1th historical blast furnace pressure value, and N is the number of at least one high pressure change layer in the blast furnace layers, the at least one high pressure change The pressure change value of the layer is greater than a pressure change threshold; determining the blast furnace warning layer according to the historical blast furnace pressure values of each of the historical blast furnace pressure data and the pipeline flow indicators; and performing a pipeline flow prediction stage, Included: providing a current blast furnace pressure data, wherein the current blast furnace pressure data includes a current pressure value of the blast furnace warning layer and a current pressure change value; performing a predicting step to determine the pressure change value according to the blast furnace warning layer and corresponding The pipeline flow indicator predicts whether the pipeline flow phenomenon occurs; and when the prediction step determines that the pipeline flow phenomenon is about to occur, a warning message is issued to warn the blast furnace operator that the pipeline flow may occur. The method for predicting the phenomenon of blast furnace pipeline flow as described in claim 1 wherein the pressure change threshold is 0.15 kg/cm 2 . The method for predicting the phenomenon of blast furnace pipeline flow as described in claim 1 wherein the warning message is a text message, a light message or a voice message. The method for predicting a blast furnace pipeline flow phenomenon as described in claim 1, wherein the predicting step comprises: calculating the blast furnace warning layer corresponding to the current pressure value of the blast furnace warning layer and the current pressure change value The pipeline flow indicator determines whether the pressure change value of the blast furnace warning layer is greater than the corresponding pipeline flow index: and when the pressure change value of the blast furnace warning layer is greater than the corresponding pipeline flow index, determining that the pipeline flow phenomenon is about to occur. The method for predicting a blast furnace pipeline flow phenomenon as described in claim 1, wherein the predicting step comprises: calculating the blast furnace warning layer corresponding to the current pressure value of the blast furnace warning layer and the current pressure change value a pipeline flow indicator; determining whether the pressure change value of the blast furnace warning layer is greater than the corresponding pipeline flow index, and providing a first judgment result: determining whether the current valve opening degree of the top of the top dust washer is greater than a predetermined opening degree The threshold value is provided, and a second judgment result is provided: and when both the first judgment result and the second judgment result are yes, it is judged that the pipeline flow phenomenon is about to occur. The method for predicting a blast furnace pipeline flow phenomenon as described in claim 5, wherein the preset opening degree threshold is 60%. The method for predicting the phenomenon of blast furnace pipeline flow according to claim 1, wherein the pressure change values of the blast furnace layers are obtained by interpolation using a neural network module. For example, in the method for predicting the blast furnace pipeline flow phenomenon described in claim 1, wherein the step of determining the blast furnace warning layer comprises: performing a blast furnace pressure change value corresponding to each of the historical valve full opening time points data. An abnormality analysis step of obtaining a plurality of candidate warning layers corresponding to the historical time points of the full valve, wherein the abnormality analysis step comprises: determining whether the blast furnace pressure change value of each of the blast furnace layers is greater than the corresponding pipeline flow And an indicator for selecting at least one abnormal blast furnace layer from the blast furnace layers, wherein the at least one blast furnace pressure change value of the at least one abnormal blast furnace layer is greater than the corresponding at least one pipeline flow index; and according to the at least one abnormal blast furnace layer Positioning to determine the candidate alert layers from the at least one abnormal blast furnace layer, wherein the candidate alert layer is one of the highest positions in the at least one abnormal blast furnace layer; and performing a blast furnace alert layer selection step to The blast furnace warning layer is selected among the candidate warning layers corresponding to the historical valve full-open time point data. The method for predicting a blast furnace pipeline flow phenomenon as described in claim 8 wherein the blast furnace alert layer selection step comprises: counting the number of occurrences of each of the candidate alert layers; and calculating each of the candidates The number of occurrences of the alert layer is selected to select the blast furnace alert layer, wherein the number of occurrences of the blast furnace alert layer is greater than the number of occurrences of each of the other candidate alert layers. The method for predicting a blast furnace pipeline flow phenomenon as described in claim 8 wherein the blast furnace warning layer selection step comprises: providing a plurality of historical time points of the historical valve not fully open; and according to the historical valve not fully open time point data The blast furnace warning layer is selected from the candidate warning layers, wherein the correlation between the blast furnace warning layer and the historical valve not fully open time points is less than the other candidate warning layers and the historical valve not fully open time points. Sex.