JP7192992B2 - Blast Furnace Abnormality Judgment Method, Stable Period Model Learning Method, Blast Furnace Operation Method, and Blast Furnace Abnormality Judgment Device - Google Patents

Description

The present invention relates to a blast furnace abnormality determination method, a stable period model learning method, a blast furnace operation method, and a blast furnace abnormality determination apparatus.

In a plant, various abnormalities may occur during operation. Therefore, in order to prevent serious abnormalities and troubles, early detection of abnormalities during operation and appropriate countermeasures are required. For example, in a blast furnace, major troubles may occur due to changes in the properties and distribution of charged raw materials, unburned pulverized coal blown from tuyeres, and the like.

Signs of plant operational abnormalities and troubles are often included in sensor values detected by sensors installed in the plant, index values calculated based on the sensor values, plant image data, and the like. Conventionally, for example, upper and lower limit values (thresholds) are simply set for each sensor value or index value, and an abnormality is determined when the range of the upper and lower limit values is exceeded. However, with this method, it is difficult to capture small changes in operating conditions in the initial state of abnormality (before serious trouble). Therefore, for example, Patent Literature 1 proposes a technique of grasping a sign of an operational abnormality using the Q statistic.

JP 2017-128805 A

However, in the method using the Q statistic as in Patent Document 1, for example, the data group characteristic in which the synchronicity in the normal state appears in the principal component value (in the same document, "the synchronous behavior characteristic in the normal state of the blast furnace shaft pressure") There is Therefore, it has been difficult to simultaneously use various data with different characteristics to determine abnormality.

The present invention has been made in view of the above, and is a blast furnace abnormality determination method and a stable period model learning method that can detect an operational abnormality in a blast furnace at an early stage by simultaneously using various data with different characteristics. An object of the present invention is to provide a method, a blast furnace operating method, and a blast furnace abnormality determination device.

In order to solve the above-described problems and achieve the object, the blast furnace abnormality determination method according to the present invention uses a plurality of operation data in the stable period of the blast furnace so that the input value and the output value are the same. An abnormality determination step of inputting a plurality of operation data to be determined to the learned stable period model and determining an operation abnormality of the blast furnace based on the difference between the input value and the output value at that time. .

Further, in the blast furnace abnormality determination method according to the present invention, in the above invention, the abnormality determination step includes each input value and each output when a plurality of operation data to be determined is input to the stable period model The integrated value of the difference from the value is calculated, and when the integrated value of the difference exceeds a preset threshold value, it is determined that there is an operational abnormality.

Further, in the blast furnace abnormality determination method according to the present invention, in the above invention, the abnormality determination step includes inputting a plurality of operation data to be determined to the stable period model, each input value and each output If the integrated value of the difference on the positive side exceeds a preset threshold value on the positive side, or if the integrated value of the difference on the negative side exceeds the preset value If the threshold on the negative side is exceeded, it is determined that there is an operational abnormality.

Further, in the blast furnace abnormality determination method according to the present invention, in the above invention, after the abnormality determination step, each input value and each It further includes a display step of displaying the integrated value for each positive and negative difference from the output value in a stacked graph.

In order to solve the above-described problems and achieve the object, the learning method of the stable period model according to the present invention inputs a plurality of operation data in the stable period of the blast furnace to an autoencoder to obtain an input value and an output value. It includes a training step that builds a plateau model trained to be the same.

Further, in the stable period model learning method according to the present invention, in the above invention, the learning step selects a plurality of operation data in the stable period of the blast furnace based on the blast flow rate, and constructs the stable period model. .

Further, in the stable period model learning method according to the present invention, in the above invention, the learning step includes, among the plurality of operation data of the blast furnace, the blast flow rate being equal to or greater than a preset threshold value, and the blast flow rate selects operation data excluding a predetermined time before and after the time when becomes equal to or greater than the threshold, and constructs the stable period model.

In order to solve the above-described problems and achieve the object, the blast furnace operation method according to the present invention changes the blast furnace operation based on the determination result of the above-described blast furnace abnormality determination method.

In order to solve the above-described problems and achieve the object, the blast furnace abnormality determination device according to the present invention uses a plurality of operation data in the stable period of the blast furnace so that the input value and the output value are the same. An abnormality determination means for inputting a plurality of operation data to be determined to the learned stable period model and determining an operation abnormality of the blast furnace based on the difference between the input value and the output value at that time. .

According to the present invention, by performing abnormality determination using a stable period model that has been learned so that the input value and the output value are the same, various data with different characteristics can be used at the same time to determine the operation abnormality of the blast furnace. Detection can be done early.

FIG. 1 is a diagram showing a schematic configuration of a blast furnace abnormality determination apparatus according to an embodiment of the present invention. FIG. 2 is a diagram showing an overview of an autoencoder used in the stable model learning method according to the embodiment of the present invention. FIG. 3 is a flow chart showing the flow of the stable model learning method according to the embodiment of the present invention. FIG. 4 is a flow chart showing the flow of the abnormality determination method according to the embodiment of the present invention. FIG. 5 is an embodiment of the blast furnace abnormality determination method according to the present invention, and is a diagram showing an example of abnormality determination using verification data for a predetermined period.

A blast furnace abnormality determination method, a stable period model learning method, a blast furnace operation method, and a blast furnace abnormality determination apparatus according to embodiments of the present invention will be described with reference to the drawings. Below, the blast furnace abnormality determination device, the learning method of the stable period model, the blast furnace abnormality determination method, and the blast furnace operation method will be described in this order. In addition, this invention is not limited to embodiment described below.

(abnormality determination device)
A configuration of a blast furnace abnormality determination apparatus according to an embodiment of the present invention will be described with reference to FIG. An abnormality determination device 1 is for determining an abnormality in a plant such as a blast furnace. The abnormality determination device 1 includes a sensor group 11, a data collection unit 12, a storage unit 13, a calculation unit 14, and a display unit 15, as shown in FIG.

The sensor group 11 consists of a plurality of sensors provided in the blast furnace, and outputs detected sensor values to the data collection unit 12 . The sensor group 11 includes, for example, a sensor group installed around the furnace body of the blast furnace.

The data collection unit 12 collects sensor values detected by the sensor group 11 and accumulates them in the storage unit 13 as operation data. The data collection unit 12 also calculates index values based on the sensor values detected by the sensor group 11, and accumulates the index values in the storage unit 13 as operation data.

The above-mentioned "index value based on the sensor value (hereinafter simply referred to as "index value")" includes an index value relating to the furnace heat of the blast furnace, an index value relating to the ventilation resistance of the blast furnace, and the like. Further, the above-mentioned "index value related to the furnace heat of the blast furnace" includes an index value calculated from the heat quantity of the furnace body and the combustion heat at the tip of the tuyere.

The storage unit 13 includes a recording medium such as an EPROM (Erasable Programmable ROM), a hard disk drive (HDD), and a removable medium. Examples of removable media include disk recording media such as USB (Universal Serial Bus) memories, CDs (Compact Discs), DVDs (Digital Versatile Discs), and BDs (Blu-ray (registered trademark) Discs).

The storage unit 13 stores a stable period model 131 constructed by the learning unit 141 and operation data (sensor values and index values) collected by the data collection unit 12 . The operation data stored in the storage unit 13 includes, for example, data (learning data) used when constructing the stable period model 131 (see FIG. 3), and when performing abnormality determination using the stable period model 131 (see FIG. 4) data (verification data) used.

The stable period model 131 is a model constructed based on a plurality of operation data in the stable period of the blast furnace. This stable period model 131 is constructed by the learning unit 141 based on the operational data collected by the data collecting unit 12 . A method of constructing the stable period model 131 will be described later.

The calculation unit 14 is realized by, for example, a processor such as a CPU (Central Processing Unit) and a memory (main storage unit) such as a RAM (Random Access Memory) and a ROM (Read Only Memory).

The computing unit 14 loads a program into the work area of the main storage unit, executes the program, and controls each component through the execution of the program, thereby realizing a function that meets a predetermined purpose. Specifically, the calculation unit 14 functions as a learning unit (learning means) 141, a difference calculation unit (difference calculation means) 142, and an abnormality determination unit (abnormality determination means) 143 through execution of programs. Note that in the present embodiment, as shown in FIG. 1, the function of each unit (the learning unit 141, the difference calculation unit 142, and the abnormality determination unit 143) is realized by one calculation unit (≈computer). The function of each part may be realized by an operation part (≈computer).

The learning unit 141 uses a plurality of operation data in the stable period of the blast furnace (hereinafter referred to as "stable period data") to perform learning so that the input value and the output value are the same, so that the stable period model 131 to build. Specifically, the learning unit 141 constructs the stable period model 131 using an autoencoder, which is one method of deep learning. Then, the learning unit 141 stores the constructed stable period model 131 in the storage unit 13 . When constructing stable period data, it is preferable to use stable period data for at least the past six months.

An overview of the autoencoder used by the learning unit 141 will be described with reference to FIG. An autoencoder is one of neural network mechanisms, and is a technique for reducing the dimensionality of input data and extracting features. In the autoencoder, as shown in FIG. 2, neurons of "the number of dimensions obtained by compressing the input data" are provided in the intermediate layer. Also, the number of dimensions of the intermediate layer is made smaller than that of the input layer in order to extract the feature amount of the input data. Then, the output value of the output layer is set to an output value that can reproduce the input value of the input data.

An autoencoder having such a configuration embeds (encodes) input data in a small dimension once, and reconstructs the input data based on the encoded data. That is, by encoding with an autoencoder, data can be expressed with a smaller number of dimensions than originally. Returning to FIG. 1, the description of the configuration of the abnormality determination device 1 is continued.

The difference calculation unit 142 inputs a plurality of operation data (verification data) to be judged to the stable period model 131, and calculates the difference between the input value and the output value at that time. Specifically, the difference calculation unit 142 calculates an integrated value of differences between each input value and each output value when a plurality of operation data are input to the stable period model 131 .

The "integrated value of the difference between each input value and each output value" indicates, for example, a value obtained by integrating the absolute values of the differences between each input value and each output value. Further, the difference calculation unit 142 may calculate, for each positive and negative, the integrated value of the difference between each input value and each output value instead of the integrated value of the absolute value of the difference between each input value and each output value. good.

The abnormality determination unit 143 determines an operational abnormality of the blast furnace based on the difference between each input value and each output value when a plurality of operation data to be determined is input to the stable period model 131 . Specifically, the abnormality determination unit 143 determines that there is an operational abnormality when the integrated value of the difference between each input value and each output value exceeds a preset threshold, and when it is less than the threshold, an operational abnormality Judged as none. In this way, by looking at the integrated value of the difference between each input value and each output value to the stable period model 131, small abnormalities in each item are amplified as accumulated values, so it is possible to detect operational abnormalities at an early stage. can be done. Note that the above threshold can be calculated empirically and experimentally in advance.

Here, when the difference calculation unit 142 described above calculates the integrated value of the difference between each input value and each output value for each positive and negative value, the following processing is performed. In this case, the abnormality determination unit 143 determines that there is an operational abnormality when the integrated value of the difference on the positive side exceeds a preset threshold value on the positive side. Further, the abnormality determination unit 143 determines that there is an operational abnormality when the integrated value of the negative side difference exceeds a preset negative side threshold value. Further, the abnormality determination unit 143 determines that there is no operational abnormality when the integrated value of the positive side difference is less than the preset positive side threshold value. Moreover, the abnormality determination unit 143 determines that there is no operational abnormality when the integrated value of the negative side difference is less than a preset negative side threshold value. In this way, in the abnormality determination device 1, by performing abnormality determination for each positive and negative, the positive side abnormality and the negative side abnormality in each item are respectively amplified as a cumulative value, so that an operational abnormality can be detected early. be able to.

The display unit 15 is implemented by a display device such as an LCD display or a CRT display, and displays various information based on display signals input from the calculation unit 14 . The information displayed on the display unit 15 includes, for example, a stacked graph (see FIG. 5) showing integrated values for each positive and negative difference between each input value and each output value, as will be described later.

(Stable period model learning method)
A method of learning a stable model (hereinafter referred to as "model learning method") executed by the abnormality determination device 1 will be described with reference to FIG. In the model learning method, a data selection step (step S1) and a model construction step (step S2) are performed in this order. Note that the construction of the stable period model 131 shown in FIG.

In the data selection step, the learning unit 141 selects stable period data, which are a plurality of operation data in the stable period of the blast furnace (step S1). In step S1, it is possible to select stable period data based on various indices. For example, if the operational state of the blast furnace deteriorates, it is generally possible to change the blast flow rate to a smaller value (reduce the wind). Therefore, it is preferable to select the stable period data based on this air flow rate.

In this case, among the plurality of operation data of the blast furnace, the blast flow rate is equal to or higher than a preset threshold, and the operation data excluding the predetermined time before and after the time when the blast flow rate is equal to or higher than the threshold is used as the stable period data. Select is preferred. This is for the following two reasons.
(1) A few hours before and after the wind reduction may include signs of anomalies, so it is not desirable as data for the stable period.
(2) Conversely, when the blast flow rate exceeds the threshold, it is assumed that the unsteadiness is very strong for several hours when the blast furnace blows up, which is not desirable as stable period data.

The above-mentioned "predetermined time" is preferably eight hours, for example. This is because it takes about 8 hours for the raw material charged from the upper part of the blast furnace to descend to the lower part of the furnace, and if it exceeds 8 hours, the possibility that there are factors in the furnace that deteriorate the operating conditions will be low. It's for.

Subsequently, in the model construction process, the learning unit 141 inputs the stable period data selected in step S1 to the autoencoder to construct the stable period model 131 (step S2). In step S2, as shown in FIG. 2, data sets of the stable period data are set to the input and the output for the number N of data, and the stable period model 131 is constructed. In step S2, in order to increase the degree of expression of the stable period model 131, it is preferable to use stable period data having a large number of data N and moderate variations within the stable range. In this way, by using an autoencoder, it is possible to obtain a stable state feature amount (that is, a digitized feature representing a normal state with reduced dimensions) of the data group of the input stable period data. can.

(Abnormality determination method)
An abnormality determination method executed by the abnormality determination device 1 will be described with reference to FIG. In the abnormality determination method, a data input process (step S11), a difference calculation process (step S12), and an abnormality determination process (steps S13 to S15) are performed in this order. In this embodiment, the model learning method (FIG. 3) and the abnormality determination method (FIG. 4) are described separately, but the abnormality determination method may be performed continuously after the model learning method.

In the data input step, the difference calculation unit 142 inputs operation data (verification data) to be determined to the stable period model 131 (step S11).

Subsequently, in the difference calculation step, the difference calculation unit 142 calculates an integrated value of differences (errors) between each input value and each output value for each input operation data (step S12). In step S12, the integrated value of the absolute values of the difference between each input value and each output value may be calculated, or the integrated value of the difference between each input value and each output value may be calculated for each positive and negative. may

Here, "to calculate the integrated value of the difference for each of the positive and negative values" means to add the values of the items whose difference is positive and to add the values of the items whose difference is negative. In this method, positive side and negative side shifts differ depending on the item, but positive side items are added on the positive side, negative side items are added on the negative side, and each is evaluated. As a result, when the deviation from the normal state increases when an abnormality occurs, the integrated value increases (or decreases) cumulatively, so that it is possible to quickly grasp the occurrence of an abnormality.

Subsequently, in the abnormality determination step, the abnormality determination unit 143 determines whether or not the integrated value of the difference between each input value and each output value exceeds a preset threshold value (step S13). Here, when the integrated value of the absolute value of the difference between each input value and each output value is calculated in step S12, the integrated value is compared with a threshold value in step S13. On the other hand, in step S12 above, when the integrated value of the difference between each input value and each output value is calculated for each positive and negative value, in step S13, the integrated value of the difference on the positive side and the threshold value on the positive side are calculated. Along with the comparison, the integrated value of the difference on the negative side and the threshold value on the negative side are compared.

When it is determined in step S13 that the integrated value of the difference exceeds the preset threshold value (Yes in step S13), the abnormality determination unit 143 determines that there is an operational abnormality (step S14), and terminates this process. On the other hand, in step S13, when it is determined that the integrated value of the difference is less than the preset threshold value (No in step S13), the abnormality determination unit 143 determines that there is no operational abnormality (step S15), and terminates this process. finish.

In the abnormality determination method, in addition to steps S11 to S15 described above, a display step of creating a stacked graph in which the difference between each input value and each output value is stacked for each positive and negative value and displaying it on the display unit 15 is performed. good too. By performing this display step, the degree of abnormality for each operation data (input item) can be visualized.

(Blast furnace operating method)
In the blast furnace operation method, the operation of the blast furnace is changed based on the determination result of the blast furnace abnormality determination method. This makes it possible to prevent serious abnormalities and troubles in the blast furnace.

According to the blast furnace abnormality determination method, the stable period model learning method, the blast furnace operation method, and the blast furnace abnormality determination device 1 according to the present embodiment described above, learning is performed so that the input value and the output value are the same. Abnormality determination is performed using the stable period model 131 thus obtained. As a result, various data with different characteristics can be used simultaneously to detect abnormal operation of the blast furnace at an early stage.

In addition, according to the present embodiment, by looking at the integrated value of the difference between each input value and each output value to the stable period model 131, small abnormalities in each item are amplified as accumulated values, so that operational abnormalities are detected. It can be detected early.

Further, according to the present embodiment, by simultaneously using various data with different characteristics to determine an abnormality, the abnormality can be detected more quickly than in the case of performing an abnormality determination using data with the same characteristics. can be done.

In addition, in the present embodiment, instead of dividing various data with different characteristics that are expected to be related to operational abnormalities and constructing models for each characteristic, a single model is constructed to determine abnormality. is performed, the abnormality determination can be performed more easily.

(Example)
An embodiment of the invention will now be described with reference to FIG. In this example, the stable period data of the operation of the blast furnace was input to the autoencoder, the stable period model was constructed, and the abnormality was determined using the stable period model. In this example, as the stable period data, the operation data when the air flow rate exceeds 75% of the normal operation was used. At that time, operation data for 8 hours before and after the air flow rate became 75% or less of the steady operation was excluded.

As the stable period data, a stable period model was constructed by using data collected in an hourly cycle for about 1.5 years for 36 items that may be related to blast furnace operational abnormalities. FIG. 5 shows the results of inputting operation data (verification data) from before the occurrence of the operational abnormality to the time of the occurrence of the operational abnormality into this stable period model. In the figure, the vertical axis is the index abnormal value indicating the abnormality of the index value, and the horizontal axis is the time.

The sensor signals used were the measured values of blast furnace exhaust gas (N ₂ , H ₂ , CO, CO ₂ ), the blast furnace gas utilization rate calculated based on these values, and the blast furnace ventilation resistance values (ventilation resistance of the entire furnace body). , Furnace lower/middle/upper ventilation), Furnace heat-related indexes (blown sensible heat amount, tuyere combustion heat amount, solution reaction heat amount, furnace top gas sensible heat amount, blast moisture decomposition heat amount, furnace body radiation heat amount , pulverized coal combustion heat quantity, pulverized coal decomposition heat quantity, slag sensible heat quantity, charged raw material sensible heat quantity, hot metal sensible heat quantity), values related to operation manipulated variables (blast flow rate, blast pressure, furnace top pressure, blast moisture content, blast temperature, amount of PCI injection), and processing values of the sensor group around the furnace body (average for each shaft pressure height, average for each temperature height of each part of the furnace body), and the like.

As verification data, operation data for about 1.5 months, excluding wind breaks, was used. Then, the verification data is input to the stable period model, the difference between the input value and the output value for each item is obtained, and the difference between the input value and the output value is integrated for each positive and negative. It is represented by a stacked graph as shown in .

As shown in FIG. 5, it can be seen that signs of abnormality appear about 10 days before the operator recognizes the operational abnormality and takes measures, and various index values increase in the abnormal direction. Then, about six days ago, it exceeded the set positive threshold and was determined to be abnormal. By presenting the information to the operator at this time, it becomes possible to take preventive measures, and the possibility of avoiding or reducing the damage is increased. In addition, in FIG. 5, the difference between the input value and the output value for each item is expressed in a stacked graph with different colors for each item, so that the operator can visually grasp which item is abnormal. becomes possible.

As described above, the blast furnace abnormality determination method, the stable period model learning method, the blast furnace operation method, and the blast furnace abnormality determination device according to the present invention have been specifically described with the embodiments and examples for carrying out the invention. The gist of the invention should not be limited to these descriptions, but should be broadly interpreted based on the descriptions of the claims. Further, it goes without saying that various changes and alterations based on these descriptions are also included in the gist of the present invention.

Reference Signs List 1 abnormality determination device 11 sensor group 12 data collection unit 13 storage unit 131 stable period model 14 calculation unit 141 learning unit 142 difference calculation unit 143 abnormality determination unit 15 display unit

Claims (8)

Using multiple operational data in the stable period of the blast furnace, input multiple operational data to be judged to the stable period model trained so that the input value and the output value are the same, and Based on the difference between the input value and the output value, an abnormality determination step for determining an operational abnormality of the blast furnace ,
The stable period model is, among the plurality of operation data of the blast furnace, the blast flow rate is equal to or higher than a preset threshold, and the operation data excluding a predetermined time before and after the time when the blast flow rate is equal to or higher than the threshold. Blast furnace abnormality judgment method built based on . The abnormality determination step includes:
Calculate the integrated value of the difference between each input value and each output value when a plurality of operation data to be judged is input to the stable period model,
2. The blast furnace abnormality determination method according to claim 1, wherein it is determined that there is an operational abnormality when the integrated value of the difference exceeds a preset threshold value. The abnormality determination step includes:
Calculate the integrated value of the difference between each input value and each output value when a plurality of operation data to be judged is input to the stable period model for each positive and negative,
If the cumulative value of the positive side difference exceeds a preset positive threshold value, or if the cumulative value of the negative side difference exceeds the preset negative threshold value, it is determined that there is an operational abnormality. The blast furnace abnormality determination method according to claim 1 or 2. After the abnormality determination step, when a plurality of operation data to be determined is input to the stable period model, the accumulated value for each positive and negative difference between each input value and each output value is displayed in a stacked graph. 4. The blast furnace abnormality determination method according to claim 3, further comprising a display step. By inputting a plurality of operation data in the stable period of the blast furnace into the autoencoder, including a learning step of constructing a stable period model learned so that the input value and the output value are the same,
In the learning step, among the plurality of operation data of the blast furnace, the blast flow rate is equal to or greater than a preset threshold, and the operation data excluding a predetermined time before and after the time when the blast flow rate becomes equal to or greater than the threshold value. A stable model learning method for building the stable model based on . 6. The method of learning a stable period model according to claim 5, wherein the learning step selects a plurality of operation data in the stable period of the blast furnace and constructs the stable period model based on the air flow rate. A method of operating a blast furnace, wherein the operation of a blast furnace is changed based on the judgment result of the method for judging abnormality of a blast furnace according to any one of claims 1 to 4. Using multiple operational data in the stable period of the blast furnace, input multiple operational data to be judged to the stable period model trained so that the input value and the output value are the same, and Abnormality determination means for determining an operational abnormality of the blast furnace based on the difference between the input value and the output value ,
The stable period model is, among the plurality of operation data of the blast furnace, the blast flow rate is equal to or higher than a preset threshold, and the operation data excluding a predetermined time before and after the time when the blast flow rate is equal to or higher than the threshold. A blast furnace abnormality determination device constructed based on

Images