Blast furnace condition diagnosis method and system based on multi-time granularity dynamic reasoning mechanism

Abstract

The invention provides a blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism, which comprises the following steps: S1, setting a plurality of time granularities, and establishing a furnace condition diagnosis rule base according to a preset rule; s2, collecting data in real time, storing the data in a dynamic database, extracting the data of the dynamic database every first time granularity, performing furnace condition diagnosis of the first time granularity according to rules of a furnace condition diagnosis rule base, and storing a diagnosis result in a process database; s3, calling data of the process database and the dynamic database every Nth time granularity, performing furnace condition diagnosis of the Nth time granularity according to rules of the furnace condition diagnosis rule base, and storing a diagnosis result in the process database; and S4, calling diagnosis results of different time granularities. According to the blast furnace condition diagnosis method with multiple time granularities, a dynamic reasoning mechanism can reduce a large amount of database read-write burden, and the calculation performance of a furnace condition diagnosis model is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of blast furnace automated diagnosis, and in particular to a blast furnace condition diagnosis method and system based on a multi-time granularity dynamic reasoning mechanism. Background Art

[0002] Blast furnace smelting is a typical "black box" process. Process personnel can only diagnose the internal operating status of the blast furnace through the blast furnace operating parameters. Judgment based on manual experience is often subjective, and it is difficult to take into account data from different sources and collection frequencies. The judgment process may delay the handling of abnormal conditions.

[0003] The Chinese patent document with publication number CN111593155A, a blast furnace diagnostic system and method, discloses a blast furnace diagnostic system and method, wherein the blast furnace diagnostic system includes a data acquisition module, a data analysis module and a real-time monitoring and alarm module, and may also include a short-term abnormal diagnosis module, a long-term abnormal diagnosis module and a production management module. Compared with the prior art, the blast furnace diagnostic system of the present invention, by establishing the time correspondence between the blast furnace input parameters, process parameters and output parameters, can realize the systematic monitoring and diagnosis of the entire blast furnace ironmaking process from raw materials, blast furnace operation to products, so that the abnormal alarms or product quality alarms occurring during production can be traced back quickly and the countermeasures can be formulated. Further, the present invention realizes all-round monitoring of the production of the blast furnace by real-time monitoring, short-term abnormal monitoring, long-term abnormal monitoring and management monitoring of the blast furnace, effectively guides the production of the blast furnace, is conducive to the stability of the blast furnace condition, and improves the economic benefits of the blast furnace. The invention mainly emphasizes which data parameters the diagnostic system uses to examine, and how to set data rules for judgment. Although it mentions that blast furnace diagnosis needs to consider short-term and long-term judgments, it does not give a specific implementation method.

[0004] The Chinese patent document with the publication number CN117094576A, a method for quantitatively evaluating the stability of blast furnace conditions, discloses a method for quantitatively evaluating the stability of blast furnace conditions, and belongs to the field of process management. The core key parameters that can characterize the operation status of the blast furnace are selected, and the evaluation dimensions that can macroscopically characterize the stability of the blast furnace are determined, and each dimension is used as a sub-index of the total stability; the calculation formula of the blast furnace stability index is constructed to realize the quantitative characterization of the stability of the furnace condition; the operation status and stability of the blast furnace are quantitatively calculated and tracked from different dimensions and time granularities; the overall stability of the blast furnace is quantitatively calculated from different time granularities, and quantitative measurement of different time granularities is realized; through the quantitative measurement of the stability of the furnace condition, the operator is helped to evaluate the furnace condition in real time and quickly, take corresponding measures, improve indicators or avoid abnormal production processes. It realizes the transformation of the stability of the blast furnace condition from qualitative to quantitative through the quantitative characterization and real-time display of the stability of the furnace condition, and provides a basis for continuous online monitoring of the furnace condition and tracking the stability trend. Although the invention also mentions different time granularities, it mainly emphasizes the quantitative method of stability of furnace condition diagnosis and does not involve how to achieve result reuse between different time granularities and how to process high-frequency data. The present invention mainly focuses on how to reduce the database reading and writing burden during program implementation and improve versatility, and can be combined with other rule bases.

[0005] The paper "Study on the Evaluation Method of Blast Furnace Condition Stability Based on Data Mining" discloses that this paper attempts to use data mining technology to propose an unsupervised furnace condition evaluation method based on cluster distance-regression to judge and evaluate the stability of blast furnace condition. The main research contents include: (1) The daily average production data of a large blast furnace in China for 7 consecutive years were analyzed, and it was found that the historical data obtained had missing values, outliers, parameter redundancy, parameter inconsistency and other problems. The relevant program was written in Python language to complete the data preprocessing process such as missing value deletion, outlier box plot identification, Z-Score standardization, data feature conversion, etc. The processed historical data met the basic requirements of data mining. (2) An unsupervised learning method for blast furnace condition evaluation based on cluster analysis-distance-regression was proposed. The parameters were clustered and analyzed using the K-means algorithm. The reference vector representing the stability of the blast furnace was obtained by the mean method and Z-Score standardization method. The deviation distance between the parameters of each day and the reference vector was calculated using the Euclidean distance. The deviation distance was fitted using the multivariate linear regression method to obtain the blast furnace condition index. This method can be trained and learned without the need for furnace condition records, avoiding the problems of unreliable labels and poor model portability of traditional models. (3) Based on the theoretical knowledge of blast furnace smelting and expert experience, the 78 selected state parameters are divided into five attribute sets: unloading, top gas, furnace body heat, air volume and pressure, and furnace cylinder. A two-step clustering-distance regression model is established (i.e., the sub-attribute set is evaluated to obtain the sub-index and then evaluated). The comparative test with the one-step clustering-distance-regression model shows that the two-step clustering-distance-regression model greatly improves the accuracy of evaluating furnace conditions. The accuracy of judging abnormal furnace conditions is improved by 29.63%. (4) The simulation results of the two-step clustering-distance-regression model are analyzed, and the parameter weights are further optimized. Finally, the accuracy of the blast furnace condition index in judging the furnace condition is 80.79%, the accuracy of judging the normal furnace condition is 86.48%, and the accuracy of judging the abnormal furnace condition is 73.06%. The paper does not consider the requirements of on-site furnace condition diagnosis at different time granularities, only makes judgments based on fixed data sources, and does not consider real-time dynamic reasoning. The dynamic reasoning mechanism provided by the present invention only focuses on how to deal with time granularity, but has no requirements on the form of the rule base. It emphasizes storing high-frequency diagnosis results and using them for low-frequency diagnosis without frequent and large-scale reading of historical data. In fact, it can also be combined with AI algorithms to perform unsupervised reasoning and judgment on furnace conditions within different time granularity periods.

[0006] Therefore, a blast furnace condition diagnosis method and system based on a multi-time granularity dynamic reasoning mechanism is needed. Summary of the invention

[0007] In view of the defects in the prior art, the object of the present invention is to provide a blast furnace condition diagnosis method and system based on a multi-time granularity dynamic reasoning mechanism.

[0008] A blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism provided by the present invention comprises:

[0009] Step S1: setting multiple time granularities and establishing a furnace condition diagnosis rule base according to preset rules;

[0010] Step S2: collecting data in real time and storing the data in a dynamic database, extracting data from the dynamic database at first time granularity, performing furnace condition diagnosis at the first time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and storing the diagnosis results in a process database;

[0011] Step S3: every Nth time granularity, calling the diagnostic results of the process database and the data of the dynamic database, performing furnace condition diagnosis at the Nth time granularity according to the diagnostic rules of the furnace condition diagnosis rule base, and storing the diagnostic results in the process database; N is an integer greater than 1;

[0012] Step S4: according to actual needs, retrieve the diagnosis results of different time granularities.

[0013] Preferably, the step S1 comprises:

[0014] Step S1.1: according to a preset rule, set the time granularity; specifically, set the time granularity in order from small to large, to the first time granularity, the second time granularity, ... the Nth time granularity;

[0015] Step S1.2: Establish a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

[0016] Preferably, the step S2 comprises:

[0017] Step S2.1: collect data in real time and store the collected data in a dynamic database;

[0018] Step S2.2: extracting data of a preset type from the dynamic database at every first time granularity according to the rules of the furnace condition diagnosis rule base, and storing the data in the buffer area as diagnostic data;

[0019] Step S2.3: traverse the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all furnace condition diagnoses within the first time granularity are completed, and enter step S2.4;

[0020] Step S2.4: Record the diagnosis result of the first time granularity into the process database, and release the diagnosis data in the buffer area.

[0021] Preferably, step S3 comprises:

[0022] Step S3.1: At every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, the data of the dynamic database and the diagnosis results of the process database are respectively extracted and stored in the cache area as the diagnosis data of this time granularity;

[0023] Step S3.2: When the same type of furnace condition diagnosis rule corresponds to both the jth time granularity and the Nth time granularity, the diagnosis result of the jth time granularity in the process database is directly called as the diagnosis result of the Nth time granularity; j is an integer less than N;

[0024] Step S3.3: traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter step S3.4;

[0025] Step S3.4: Record the diagnostic result of the Nth time granularity into the process database, and release the diagnostic data in the buffer area.

[0026] Preferably, in step S3,

[0027] When performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnosis score; if the average value of the data is not within the set range, the diagnosis score is reduced; otherwise, it is increased;

[0028] Based on the obtained diagnostic scores, a unified evaluation standard is formed between different diagnostic results.

[0029] A blast furnace condition diagnosis system based on a multi-time granularity dynamic reasoning mechanism provided by the present invention comprises:

[0030] Module M1: Set multiple time granularities and establish a furnace condition diagnosis rule base according to preset rules;

[0031] Module M2: collects data in real time and stores the data in a dynamic database. At each first time granularity, extracts data from the dynamic database, performs furnace condition diagnosis at the first time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and stores the diagnosis results in the process database.

[0032] Module M3: every Nth time granularity, calling the diagnostic results of the process database and the data of the dynamic database, performing furnace condition diagnosis at the Nth time granularity according to the diagnostic rules of the furnace condition diagnosis rule base, and storing the diagnostic results in the process database; N is an integer greater than 1;

[0033] Module M4: retrieve diagnostic results of different time granularities according to actual needs.

[0034] Preferably, the module M1 comprises:

[0035] Module M1.1: according to the preset rules, set the time granularity; specifically, set the time granularity in the order of time from small to large, to the first time granularity, the second time granularity, and so on, the Nth time granularity;

[0036] Module M1.2: Establish a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

[0037] Preferably, the module M2 comprises:

[0038] Module M2.1: collect data in real time and store the collected data in a dynamic database;

[0039] Module M2.2: At every first time granularity, according to the rules of the furnace condition diagnosis rule base, extract data of a preset type from the dynamic database and store it in the cache as diagnostic data;

[0040] Module M2.3: traverse the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result until all furnace condition diagnoses within the first time granularity are completed, and enter module M2.4;

[0041] Module M2.4: Record the diagnostic results of the first time granularity into the process database and release the diagnostic data in the buffer area.

[0042] Preferably, the module M3 includes:

[0043] Module M3.1: Every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, extract the data of the dynamic database and the diagnosis results of the process database respectively, and store them in the cache area as the diagnosis data of this time granularity;

[0044] Module M3.2: When the same type of furnace condition diagnosis rules correspond to both the jth time granularity and the Nth time granularity, the diagnosis result of the jth time granularity in the process database is directly called as the diagnosis result of the Nth time granularity; j is an integer less than N;

[0045] Module M3.3: Traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis results until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter module M3.4;

[0046] Module M3.4: Record the diagnostic results of the Nth time granularity into the process database and release the diagnostic data in the buffer area.

[0047] Preferably, in the module M3,

[0048] When performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnosis score; if the average value of the data is not within the set range, the diagnosis score is reduced; otherwise, it is increased;

[0049] Based on the obtained diagnostic scores, a unified evaluation standard is formed between different diagnostic results.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention constructs a blast furnace condition diagnosis method with multiple time granularities, which can respectively characterize the evolution of the blast furnace condition in the short, medium and long term, and provide diagnosis, early warning and operation guidance for the blast furnace smelting process in real time and accurately through the inference of the blast furnace condition;

[0052] 2. The dynamic reasoning mechanism designed by the present invention combines the diagnostic data reading operations at different time granularities and reduces repeated diagnostic calculations, which can effectively reduce the database reading and writing burden and computing resource consumption, and improve the computing performance of the furnace condition diagnosis model;

[0053] 3. The present invention stores the diagnostic results of different time granularities in the process database, which facilitates subsequent data management and query. Users can retrieve the diagnostic results of different time granularities according to actual needs for further analysis and research. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0055] Figure 1 It is a flow chart of a blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism of the present invention;

[0056] Figure 2 This is a schematic diagram of the method of Example 2 of the present invention. DETAILED DESCRIPTION

[0057] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0058] The present invention proposes a blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism, which is used for real-time and accurate diagnosis, early warning and operation guidance of the blast furnace smelting process.

[0059] Key points: 1. Divide furnace condition judgment into six time granularities: minute level, ten-minute level, hour level, shift level, daily average level and weekly average level. By collecting data in different time ranges, the short-, medium- and long-term change trends of furnace conditions are diagnosed respectively; 2. By building a dynamic reasoning mechanism, only a single data read is performed in the traversal of multiple rule items at each time granularity, and multiple furnace condition diagnosis processes are merged to reduce the database read and write load of massive data; 3. By recording the reasoning results of high-frequency data, memory space is released in time, thereby reducing the computational complexity of low-frequency reasoning processes.

[0060] Example 1

[0061] A blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism comprises:

[0062] Step S1: setting multiple time granularities and establishing a furnace condition diagnosis rule base according to preset rules;

[0063] Step S2: collecting data in real time and storing the data in a dynamic database, extracting data from the dynamic database at every second time granularity, performing furnace condition diagnosis at the second time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and storing the diagnosis results in a process database;

[0064] Step S3: every Nth time granularity, calling the diagnostic results of the process database and the data of the dynamic database, performing furnace condition diagnosis at the Nth time granularity according to the rules of the furnace condition diagnosis rule base, and storing the diagnostic results in the process database; N is an integer greater than 1;

[0065] Step S4: According to actual needs, retrieve the diagnostic results of different time granularities to conduct short-, medium- and long-term furnace condition evaluation.

[0066] Specifically, the step S1 includes:

[0067] Step S1.1: according to the preset rules, set the time granularity; specifically, set the time granularity in the order of time from small to large, to the first time granularity, the second time granularity ... the Nth time granularity; wherein the Nth time granularity includes the time lengths in M ​​(N-1)th time granularities, and N and M are both integers greater than 1; the time types corresponding to each time granularity include one minute level, ten minutes level, one hour level, shift level, daily average level, etc.; time granularity refers to the inference frequency;

[0068] Step S1.2: Establish a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

[0069] Specifically, step S2 includes:

[0070] Step S2.1: collect data in real time and store the collected data in a dynamic database;

[0071] Step S2.2: extracting data of a preset type from the dynamic database at every first time granularity according to the rules of the furnace condition diagnosis rule base, and storing the data in the buffer area as diagnostic data;

[0072] Step S2.3: traverse the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all furnace condition diagnoses within the first time granularity are completed, and enter step S2.4;

[0073] Step S2.4: Record the diagnostic result of the first time granularity into the process database, and release the diagnostic data in the buffer area.

[0074] Specifically, step S3 includes:

[0075] Step S3.1: At every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, the real-time data of the dynamic database and the diagnosis results of the process database are extracted respectively, and stored in the cache area as the diagnosis data of this time granularity;

[0076] Step S3.2: When the same type of furnace condition diagnosis rule corresponds to both the jth time granularity and the Nth time granularity, the diagnosis result of the jth time granularity in the process database is directly called as the diagnosis result of the Nth time granularity; j is an integer less than N;

[0077] For example, when different time granularities are used to diagnose the same type of diagnostic rules, the diagnostic results of the short time granularity can be directly called; specifically, assuming that the hydrogen content deviation of the furnace top gas needs to be diagnosed during the diagnosis process of the second time granularity and the third time granularity, then when diagnosing at the third time granularity, multiple diagnostic results of the second time granularity can be directly called without re-diagnosis, thereby reducing the amount of calculation and the read-write burden.

[0078] That is, the Nth time granularity can retrieve the diagnosis results of the 1st to (N-1)th time granularities.

[0079] Step S3.3: traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter step S3.4;

[0080] Step S3.4: Record the diagnostic result of the Nth time granularity into the process database, and release the diagnostic data in the buffer area.

[0081] Specifically, in step S3,

[0082] When performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnosis score; if the average value of the data is not within the set range, the diagnosis score is reduced; otherwise, it is increased;

[0083] Based on the obtained diagnostic scores, a unified evaluation standard is formed between different diagnostic results.

[0084] The score calculation process is called in each rule of each time granularity. For example, if there are 183 rules in 5 time granularities, 183 diagnostic scores will be calculated, which is the formula listed below. At the same time, scores for different time granularities will also be formed. For example, if there are 30 rules in this time granularity, the score of this time granularity is the weighted sum of the 30 diagnostic scores (the weight is set manually).

[0085] In a preferred embodiment, the diagnostic score is calculated as follows:

[0086] According to the mean value x of the data required by the furnace condition diagnosis rule of the Nth time granularity, the diagnosis score y is calculated:

[0087]

[0088] Among them, in the diagnosis rule corresponding to the Nth time granularity, set a min is the minimum value of the data, a is the preferred lower limit of the data, b is the preferred upper limit of the data, and b max is the maximum value of the data.

[0089] After obtaining the diagnostic score, a unified evaluation standard is formed. The diagnostic results of different time granularities can be compared and a passing score line can be defined. When the score is less than the passing score line, an alarm or mark can be issued for unified monitoring.

[0090] The present invention sets up two databases, puts the real-time collected data into the dynamic database, and puts the past diagnosis results into the process database. In the current diagnosis stage, if there are the same type of diagnosis results in the process database, the diagnosis results can be directly called; if there are no same type of diagnosis results, the data is retrieved from the dynamic database for diagnosis. The same type of diagnosis rules are only executed once, without repeated execution. When the diagnosis results are obtained, the diagnosis data cache is released, and the diagnosis round of this time granularity ends.

[0091] Example 2

[0092] The time granularity is set to a more specific time length to perform reasoning on blast furnace condition diagnosis.

[0093] Step S1: Establish a furnace condition diagnosis rule base according to preset rules and classify according to different time granularities;

[0094] Step S2: Perform furnace condition diagnosis at the minute level based on the second-level data in the dynamic database, and store the furnace condition diagnosis results at the minute level in the process database;

[0095] Among them, the second-level data here refers to the frequency of data collection;

[0096] Step S3: for furnace condition diagnosis at a time granularity of ten minutes, use the furnace condition diagnosis at a time granularity of minute and minute data in the dynamic database, and store the furnace condition diagnosis result at a time granularity of ten minutes in the process database;

[0097] Step S4: For furnace condition diagnosis at a larger time granularity (hour, shift, day, week), as in the above process, the diagnosis results of a smaller time granularity in the process database and the minute-level data in the dynamic database are repeatedly used to perform furnace condition diagnosis at this time granularity, and the furnace condition diagnosis results of this time granularity are stored in the process database;

[0098] Step S5: Call the diagnostic results stored in the process database according to the model requirements.

[0099] Specifically, the data of high-frequency calculations is saved for use in low-frequency calculations with a longer time span (otherwise, if the daily average needs to review all 24 hours of data, the amount of calculation is large, which will cause delays and performance degradation. However, since the process data is retained in the calculations every minute, every ten minutes, and every hour, the results stored in the process database can be directly called)

[0100] Generally speaking, rule items with small time granularity change frequently, such as the deviation of hydrogen content in furnace top gas. A large deviation in a short period of time may indicate a water leakage accident, but at large time granularity above the shift level, the importance of this parameter is not high. On the contrary, rule items with large time granularity change slowly, and only when they are high for a long time will they be considered to have a fault. For example, zinc load, alkali metal load, etc., as an accumulated value, only when the zinc load is high for a period of time will it have an adverse effect on the furnace condition.

[0101] See also Figure 2 , the same rule (one rule corresponds to one parameter) may span reasoning at multiple time granularities.

[0102] The furnace condition diagnosis (or furnace condition) here refers to a collection of rules. The shift-level, daily-average, and weekly-average furnace condition diagnosis refers to the overall furnace condition over the past shift, day, and week (which means that all historical values ​​over the past period of time should be considered, consisting of several rules with certain weights).

[0103] The step 1 comprises the following steps:

[0104] Step 1.1: Experts determine several rules and related data items for judging furnace conditions, such as the rule of difficult material flow, which requires the second-level data of material height, and the rule of pressure difference exceeding limit, which requires the minute-level average value of the total pressure difference of the blast furnace, etc. Such rules can be determined based on the experience of on-site experts.

[0105] Step 1.2: Based on experience, the furnace condition judgment is mapped to different time granularities. For example, the influence period of the blast furnace pressure difference is short, and usually only short-term parameter changes need to be paid attention to. Therefore, it is divided into minute-level furnace condition judgment (inference once every ten minutes) and hour-level furnace condition judgment (inference once every hour). Each furnace condition judgment rule may correspond to one or more time granularities.

[0106] Step 1.3: According to steps 1.1 and 1.2, establish a complete expert rule base for subsequent calls.

[0107] The step 2 comprises the following steps:

[0108] Step 2.1: The data items in minute-level diagnosis usually use high-frequency second-level data. The dynamic database stores them by sending 60 values ​​per minute. Therefore, minute-level diagnosis reads the latest 60 values.

[0109] Step 2.2: Perform logical judgment on the second-level data. The specific details of the logical judgment can be determined based on experience. In the practice of the present invention, taking the judgment of material difficulty as an example, it can be simply described as the material ruler height remains unchanged for 30 seconds. Since the material ruler height at the next moment cannot be known during the logical judgment every minute, the length of the head and tail difficulty of each minute is stored. When the continuous length of the head and tail of two adjacent moments exceeds 30 seconds, it can be recorded as material difficulty, thereby realizing cross-cycle logical judgment.

[0110] Step 2.3: Traverse the minute-level diagnostic rules in the expert rule base. The second-level data obtained in step 2.1 is stored in the cache for a short time. Repeat steps 2 and 2 to perform logical judgment on other diagnostic rules.

[0111] Step 2.4: Record minute-level diagnostic results into the process database.

[0112] The step 3 comprises the following steps:

[0113] Step 3.1: The data items used for the ten-minute furnace condition diagnosis usually come from the minute-level blast furnace operation data in the dynamic database and the minute-level diagnosis results stored in the process database. The data from the two sources are read separately.

[0114] Step 3.2: The data obtained in step 3.1 is temporarily stored in the cache, and the ten-minute furnace condition diagnosis rules in the expert rule base are traversed for numerical calculation. When the time granularity corresponding to the relevant rules includes both minute and ten-minute levels, the minute-level diagnosis results can be directly called without repeated calculation.

[0115] Step 3.3: Record the ten-minute furnace condition diagnosis results calculated in step 3.2 into the process database.

[0116] The step 4 comprises the following steps:

[0117] Step 4.1: The data items used for hourly furnace condition diagnosis usually come from the minute-level blast furnace operation data of the dynamic database and the minute-level or ten-minute-level diagnosis results stored in the process database, and the data from different sources are read separately.

[0118] Step 4.2: The data obtained in step 4.1 is temporarily stored in the cache, and the hourly furnace condition diagnosis rules in the expert rule base are traversed for numerical calculation. When the relevant rules correspond to multiple time granularities, the minute-level and ten-minute-level high-frequency diagnosis results can be directly called without repeated calculation.

[0119] Step 4.3: Record the hourly furnace condition diagnosis results calculated in step 4.2 into the process database.

[0120] The step 5 comprises the following steps:

[0121] Step 5.1: The data items used for shift-level, daily-average and weekly-average furnace condition diagnosis usually come from the minute-level blast furnace operation data in the dynamic database and the minute-level, ten-minute-level and hour-level diagnosis results stored in the process database, reading data from different sources respectively.

[0122] Step 5.2: The data obtained in step 3.1 is temporarily stored in the cache, and the furnace condition diagnosis rules of various time granularities in the expert rule base are traversed for numerical calculation. When the furnace condition diagnosis rules correspond to multiple time granularities, the minute-level, ten-minute-level, and hour-level high-frequency diagnosis results can be directly called without repeated calculations.

[0123] The present invention also provides a blast furnace condition diagnosis system based on a multi-time granularity dynamic reasoning mechanism. The blast furnace condition diagnosis system based on a multi-time granularity dynamic reasoning mechanism can be realized by executing the process steps of the blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism, that is, technical personnel in this field can understand the blast furnace condition diagnosis method based on a multi-time granularity dynamic reasoning mechanism as a preferred implementation of the blast furnace condition diagnosis system based on a multi-time granularity dynamic reasoning mechanism.

[0124] A blast furnace condition diagnosis system based on a multi-time granularity dynamic reasoning mechanism comprises: a module M1: setting multiple time granularities, and establishing a furnace condition diagnosis rule base according to preset rules; a module M2: collecting data in real time and storing the data in a dynamic database, extracting data from the dynamic database at every first time granularity, performing furnace condition diagnosis at the first time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and storing the diagnosis result in a process database; a module M3: calling the diagnosis result of the process database and the data of the dynamic database at every Nth time granularity, performing furnace condition diagnosis at the Nth time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and storing the diagnosis result in the process database; N is an integer greater than 1; a module M4: calling the diagnosis results of different time granularities according to actual needs.

[0125] Specifically, the module M1 includes: module M1.1: setting the time granularity according to preset rules; specifically, setting the time granularity in order from small to large, as the first time granularity, the second time granularity...the Nth time granularity; module M1.2: establishing a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and the time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

[0126] Specifically, the module M2 includes: module M2.1: real-time data collection, and storing the collected data in a dynamic database; module M2.2: at every first time granularity, according to the rules of the furnace condition diagnosis rule library, extracting preset types of data from the dynamic database, and storing them in the cache as diagnostic data; module M2.3: traversing the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnosing the furnace condition according to the diagnostic data, and obtaining the diagnostic result, until all furnace condition diagnoses within the first time granularity are completed, and entering module M2.4; module M2.4: recording the diagnostic result of the first time granularity into the process database, and releasing the diagnostic data in the cache.

[0127] Specifically, the module M3 includes: module M3.1: every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, respectively extract the data of the dynamic database and the diagnosis results of the process database, and store them in the cache area as the diagnosis data of this time granularity; module M3.2: when the same type of furnace condition diagnosis rules correspond to the jth time granularity and the Nth time granularity at the same time, directly call the diagnosis result of the jth time granularity in the process database as the diagnosis result of the Nth time granularity; j is an integer less than N; module M3.3: traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter module M3.4; module M3.4: record the diagnosis result of the Nth time granularity to the process database, and release the diagnosis data in the cache area.

[0128] Specifically, in the module M3, when performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnostic score; if the average value of the data is not within the set range, the diagnostic score is reduced; otherwise, it is increased; based on the obtained diagnostic score, a unified evaluation standard is formed between different types of diagnostic results.

[0129] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0130] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A blast furnace condition diagnosis method based on multi-time granularity dynamic reasoning mechanism, characterized in that: include: Step S1: setting multiple time granularities and establishing a furnace condition diagnosis rule base according to preset rules; Step S2: collecting data in real time and storing the data in a dynamic database, extracting data from the dynamic database at first time granularity, performing furnace condition diagnosis at the first time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and storing the diagnosis results in a process database; Step S3: every Nth time granularity, calling the diagnostic results of the process database and the data of the dynamic database, performing furnace condition diagnosis at the Nth time granularity according to the diagnostic rules of the furnace condition diagnosis rule base, and storing the diagnostic results in the process database; N is an integer greater than 1; Step S4: according to actual needs, retrieve the diagnosis results of different time granularities.

2. The blast furnace condition diagnosis method based on multi-time granularity dynamic reasoning mechanism according to claim 1 is characterized in that: The step S1 comprises: Step S1.1: according to a preset rule, set the time granularity; specifically, set the time granularity in order from small to large, to the first time granularity, the second time granularity, ... the Nth time granularity; Step S1.2: Establish a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

3. The blast furnace condition diagnosis method based on multi-time granularity dynamic reasoning mechanism according to claim 1 is characterized in that: The step S2 comprises: Step S2.1: collect data in real time and store the collected data in a dynamic database; Step S2.2: extracting data of a preset type from the dynamic database at every first time granularity according to the rules of the furnace condition diagnosis rule base, and storing the data in the buffer area as diagnostic data; Step S2.3: traverse the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all furnace condition diagnoses within the first time granularity are completed, and enter step S2.4; Step S2.4: Record the diagnostic result of the first time granularity into the process database, and release the diagnostic data in the buffer area.

4. The blast furnace condition diagnosis method based on multi-time granularity dynamic reasoning mechanism according to claim 1 is characterized in that: The step S3 comprises: Step S3.1: At every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, the data of the dynamic database and the diagnosis results of the process database are respectively extracted and stored in the cache area as the diagnosis data of this time granularity; Step S3.2: When the same type of furnace condition diagnosis rule corresponds to both the jth time granularity and the Nth time granularity, the diagnosis result of the jth time granularity in the process database is directly called as the diagnosis result of the Nth time granularity; j is an integer less than N; Step S3.3: traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result, until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter step S3.4; Step S3.4: Record the diagnostic result of the Nth time granularity into the process database, and release the diagnostic data in the buffer area.

5. The blast furnace condition diagnosis method based on multi-time granularity dynamic reasoning mechanism according to claim 1 is characterized in that: In the step S3, When performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnosis score; if the average value of the data is not within the set range, the diagnosis score is reduced; otherwise, it is increased; Based on the obtained diagnostic scores, a unified evaluation standard is formed between different diagnostic results.

6. A blast furnace condition diagnosis system based on multi-time granularity dynamic reasoning mechanism, characterized in that: include: Module M1: Set multiple time granularities and establish a furnace condition diagnosis rule base according to preset rules; Module M2: collects data in real time and stores the data in a dynamic database. At each first time granularity, extracts data from the dynamic database, performs furnace condition diagnosis at the first time granularity according to the diagnosis rules of the furnace condition diagnosis rule base, and stores the diagnosis results in the process database. Module M3: every Nth time granularity, calling the diagnostic results of the process database and the data of the dynamic database, performing furnace condition diagnosis at the Nth time granularity according to the diagnostic rules of the furnace condition diagnosis rule base, and storing the diagnostic results in the process database; N is an integer greater than 1; Module M4: retrieve diagnostic results of different time granularities according to actual needs.

7. The blast furnace condition diagnosis system based on multi-time granularity dynamic reasoning mechanism according to claim 6 is characterized in that: The module M1 comprises: Module M1.1: according to the preset rules, set the time granularity; specifically, set the time granularity in the order of time from small to large, to the first time granularity, the second time granularity, and so on, the Nth time granularity; Module M1.2: Establish a furnace condition diagnosis rule base, in which different types of furnace condition diagnosis rules are preset, and the correspondence between the furnace condition diagnosis rules and time granularity is set, so that each type of furnace condition diagnosis rule corresponds to one or more time granularities.

8. The blast furnace condition diagnosis system based on multi-time granularity dynamic reasoning mechanism according to claim 6 is characterized in that: The module M2 comprises: Module M2.1: collect data in real time and store the collected data in a dynamic database; Module M2.2: At every first time granularity, according to the rules of the furnace condition diagnosis rule base, extract data of a preset type from the dynamic database and store it in the cache as diagnostic data; Module M2.3: traverse the furnace condition diagnosis rules corresponding to the first time granularity in the furnace condition diagnosis rule library, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis result until all furnace condition diagnoses within the first time granularity are completed, and enter module M2.4; Module M2.4: Record the diagnostic results of the first time granularity into the process database and release the diagnostic data in the buffer area.

9. The blast furnace condition diagnosis system based on multi-time granularity dynamic reasoning mechanism according to claim 6 is characterized in that: The module M3 comprises: Module M3.1: Every Nth time granularity, according to the rules of the furnace condition diagnosis rule base, extract the data of the dynamic database and the diagnosis results of the process database respectively, and store them in the cache area as the diagnosis data of this time granularity; Module M3.2: When the same type of furnace condition diagnosis rules correspond to both the jth time granularity and the Nth time granularity, the diagnosis result of the jth time granularity in the process database is directly called as the diagnosis result of the Nth time granularity; j is an integer less than N; Module M3.3: Traverse the furnace condition diagnosis rules of the Nth time granularity in the furnace condition diagnosis rule base, diagnose the furnace condition according to the diagnosis data, and obtain the diagnosis results until all the furnace condition diagnoses corresponding to the Nth time granularity are completed, and enter module M3.4; Module M3.4: Record the diagnostic results of the Nth time granularity into the process database and release the diagnostic data in the buffer area.

10. The blast furnace condition diagnosis system based on multi-time granularity dynamic reasoning mechanism according to claim 6 is characterized in that: In the module M3, When performing furnace condition diagnosis at the Nth time granularity, the average value of the preset type of data in the dynamic database is calculated, and the average value is calculated by a preset algorithm to obtain a diagnosis score; if the average value of the data is not within the set range, the diagnosis score is reduced; otherwise, it is increased; Based on the obtained diagnostic scores, a unified evaluation standard is formed between different diagnostic results.

Images