Blast furnace process fault diagnosis method and device based on dynamic canonical correlation analysis

Abstract

The invention provides a blast furnace process fault diagnosis method and device based on dynamic canonical correlation analysis, and relates to the technical field of blast furnace process fault diagnosis. The method comprises the following steps: acquiring blast furnace offline data and blast furnace online data; based on a data driving method, performing data identification processing according to the blast furnace offline data and the blast furnace online data to obtain an offline zero average vector and an online zero average vector; based on a preset fault threshold value, according to the offline zero average vector and the online zero average vector, fault verification is carried out by adopting an SD-CCA method, and a residual vector based on SD-CCA and a fault detection result are obtained; and when the fault detection result is a fault, performing fault retrieval according to the SD-CCA-based residual vector based on the fault database to obtain a fault type. The method is an accurate and efficient blast furnace process fault diagnosis method based on dynamic canonical correlation analysis.

Description

Technical Field

[0001] The invention relates to the technical field of blast furnace process fault diagnosis, and in particular to a blast furnace process fault diagnosis method and device based on dynamic canonical correlation analysis. Background Art

[0002] Steel is the raw material in industry, supporting the development of industrialization and modernization. With the development of economy and society, the demand for high-quality steel and the requirements for production safety are also increasing. The steel production process can be divided into two processes: steelmaking and steel rolling. Steelmaking mainly makes raw materials into molten steel, and steel rolling mainly casts molten steel into billets and processes them into various steels through rolling mills. The blast furnace process is the core of the steelmaking process. Its emissions and energy consumption account for the largest proportion in the entire steel production process, and the production cost accounts for about 35% of the total cost. Because it is in the upstream of the production process, the quality of the output molten iron directly affects the subsequent steel rolling process. Due to the important position of the blast furnace process in the entire production process, real-time monitoring of the blast furnace process and timely detection and determination of the type of blast furnace faults are particularly important. Once the fault is not discovered and handled in time, it will cause significant losses of resources and equipment, and may even cause accidents and casualties. Therefore, designing a suitable fault diagnosis algorithm for the blast furnace system, timely detecting and locating faults, predicting potential risks, and optimizing production parameters are of great significance for the production safety and economic benefits of the blast furnace ironmaking process.

[0003] The blast furnace system is mainly composed of the blast furnace body, the feeding system, the pulverized coal injection system, the iron-making system, the hot air system and the gas treatment system. The blast furnace process is to input raw materials and fuel and blow air into the blast furnace, carry out a series of reactions in the blast furnace, and output molten iron, slag, gas and other products. Molten iron and slag are the products of long-term reactions, and there is a certain lag in judging the blast furnace condition. Since pulverized coal is solid and has poor fluidity, the pulverized coal control method also has the problem of response lag in blast furnace control. In contrast, the wind flows faster, which can not only reflect the blast furnace status more accurately, but also achieve faster control of the blast furnace status by adjusting parameters such as air volume and air temperature.

[0004] If the model of the blast furnace process is known, a model-based detection method can be used. However, the blast furnace process is a very complex dynamic system, which is a series of complex physical and chemical processes. The mechanism is difficult to analyze, and it is difficult to detect faults through an accurate mechanism model. With the development of instrumentation science, a large amount of blast furnace process data has been collected through high-quality sensors in the blast furnace, which provides data support for data-driven fault detection methods. Data-driven fault detection methods refer to methods that use real-time data and machine learning, statistical analysis and other technologies to identify anomalies or faults in equipment or systems. Unlike physical model-based fault detection methods, data-driven methods do not rely on detailed physical models of equipment, but predict and detect faults through pattern or feature changes in historical data and real-time data. This can well solve the problem of complex blast furnace system models and difficult modeling.

[0005] The diagnosis method based on multivariate statistical analysis is a data-driven method and has been widely used in the field of fault detection. This method uses multivariate statistical technology to analyze the operating data of industrial processes and identify abnormal conditions from the data for fault detection. Its core is to analyze the relationship and coordinated change pattern between multiple variables. Commonly used methods include principal component analysis (PCA), canonical correlation analysis (CCA) and partial least squares regression (PLS). The CCA-based fault diagnosis method performs fault diagnosis by analyzing the correlation between input and output. This method has no model dependence and can perform fault diagnosis on process data. However, the traditional CCA fault diagnosis method can only process static data, and its applicable condition is an industrial process running under a single stable condition. Modern industrial processes such as blast furnace processes are dynamic processes and have time-varying problems. During the operation of blast furnaces, the operating conditions of industrial processes often change due to adjustments in production demand, fluctuations in raw materials, aging of equipment, and other reasons. If only the traditional CCA fault diagnosis method is used, it is impossible to accurately capture the dynamic characteristics of the system, resulting in delayed diagnosis, and it is impossible to eliminate the impact of changes in reference variables, which seriously interferes with the judgment of blast furnace faults.

[0006] In the prior art, there is a lack of an accurate and efficient blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. Summary of the invention

[0007] In order to solve the technical problems of data dynamics and time-varying in blast furnace fault diagnosis in the prior art, the embodiment of the present invention provides a blast furnace process fault diagnosis method and device based on dynamic canonical correlation analysis. The technical solution is as follows:

[0008] On the one hand, a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis is provided, the method is implemented by a blast furnace process fault diagnosis device, and the method comprises:

[0009] Acquire blast furnace offline data and blast furnace online data; perform data recognition processing based on the blast furnace offline data and the blast furnace online data based on a data-driven method to obtain an offline zero-average vector and an online zero-average vector;

[0010] Based on a preset fault threshold, according to the offline zero mean value vector and the online zero mean value vector, a SD-CCA method is used to perform fault verification to obtain a residual vector based on SD-CCA and a fault detection result;

[0011] When the fault detection result is a fault, a fault search is performed based on the fault database and the SD-CCA-based residual vector to obtain the fault type.

[0012] On the other hand, a blast furnace process fault diagnosis device based on dynamic typical correlation analysis is provided, and the device is applied to a blast furnace process fault diagnosis method based on dynamic typical correlation analysis, and the device comprises:

[0013] A blast furnace data acquisition module is used to acquire blast furnace offline data and blast furnace online data; based on a data-driven method, data recognition processing is performed according to the blast furnace offline data and the blast furnace online data to obtain an offline zero average value vector and an online zero average value vector;

[0014] A fault detection module, configured to perform fault detection using an SD-CCA method based on a preset fault threshold, the offline zero mean value vector and the online zero mean value vector, and obtain a residual vector based on SD-CCA and a fault detection result;

[0015] The fault type retrieval module is used to perform fault retrieval based on the SD-CCA-based residual vector based on the fault database to obtain the fault type when the fault detection result is a fault.

[0016] On the other hand, a blast furnace process fault diagnosis device is provided, which includes: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the above-mentioned blast furnace process fault diagnosis methods based on dynamic canonical correlation analysis is implemented.

[0017] On the other hand, a computer-readable storage medium is provided, in which at least one instruction is stored. The at least one instruction is loaded and executed by a processor to implement any one of the above-mentioned blast furnace process fault diagnosis methods based on dynamic canonical correlation analysis.

[0018] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0019] The present invention proposes a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. The image characterization method uses offline data to identify a stable image characterization operator to eliminate the dynamics of the closed-loop blast furnace system; the SD-CCA method is used to perform fault verification, and the vector of the zero mean is converted into the SD-CCA residual to achieve fault detection. The present invention can well deal with the dynamic problem of the blast furnace system. In fault detection, the proposed method achieves better fault detection performance when facing the dynamic and time-varying problems of the blast furnace system. In view of the problem of fault classification, on the basis of the stable image characterization operator, a fault classification method based on the direction angle is used, and a fault library is established through historical fault data. When a new fault is detected, it is compared with the vector in the fault library, and the type of fault is determined according to the size of the angle. The method can accurately classify faults and realize the identification of faults in the blast furnace process. Provide guidance for subsequent fault elimination processing. The present invention is an accurate and efficient blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 It is a flow chart of a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis provided by an embodiment of the present invention;

[0022] Figure 2 It is a structural schematic diagram of a blast furnace ironmaking process provided by an embodiment of the present invention;

[0023] Figure 3 This is a schematic diagram of a fault detection result when no fault occurs provided by an embodiment of the present invention;

[0024] Figure 4 It is a schematic diagram of a fault detection result when a blast furnace pipeline travel fault occurs provided by an embodiment of the present invention;

[0025] Figure 5It is a schematic diagram of a fault detection result when a blast furnace hanging material fault occurs provided by an embodiment of the present invention;

[0026] Figure 6 It is a schematic diagram of a fault detection result when a blast furnace cooling fault occurs provided by an embodiment of the present invention;

[0027] Figure 7 It is a block diagram of a blast furnace process fault diagnosis device based on dynamic canonical correlation analysis provided by an embodiment of the present invention;

[0028] Figure 8 It is a structural schematic diagram of a blast furnace process fault diagnosis device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0030] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0031] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0032] In the embodiments of the present invention, sometimes the subscripts such as W 1 It may be written in non-subscript form such as W1. When the difference is not emphasized, the meaning is the same.

[0033] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0034] The embodiment of the present invention provides a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis, which can be implemented by a blast furnace process fault diagnosis device, which can be a terminal or a server. Figure 1 The flowchart of the blast furnace process fault diagnosis method based on dynamic canonical correlation analysis is shown in FIG. The processing flow of the method may include the following steps:

[0035] S1. Acquire blast furnace offline data and blast furnace online data; perform data recognition and processing based on the blast furnace offline data and blast furnace online data based on a data-driven method to obtain an offline zero-average vector and an online zero-average vector.

[0036] Optionally, based on a data-driven method, data recognition processing is performed according to the blast furnace offline data and the blast furnace online data to obtain an offline zero-average vector and an online zero-average vector, including:

[0037] According to the blast furnace offline data and the preset reference vector drive, the data is processed by using the data-driven method to obtain the offline zero-mean vector;

[0038] According to the blast furnace online data and the preset reference vector drive, the data is processed by using the data driven method to obtain the online zero mean value vector.

[0039] In a feasible implementation, in order to eliminate the influence of time-varying characteristics of the blast furnace system, the present invention uses offline data and a data-driven method to identify a stable image characterization operator. For the online data of the blast furnace, a data-driven method is used to obtain a zero mean value vector of the online data to eliminate the influence of time-varying characteristics.

[0040] Blast furnace system such as Figure 2 As shown, the closed-loop system is as follows (1) and (2):

[0041] (1);

[0042] (2);

[0043] Where K(z) is the controller, G(z) is the controlled object, w, u, and y are the reference input, process input, and process output of the closed-loop system, respectively. In the blast furnace process, the set value of the blast furnace parameter is the reference input w, the actual measured internal state of the blast furnace is the process input u, and the output of the blast furnace is the output y. Where w, u, and y at any time k can be written as , , In the form of a data set with a length of s, the mathematical expressions are as follows (3), (4), and (5):

[0044] (3);

[0045] (4);

[0046] (5);

[0047] Optionally, according to the blast furnace offline data and a preset reference vector drive, a data-driven method is used to perform data processing to obtain an offline zero-mean vector, including:

[0048] Using the blast furnace parameter setting value as the reference input, the actual measured internal state of the blast furnace as the process input and the output of the blast furnace as the process output, a sample Hankel matrix is ​​obtained according to the blast furnace offline data;

[0049] According to the sample Hankel matrix, factorization is performed through low-rank decomposition to obtain a stable image representation operator;

[0050] Based on the stable image characterization operator, a non-zero mean vector of the reference vector drive is obtained by performing calculation according to the preset reference vector drive;

[0051] According to the non-zero mean vector driven by the reference vector, the time-varying effects of multiple image characterization operators are eliminated to obtain an offline zero mean vector.

[0052] In a feasible implementation, in a closed system, w, u, and y are written in the form of the following Hankel matrix. The Hankel matrix form structure makes each column or row shifted in time relative to the previous column or row. This structure is particularly suitable for revealing the time dependence and dynamic characteristics in sequence data. According to the different subscripts, the Hankel matrix is ​​divided into two forms with subscripts f and p, which store past and future information respectively. Write u and w in the form of a Hankel matrix containing N samples , , , as shown in equations (6), (7), and (8):

[0053] (6);

[0054] (7);

[0055] (8);

[0056] System Input and output By reference vector This expression is called stable image representation, as shown in formula (9) where is called the stable image characterization operator:

[0057] (9);

[0058] Will The constructed vector is subjected to low-rank decomposition, and the decomposed factors are as follows (10):

[0059] (10);

[0060] in, and They are the corresponding lower triangular matrix and orthogonal matrix after low-rank decomposition respectively. and i,j,n=1,2,3 are and The block matrix of .

[0061] therefore, It can be approximately expressed by the following formula (11):

[0062] (11);

[0063] According to the definition of image representation, the stable image representation operator Approximately the following formula (12):

[0064] (12);

[0065] After identifying the stable image characterization operator, a vector with zero average value can be obtained by driving the reference vector , eliminate the influence of time variability on the system, offline zero mean vector As shown in formula (13):

[0066] (13);

[0067] Offline zero mean vector The zero average value calculation method of blast furnace online data is the same as that of offline data, which will not be repeated here.

[0068] S2. Based on the preset fault threshold, according to the offline zero mean value vector and the online zero mean value vector, the SD-CCA method is used to perform fault verification to obtain the residual vector based on SD-CCA and the fault detection result.

[0069] Optionally, based on a preset fault threshold, according to an offline zero mean value vector and an online zero mean value vector, a SD-CCA method is used to perform fault verification, and a residual vector based on SD-CCA and a fault detection result are obtained, including:

[0070] According to the offline zero mean value vector, the covariance matrix is ​​obtained by calculating through the canonical correlation analysis fault detection method;

[0071] Perform singular value decomposition on the covariance matrix to obtain a left singular vector matrix, a singular value matrix, and a right singular vector matrix;

[0072] Calculations are performed based on the left singular vector matrix, the singular value matrix, and the right singular vector matrix to obtain parameters based on SD-CCA;

[0073] Calculate based on the SD-CCA based parameters and the online zero mean vector to obtain the SD-CCA based residual vector;

[0074] Fault detection is performed according to the preset fault threshold and the residual vector based on SD-CCA to obtain the fault detection result.

[0075] In a feasible implementation, fault detection based on stable image characterization and dynamic canonical correlation analysis. According to the method of canonical correlation analysis fault detection, the covariance matrix of the residual is calculated, and various parameters are obtained by singular value decomposition, and the residual based on SD-CCA is calculated. The residual based on SD-CCA is calculated in the form of a statistic and compared with a threshold to achieve fault detection.

[0076] After eliminating the time-varying effects in the above steps, we get the vector with zero mean value offline: and According to the vector obtained from offline data, the parameters can be calculated. Refer to the steps of CCA fault detection to calculate and Covariance matrix, perform singular value decomposition on their product to obtain the left singular vector matrix Singular value matrix Right Singular Vector Matrix Three parameter matrices, in canonical correlation analysis, these matrices are used to express and explain the key linear relationship between two sets of data. The mathematical expression is as follows (14):

[0077] (14);

[0078] Where n represents the degree of freedom, , express and covariance and cross-covariance of .

[0079] Based on formula (14), the optimal parameters are: , , , , calculated as follows (15):

[0080] (15);

[0081] By online zero mean vector and As well as the four optimal parameters, the residuals based on SD-CCA can be calculated , The vectors are as follows (16) and (17):

[0082] (16);

[0083] (17);

[0084] The SD-CCA residual vector is written as The form of the statistics is as follows (18) and (19):

[0085] (18);

[0086] (19);

[0087] in , Residual vector , Covariance of the vector.

[0088] According to the set confidence level , degrees of freedom n and sample size N, the threshold can be calculated using the F distribution , as shown in formula (20):

[0089] (20);

[0090] The statistics and threshold By comparison, fault detection can be performed. The mathematical expression of this process is as follows (21):

[0091] (twenty one);

[0092] Among them, the SD-CCA method refers to a method combining signal detection threshold and canonical correlation analysis.

[0093] In a feasible implementation, the image representation-based dynamic canonical correlation analysis method (Stable Image Representation Aided Dynamic CCA, SD-CCA) is a method that combines signal detection threshold and canonical correlation analysis for channel state assessment and fault detection in a dynamic environment. This method enhances the interpretability of canonical variables by introducing sparsity in feature extraction. By extracting sparse features in a dynamic process, the ability to explore the relationship between input and output is enhanced. This method also shows its superiority in fields such as industrial process monitoring.

[0094] S3. When the fault detection result is a fault, based on the fault database, a fault search is performed according to the residual vector based on SD-CCA to obtain the fault type.

[0095] Optionally, when the fault detection result is a fault, based on the fault database, a fault search is performed according to the residual vector based on SD-CCA to obtain the fault type, including:

[0096] According to the fault database, all types of fault vectors are obtained;

[0097] Based on all types of fault vectors, the residual vector based on SD-CCA is calculated to obtain the vector angle data set;

[0098] Select the minimum value in the vector angle data set as the fault vector angle;

[0099] Based on the fault database, the fault type is determined according to the fault vector angle.

[0100] In a feasible implementation, the detected fault is separated based on stable image characterization and dynamic canonical correlation analysis. This method belongs to the fault isolation method based on residual direction vector. The historical fault data is collected to build a fault library, and each vector in the fault library represents a fault type. When a fault is detected, according to the angle between the fault library and the fault residual vector, if the angle with a vector in the fault library is the smallest, it is determined to be the fault type.

[0101] The detected fault residual vector Compare with each fault type, that is, each column in the fault library B. They all represent the direction vector of a fault type. With a certain fault type When the direction angle j is the smallest, the newly detected fault is considered to be of this fault type. The calculation formula of this process is as follows (22):

[0102] (twenty two);

[0103] The fault database is constructed based on historical fault data; each vector in the fault database represents a fault type.

[0104] In a feasible implementation manner, in the present invention, different faults need to be collected in advance, based on the SD-CCA residual vector Can be used and Simplified expression, further written as the following formula (23):

[0105] (twenty three);

[0106] in, , .

[0107] For each fault sample , measured value Fault Items and the noise term Two parts, formula (23) can be transformed into the following formula (24):

[0108] (twenty four);

[0109] in, It is expressed as the FS direction at time k, assuming that the residual vectors of different faults are not collinear.

[0110] For each fault type The sum of the fault samples is , and obtain the direction vector of each fault type. fault types, forming a fault library B. Each column in the fault library It corresponds to the direction of each fault type. Finally, the fault database B can be obtained as follows (25):

[0111] (25)

[0112] In a feasible implementation manner, the present invention takes the furnace condition of a blast furnace of Liuzhou Iron and Steel as an example, sets controllable standard wind speed, oxygen-enriched flow, cold air pressure, and hot air temperature as reference inputs, uses the measured blast kinetic energy, total pressure difference, and actual wind speed inside the blast furnace as process inputs, uses the bosh gas volume and bosh gas index of the blast furnace output gas as process outputs, and uses the data of its normal operation as offline data. The blast furnace data is sampled approximately every 10 seconds.

[0113] There are three typical failure scenarios in the blast furnace process. The first failure is a pipeline travel failure, which means that the air permeability of a certain area of ​​the blast furnace charge column on the cross section is particularly strong, causing the gas flow in this area to develop abnormally like in a pipeline. At this time, the wind pressure will gradually decrease and the air volume will gradually increase. The second failure is a suspended material failure, which means that the blast furnace charge stops falling. At this time, the wind pressure will increase, the air volume will decrease, and the permeability index will decrease. The third failure is a furnace cooling failure, which means that the blast furnace temperature drops sharply, which is specifically manifested as unstable air volume and air pressure, and low gas pressure and temperature on the top of the furnace. In a blast furnace of Liugang, 2000, 1000, and 2500 fault samples were collected respectively.

[0114] When no fault occurs, the test results are as follows: Figure 3 As shown in Figure 2, we can see that the statistics are basically below the curve. The results of the three types of fault detection are shown in Figure 2. Figure 4 , Figure 5 , Figure 6 As shown. The false alarm rate is set to 0.01. It can be seen that this fault detection method can detect faults in a timely manner. After the fault is detected, the residuals of the 1000 fault samples are classified, and the results shown in Table 1 (fault classification result table) can be obtained. It can be seen that the three faults can be well classified, and the classification success rate reaches 99.8%.

[0115] Table 1

[0116]

[0117] The present invention proposes a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. The image characterization method uses offline data to identify a stable image characterization operator to eliminate the dynamics of the closed-loop blast furnace system; the SD-CCA method is used to perform fault verification, and the vector of the zero mean is converted into the SD-CCA residual to achieve fault detection. The present invention can well deal with the dynamic problem of the blast furnace system. In fault detection, the proposed method achieves better fault detection performance when facing the dynamic and time-varying problems of the blast furnace system. In view of the problem of fault classification, on the basis of the stable image characterization operator, a fault classification method based on the direction angle is used, and a fault library is established through historical fault data. When a new fault is detected, it is compared with the vector in the fault library, and the type of fault is determined according to the size of the angle. The method can accurately classify faults and realize the identification of faults in the blast furnace process. Provide guidance for subsequent fault elimination processing. The present invention is an accurate and efficient blast furnace process fault diagnosis method based on dynamic canonical correlation analysis.

[0118] Figure 7 The block diagram of a blast furnace process fault diagnosis device based on dynamic canonical correlation analysis according to an exemplary embodiment is shown. The device is used for a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. Figure 7 The device includes a blast furnace data acquisition module 710, a fault detection module 720 and a fault type retrieval module 730. Among them:

[0119] The blast furnace data acquisition module 710 is used to acquire the blast furnace offline data and the blast furnace online data; based on the data-driven method, data recognition processing is performed according to the blast furnace offline data and the blast furnace online data to obtain the offline zero average value vector and the online zero average value vector;

[0120] A fault detection module 720 is used to perform fault detection using an SD-CCA method based on a preset fault threshold, an offline zero mean value vector and an online zero mean value vector, and obtain a residual vector based on SD-CCA and a fault detection result;

[0121] The fault type retrieval module 730 is used to perform fault retrieval based on the residual vector based on SD-CCA based on the fault database to obtain the fault type when the fault detection result is a fault.

[0122] Optionally, the blast furnace data acquisition module 710 is further used to:

[0123] According to the blast furnace offline data and the preset reference vector drive, the data is processed by using the data-driven method to obtain the offline zero-mean vector;

[0124] According to the blast furnace online data and the preset reference vector drive, the data is processed by using the data driven method to obtain the online zero mean value vector.

[0125] Optionally, the blast furnace data acquisition module 710 is further used to:

[0126] Using the blast furnace parameter setting value as the reference input, the actual measured internal state of the blast furnace as the process input and the output of the blast furnace as the process output, a sample Hankel matrix is ​​obtained according to the blast furnace offline data;

[0127] According to the sample Hankel matrix, factorization is performed through low-rank decomposition to obtain a stable image representation operator;

[0128] Based on the stable image characterization operator, a non-zero mean vector of the reference vector drive is obtained by performing calculation according to the preset reference vector drive;

[0129] According to the non-zero mean vector driven by the reference vector, the time-varying effects of multiple image characterization operators are eliminated to obtain an offline zero mean vector.

[0130] Optionally, the fault detection module 720 is further configured to

[0131] According to the offline zero mean value vector, the covariance matrix is ​​obtained by calculating through the canonical correlation analysis fault detection method;

[0132] Perform singular value decomposition on the covariance matrix to obtain a left singular vector matrix, a singular value matrix, and a right singular vector matrix;

[0133] Calculations are performed based on the left singular vector matrix, the singular value matrix, and the right singular vector matrix to obtain parameters based on SD-CCA;

[0134] Calculate based on the SD-CCA based parameters and the online zero mean vector to obtain the SD-CCA based residual vector;

[0135] Fault detection is performed according to the preset fault threshold and the residual vector based on SD-CCA to obtain the fault detection result.

[0136] Among them, the SD-CCA method refers to a method combining signal detection threshold and canonical correlation analysis.

[0137] Optionally, the fault type retrieval module 730 is further configured to:

[0138] According to the fault database, all types of fault vectors are obtained;

[0139] Based on all types of fault vectors, the residual vector based on SD-CCA is calculated to obtain the vector angle data set;

[0140] Select the minimum value in the vector angle data set as the fault vector angle;

[0141] Based on the fault database, the fault type is determined according to the fault vector angle.

[0142] The fault database is constructed based on historical fault data; each vector in the fault database represents a fault type.

[0143] The present invention proposes a blast furnace process fault diagnosis method based on dynamic canonical correlation analysis. The image characterization method uses offline data to identify a stable image characterization operator to eliminate the dynamics of the closed-loop blast furnace system; the SD-CCA method is used to perform fault verification, and the vector of the zero mean is converted into the SD-CCA residual to achieve fault detection. The present invention can well deal with the dynamic problem of the blast furnace system. In fault detection, the proposed method achieves better fault detection performance when facing the dynamic and time-varying problems of the blast furnace system. In view of the problem of fault classification, on the basis of the stable image characterization operator, a fault classification method based on the direction angle is used, and a fault library is established through historical fault data. When a new fault is detected, it is compared with the vector in the fault library, and the type of fault is determined according to the size of the angle. The method can accurately classify faults and realize the identification of faults in the blast furnace process. Provide guidance for subsequent fault elimination processing. The present invention is an accurate and efficient blast furnace process fault diagnosis method based on dynamic canonical correlation analysis.

[0144] Figure 8 is a schematic diagram of the structure of a blast furnace process fault diagnosis device provided by an embodiment of the present invention, such as Figure 8 As shown, the blast furnace process fault diagnosis equipment may include the above Figure 7 The blast furnace process fault diagnosis device based on dynamic canonical correlation analysis is shown. Optionally, the blast furnace process fault diagnosis device 810 may include a first processor 2001 .

[0145] Optionally, the blast furnace process fault diagnosis device 810 may further include a memory 2002 and a transceiver 2003 .

[0146] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0147] Combine the following Figure 8 The various components of the blast furnace process fault diagnosis equipment 810 are introduced in detail:

[0148] The first processor 2001 is the control center of the blast furnace process fault diagnosis device 810, and can be a processor or a general term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiment of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0149] Optionally, the first processor 2001 can perform various functions of the blast furnace process fault diagnosis device 810 by running or executing a software program stored in the memory 2002 and calling data stored in the memory 2002.

[0150] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 8 CPU0 and CPU1 are shown in FIG.

[0151] In a specific implementation, as an embodiment, the blast furnace process fault diagnosis device 810 may also include multiple processors, such as Figure 8 The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0152] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled to be executed by the first processor 2001. The specific implementation method can refer to the above method embodiment, which will not be repeated here.

[0153] Optionally, the memory 2002 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001, or may exist independently and access the first processor 2001 through the interface circuit ( Figure 8 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0154] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0155] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 8 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0156] Optionally, the transceiver 2003 may be integrated with the first processor 2001, or may exist independently and communicate with the first processor 2001 through the interface circuit ( Figure 8 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0157] It should be noted that Figure 8 The structure of the blast furnace process fault diagnosis device 810 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure identification device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0158] In addition, the technical effects of the blast furnace process fault diagnosis device 810 can refer to the technical effects of the blast furnace process fault diagnosis method based on dynamic canonical correlation analysis described in the above method embodiment, and will not be repeated here.

[0159] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0160] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0161] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0162] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0163] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0164] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0165] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0166] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0167] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0170] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0171] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A blast furnace process fault diagnosis method based on dynamic canonical correlation analysis, characterized in that: The method comprises: Acquire blast furnace offline data and blast furnace online data; perform data recognition processing based on the blast furnace offline data and the blast furnace online data based on a data-driven method to obtain an offline zero-average vector and an online zero-average vector; Based on a preset fault threshold, according to the offline zero mean value vector and the online zero mean value vector, a SD-CCA method is used to perform fault verification to obtain a residual vector based on SD-CCA and a fault detection result; When the fault detection result is a fault, a fault search is performed based on the SD-CCA-based residual vector based on the fault database to obtain the fault type.

2. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 1 is characterized in that: The data-driven method performs data recognition processing according to the blast furnace offline data and the blast furnace online data to obtain an offline zero-average vector and an online zero-average vector, including: According to the blast furnace offline data and a preset reference vector drive, a data-driven method is used to perform data processing to obtain an offline zero-average vector; According to the blast furnace online data and preset reference vector drive, a data-driven method is used to perform data processing to obtain an online zero-average value vector.

3. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 2 is characterized in that: The method of driving according to the blast furnace offline data and a preset reference vector, using a data-driven method to process data, and obtaining an offline zero-average vector, includes: Using the blast furnace parameter setting value as a reference input, the actually measured internal state of the blast furnace as a process input and the output of the blast furnace as a process output, and obtaining a sample Hankel matrix according to the blast furnace offline data; According to the sample Hankel matrix, factorization is performed through low-rank decomposition to obtain a stable image characterization operator; Based on the stable image characterization operator, a calculation is performed according to a preset reference vector drive to obtain a non-zero mean vector of the reference vector drive; According to the non-zero mean vector driven by the reference vector, time-varying effects are eliminated on the multiple image characterization operators to obtain an offline zero mean vector.

4. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 1 is characterized in that: The method of performing fault detection based on the preset fault threshold and the offline zero mean value vector and the online zero mean value vector by using the SD-CCA method to obtain the residual vector and the fault detection result based on the SD-CCA includes: According to the offline zero mean value vector, a covariance matrix is ​​obtained by calculating through a method of canonical correlation analysis fault detection; Performing singular value decomposition on the covariance matrix to obtain a left singular vector matrix, a singular value matrix, and a right singular vector matrix; Calculating according to the left singular vector matrix, the singular value matrix and the right singular vector matrix to obtain parameters based on SD-CCA; Calculating according to the SD-CCA-based parameters and the online zero mean value vector to obtain a SD-CCA-based residual vector; Fault detection is performed according to a preset fault threshold and the residual vector based on SD-CCA to obtain a fault detection result.

5. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 4 is characterized in that: The SD-CCA method refers to a method combining signal detection threshold and canonical correlation analysis.

6. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 1 is characterized in that: When the fault detection result is a fault, based on the fault database, a fault search is performed according to the residual vector based on SD-CCA to obtain the fault type, including: According to the fault database, all types of fault vectors are obtained; Based on the fault vectors of all types, calculation is performed according to the residual vector based on SD-CCA to obtain a vector angle data set; Selecting the minimum value in the vector angle data set as the fault vector angle; Based on the fault database, a fault type is determined according to the fault vector angle.

7. The blast furnace process fault diagnosis method based on dynamic canonical correlation analysis according to claim 6 is characterized in that: The fault database is constructed based on historical fault data; each vector in the fault database represents a fault type.

8. A blast furnace process fault diagnosis device based on dynamic canonical correlation analysis, the blast furnace process fault diagnosis device based on dynamic canonical correlation analysis is used to implement the blast furnace process fault diagnosis method based on dynamic canonical correlation analysis as claimed in any one of claims 1 to 7, characterized in that: The device comprises: A blast furnace data acquisition module is used to acquire blast furnace offline data and blast furnace online data; based on a data-driven method, data recognition processing is performed according to the blast furnace offline data and the blast furnace online data to obtain an offline zero average value vector and an online zero average value vector; A fault detection module, configured to perform fault detection using an SD-CCA method based on a preset fault threshold, the offline zero mean value vector and the online zero mean value vector, and obtain a residual vector based on SD-CCA and a fault detection result; The fault type retrieval module is used to perform fault retrieval based on the SD-CCA-based residual vector based on the fault database to obtain the fault type when the fault detection result is a fault.

9. A blast furnace process fault diagnosis device, characterized in that: The blast furnace process fault diagnosis equipment comprises: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, which can be called by a processor to execute the method according to any one of claims 1 to 7.

Images