Data monitoring-based intelligent supervision system and method for abnormal data

Abstract

The present invention relates to the technical field of data analysis, and in particular to a data monitoring-based intelligent supervision system and method for abnormal data. The system comprises a data transceiving module, a matrix processing module, a feature parsing module, an anomaly localization module and a part scaling module; the data transceiving module is configured to send and receive matrix data; the matrix processing module is configured to fit a composite matrix and output a processed matrix; the feature parsing module is configured to determine whether the final processed matrix has a data processing error or not; the anomaly localization module is configured to locate a node device in which an error step has occurred; and the part scaling module is configured to scale a processed matrix of an abnormal device. The present invention can find abnormal data, and locate an abnormal device, thereby reducing the effect of data anomalies on production processes and device performance, helping industrial control systems to discover and solve problems in time, and improving the production efficiency.

Description

An abnormal data intelligent supervision system and method based on data monitoring Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a system and method for intelligently supervising abnormal data based on data monitoring. Background Art

[0002] Intelligent industrial control refers to technologies and methods that achieve automation and optimization in industrial production through digital monitoring and adjustment of equipment, systems, or processes. These technologies encompass the control, measurement, and adjustment of various production equipment. The process of intelligent industrial control is as follows: After industrial control probes acquire data, the data is sent to each node device in a matrix format according to a process flow. The final processed matrix is ​​then executed at the industrial control terminal.

[0003] Since the underlying operations of industrial control systems when performing data processing are all performed in a matrix manner, the process of node device processing data can be expressed as the process of multiplying the input data by the processing matrix to obtain the output matrix. However, in a mature intelligent industrial control system, there are often dozens or even hundreds of node devices. Errors can easily occur during the matrix processing process, resulting in data misalignment, data loss, or garbled data.

[0004] However, due to the linear nature of data processing, information can only be transmitted in a sequential, one-way manner. Most node devices only process and transmit data and cannot identify anomalies in input data, leading to confusion in the execution of instructions by industrial control terminals. Conventional anomaly monitoring systems reduce the probability of accidental errors through multiple verifications, but lack effective management methods for anomalies caused by device failures. They are unable to locate and troubleshoot the node devices causing the anomaly, nor can they mitigate the impact of abnormal data. Summary of the Invention

[0005] The purpose of the present invention is to provide a system and method for intelligent supervision of abnormal data based on data monitoring to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an abnormal data intelligent supervision system based on data monitoring, comprising: a data transceiver module, a matrix processing module, a feature analysis module, an abnormality positioning module and a component scaling module;

[0007] The data transceiver module is used to extract matrix features, send and receive matrix data between node devices, and extract matrix row and column widths;

[0008] The matrix processing module is used to perform a multiplication operation on the data matrix and the node device characteristic matrix after receiving the data matrix sent by the previous process to obtain a composite matrix, and complete the node device's own processing process on the composite matrix;

[0009] The feature analysis module is used to analyze the eigenvalues ​​and eigenvectors of the matrix after the industrial control terminal receives the final data matrix, and perform step-by-step inverse operations on the data matrix according to the characteristic matrix of each node device to obtain a multi-dimensional transfer matrix. Then, by calculating the eigenvalues ​​of the transfer matrix, it is determined whether there is a data processing error. If no error exists, the final transfer matrix is ​​input to the execution terminal for execution;

[0010] The abnormality locating module is used to locate the node device where the error step occurred according to the number of inverse operations on the data matrix when a data processing error is found, calculate the deviation value of the device, add an abnormality record of the device, and notify the maintenance personnel of the number of the abnormal device;

[0011] The component scaling module is used to calculate the degradation coefficient of the abnormal device according to the deviation value and abnormal frequency of the abnormal device, and scale the processing matrix of the device according to the degradation coefficient to reduce the impact of the device abnormality on the final result.

[0012] Furthermore, the data transceiver module includes: a matrix extraction unit and a signal transceiver unit;

[0013] The matrix extraction unit is used to obtain the initial matrix through the industrial control probe and extract the row and column widths of the matrix;

[0014] The signal transceiver unit is used to receive the data matrix sent by the previous device and send the processed matrix information to the next device.

[0015] Furthermore, the matrix processing module includes: a feature fitting unit and an engineering operation unit;

[0016] The feature fitting unit is used to, after receiving the data matrix, multiply the device's own feature matrix by the data matrix to fit the device features into the data matrix to obtain a composite matrix;

[0017] The engineering operation unit is used to take the composite matrix as an input matrix, input the node device to perform data processing, and output the processed data matrix.

[0018] Furthermore, the feature analysis module includes: an inverse feature unit, a solution fitting unit and an abnormality judgment unit;

[0019] The inverse characteristic unit is used to calculate the inverse matrix of the characteristic matrix of all node devices, and calculate the eigenvalues ​​and eigenvectors of these inverse matrices;

[0020] The de-fitting unit is used to calculate the eigenvalues ​​of the data matrix after the industrial control terminal receives the final data matrix, and multiply the eigenvalues ​​of the data matrix with the inverse matrix of the characteristic matrix of each node device one by one to obtain multiple restored matrices;

[0021] The abnormality judgment unit is used to calculate the eigenvalue of the reduction matrix every time a reduction matrix is ​​fitted, compare the eigenvalue of the reduction matrix, the eigenvalue of the characteristic matrix and the eigenvalue of the previous reduction matrix, and judge whether there is an abnormality in the processing process.

[0022] Furthermore, the anomaly location module includes: an inversion analysis unit, a feature recovery unit and a deviation recording unit;

[0023] The inverter analysis unit is used to determine the number of the abnormal device according to the number of restoration steps of the restoration matrix after discovering the abnormality, and notify the maintenance personnel of the number of the abnormal device;

[0024] The feature recovery unit is used to request the previous node device of the abnormal device to send a data matrix, continue to analyze the data matrix, and determine the abnormal conditions of other devices;

[0025] The deviation recording unit is used to add an abnormal record of an abnormal device and calculate the deviation coefficient of the abnormal device according to the eigenvalue of the restoration matrix.

[0026] Furthermore, the component scaling module includes: a degradation calculation unit and a simplification processing unit;

[0027] The degradation calculation unit is used to calculate the degradation matrix of the device according to the abnormal frequency and deviation coefficient of the abnormal device;

[0028] The simplified processing unit is used to multiply the matrix to be input into the abnormal device with the degradation matrix. After the node device obtains the final data matrix, it is multiplied by the inverse matrix of the degradation matrix to restore the data of the final data matrix, thereby reducing the impact of the abnormal device on the final result before the abnormal device is repaired.

[0029] A method for intelligent supervision of abnormal data based on data monitoring, comprising the following steps:

[0030] S100. All node devices are numbered in the order of data processing flow, and the industrial control terminal assigns a unique characteristic matrix to each node device and calculates the inverse matrix of these characteristic matrices;

[0031] S200. After the industrial control probe obtains the initial data, it sends the matrix recording the initial data to the first node device according to the node device number. After the first node device receives the initial data matrix, it multiplies its own characteristic matrix by the initial data matrix to obtain a composite matrix. The composite matrix is ​​input as an input matrix to the device for processing and outputs a processed matrix.

[0032] S300. The first node device sends the first processing matrix to the second node device in order of number. The second node device uses the first processing matrix as the initial data matrix and repeats step S200 to obtain the second processing matrix. This process is repeated in this way until all node devices have completed the processing and obtain the final processing matrix.

[0033] S400. The control terminal receives the final processing matrix, calculates the eigenvalue of the final processing matrix, and then multiplies the final processing matrix by the inverse matrix of the characteristic matrix of each node device in reverse order of the number to obtain multiple reduction matrices. All reduction matrices are arranged in ascending order of the number of reductions, and the ratio of the eigenvalue of each reduction matrix to the eigenvalue of the previous reduction matrix is ​​calculated;

[0034] S500. Determine whether the data is abnormal based on the ratio and the eigenvalue of the characteristic matrix of each node device. When the data is determined to be abnormal, give the number of the abnormal device and calculate the degradation coefficient of the abnormal device. According to the degradation coefficient, obtain the degradation matrix, and then use the degradation matrix to scale the data processing process of the abnormal device.

[0035] Furthermore, step S100 includes:

[0036] Step S101. Number all node devices in the order of data processing, and record the numbering result as data set X, where X = {1, 2, ..., n}, where n is the number of node devices, n is an integer and n ≥ 2;

[0037] Step S102. The industrial control terminal generates n different characteristic matrices, wherein the characteristic matrices are reversible matrices, and the eigenvalues ​​of all characteristic matrices are different;

[0038] The feature matrix is ​​assigned to each node device, and the assignment result is recorded as data set Y, Y={A1, A2, ..., A n}, the dataset Y and dataset X form a mapping relationship, A1, A2 and A n They represent the feature matrices assigned to node devices numbered 1, 2, and n, respectively, and the mapping relationships of the remaining elements are similar;

[0039] Step S103. The industrial control terminal calculates the inverse matrix of each characteristic matrix, and the calculation results form the data set Z, Z={A1 -1 , A2 -1 ,…,An -1}, the dataset Z and dataset Y form a mapping relationship, A1 -1 , A2 -1 , and A n -1 Represent the feature matrices A1, A2 and A respectively n The inverse matrix of , and the mapping relationship of the remaining elements is similar;

[0040] Step S104. Calculate the eigenvalues ​​of each characteristic matrix, and record the calculation results as data set R, R={r1, r2, ..., r n}, the dataset R and the dataset Y form a mapping relationship, r1, r2, and r n Represent the feature matrices A1, A2 and A respectively n The mapping relationships of the remaining elements are similar.

[0041] Furthermore, step S200 includes:

[0042] Step S201: The industrial control probe obtains initial data and sends the initial data in the form of a matrix to the node device numbered 1. The initial data matrix is ​​denoted as Q.

[0043] Step S202: After receiving the initial data matrix, node device 1 performs operation T1=Q·A1, where T1 is the composite matrix of node device 1. T1 is used as the input matrix and the operation is performed according to the operation program of node device 1 itself to obtain a processing matrix Q1.

[0044] This step performs feature matrix fitting at the input end, effectively avoiding the situation where feature fitting fails due to equipment failure.

[0045] Furthermore, step S300 includes:

[0046] Step S301. Node device 1 sends the primary processing matrix to node device 2. Node device 2 performs the operation T2 = Q1·A2, where T2 is the composite matrix of node device 2. Node device 2 uses T2 as the input matrix and performs the operation according to its own operation program to obtain the secondary processing matrix Q2.

[0047] Step S302: Repeat the above steps until node device n completes matrix processing and obtains the final processing matrix Q. n , the matrix Q n Send to the industrial control terminal.

[0048] The design of data verification at the industrial control terminal of the present invention also ensures that all commands submitted for execution are authentic and valid, avoids the occurrence of command confusion and error reporting, helps the industrial control system to discover and solve problems in a timely manner, improves production efficiency, and reduces abnormal risks.

[0049] Furthermore, step S400 includes:

[0050] Step S401. The industrial control terminal receives Q n Then, calculate Q n The eigenvalue t n+1 , and then perform the following operations to calculate the restored matrix: ;

[0051] Among them, P m Represents the reduction matrix corresponding to the node device numbered m, where m∈X, Represents the ni-th element in the data set Z;

[0052] Step S402. Calculate the eigenvalues ​​of each restored matrix, and record the calculation results as data set T, T={t1,t2,…t m ,…,t n}, where t m Represents the restored matrix P m The corresponding eigenvalues;

[0053] Step S403: Verify the accuracy of the eigenvalue processing according to the following formula:

[0054] Among them, N m Represents the status parameters of node device m, the N m ∈{0,1}, IFS is a logical judgment function, is the judgment formula in the logical judgment function. If the judgment formula is established, then N m =1, if the judgment is not true, then N m =0;

[0055] Generate a verification sequence K in reverse order of numbering, where K=[n,n-1,…,1], and substitute the elements in sequence K into m in the above formula in turn to obtain the result state sequence L, where L=[N n ,N n-1 ,…,N1], the elements in the sequence L correspond one to one with the elements in the sequence K;

[0056] Step S404. Determine in order whether the elements in the sequence L are 1. If all the elements in the sequence L are 1, it is determined that there is no abnormality in the data, and the matrix P is restored. n It is sent as the final execution matrix to the execution terminal for execution;

[0057] When it is found that the value of an element in the sequence L is 0, it is determined that there is an abnormality in the data, and the position of this element in the sequence L is recorded as g, where g∈X. The node device numbered g is marked as an abnormal device, the maintenance personnel are notified to repair the device g, and an abnormal record of the node device g is added, and go to step S500.

[0058] Furthermore, step S500 includes:

[0059] Step S501. Calculate the deviation value e of the node device g, , where t g+1 Represents the restored matrix eigenvalue corresponding to the node device numbered g+1, r g Represents the eigenvalue of the characteristic matrix of node device g;

[0060] Obtain the historical abnormality records of node device g. Within the preset time range T, the total number of records is recorded as b, and the abnormality frequency of node device g is V=b / T;

[0061] Step S502: Calculate the degradation coefficient J of node device g according to the following formula: ;

[0062] Among them, V0 is the system's preset fault-tolerant frequency;

[0063] Step S503: Calculate the degradation matrix W, where W = J·E·A g , where E is the identity matrix, A g is the characteristic matrix of node device g, and the restored degradation matrix W -1 =1 / J·A g -1 , where Ag -1 is the inverse matrix of the device g feature matrix;

[0064] Before the abnormal device g is repaired, the characteristic matrix of the node device g is replaced by W and the characteristic matrix of the node device g is replaced by W. -1 Replace the inverse matrix of the characteristic matrix of node device g, and in the subsequent execution of step S404, let N g =1;

[0065] The processed restored matrix P n It is sent as the final execution matrix to the execution terminal for execution.

[0066] The present invention can scale the processing of abnormal equipment according to the historical processing conditions and data deviation values ​​of the abnormal equipment, thereby reducing the impact of abnormal node equipment on production technology and equipment performance.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] The present invention can provide an operational identity feature for each data processing device, determine whether there is an anomaly in the data matrix at the industrial control terminal, and locate the device with the anomaly. By checking the abnormal data, the abnormal device is discovered and located, and maintenance personnel are notified to investigate the cause of the anomaly. The operation problems of the equipment are discovered in time and corresponding maintenance and optimization measures are taken, thereby reducing the impact of data processing anomalies on the production process and equipment performance.

[0069] The present invention does not require step-by-step calculations on the data, nor does it require the installation of data accounting devices on the node devices. The hardware requirements are lower, and the design of data verification at the industrial control terminal also ensures that all commands submitted for execution are true and valid, avoiding the occurrence of instruction confusion and error reporting, helping the industrial control system to discover and solve problems in a timely manner, improve production efficiency, and reduce abnormal risks.

[0070] The present invention can scale the processing of abnormal equipment based on its historical processing conditions, thereby reducing the impact of abnormal node equipment on production processes and equipment performance. It can also analyze and mine data in industrial control systems and take corresponding improvement measures to improve production efficiency and equipment utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0072] FIG1 is a schematic structural diagram of an abnormal data intelligent supervision system based on data monitoring according to the present invention;

[0073] FIG2 is a schematic diagram of the steps of a method for intelligently supervising abnormal data based on data monitoring according to the present invention. DETAILED DESCRIPTION

[0074] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0075] Please refer to FIG1 . The present invention provides a technical solution: an abnormal data intelligent supervision system based on data monitoring, comprising: a data transceiver module, a matrix processing module, a feature analysis module, an abnormality location module, and a component scaling module;

[0076] The data transceiver module is used to extract matrix features, send and receive matrix data between node devices, and extract matrix row and column widths;

[0077] The data transceiver module includes: a matrix extraction unit and a signal transceiver unit;

[0078] The matrix extraction unit is used to obtain the initial matrix through the industrial control probe and extract the row and column widths of the matrix;

[0079] The signal transceiver unit is used to receive the data matrix sent by the previous device and send the processed matrix information to the next device.

[0080] The matrix processing module is used to perform a multiplication operation on the data matrix and the node device characteristic matrix after receiving the data matrix sent by the previous process to obtain a composite matrix, and complete the node device's own processing process on the composite matrix;

[0081] The matrix processing module includes: a feature fitting unit and an engineering operation unit;

[0082] The feature fitting unit is used to, after receiving the data matrix, multiply the device's own feature matrix by the data matrix to fit the device features into the data matrix to obtain a composite matrix;

[0083] The engineering operation unit is used to take the composite matrix as an input matrix, input the node device to perform data processing, and output the processed data matrix.

[0084] The feature analysis module is used to analyze the eigenvalues ​​and eigenvectors of the matrix after the industrial control terminal receives the final data matrix, and perform step-by-step inverse operations on the data matrix according to the characteristic matrix of each node device to obtain a multi-dimensional transfer matrix. Then, by calculating the eigenvalues ​​of the transfer matrix, it is determined whether there is a data processing error. If no error exists, the final transfer matrix is ​​input to the execution terminal for execution;

[0085] The feature analysis module includes: an inversion feature unit, a solution fitting unit and an abnormality judgment unit;

[0086] The inverse characteristic unit is used to calculate the inverse matrix of the characteristic matrix of all node devices, and calculate the eigenvalues ​​and eigenvectors of these inverse matrices;

[0087] The de-fitting unit is used to calculate the eigenvalues ​​of the data matrix after the industrial control terminal receives the final data matrix, and multiply the eigenvalues ​​of the data matrix with the inverse matrix of the characteristic matrix of each node device one by one to obtain multiple restored matrices;

[0088] The abnormality judgment unit is used to calculate the eigenvalue of the reduction matrix every time a reduction matrix is ​​fitted, compare the eigenvalue of the reduction matrix, the eigenvalue of the characteristic matrix and the eigenvalue of the previous reduction matrix, and judge whether there is an abnormality in the processing process.

[0089] The abnormality locating module is used to locate the node device where the error step occurred according to the number of inverse operations on the data matrix when a data processing error is found, calculate the deviation value of the device, add an abnormality record of the device, and notify the maintenance personnel of the number of the abnormal device;

[0090] The anomaly location module includes: an inversion analysis unit, a feature recovery unit and a deviation recording unit;

[0091] The inverter analysis unit is used to determine the number of the abnormal device according to the number of restoration steps of the restoration matrix after discovering the abnormality, and notify the maintenance personnel of the number of the abnormal device;

[0092] The feature recovery unit is used to request the previous node device of the abnormal device to send a data matrix, continue to analyze the data matrix, and determine the abnormal conditions of other devices;

[0093] The deviation recording unit is used to add an abnormal record of an abnormal device and calculate the deviation coefficient of the abnormal device according to the eigenvalue of the restoration matrix.

[0094] The component scaling module is used to calculate the degradation coefficient of the abnormal device according to the deviation value and abnormal frequency of the abnormal device, and scale the processing matrix of the device according to the degradation coefficient to reduce the impact of the device abnormality on the final result.

[0095] The component scaling module includes: a degradation calculation unit and a simplification processing unit;

[0096] The degradation calculation unit is used to calculate the degradation matrix of the device according to the abnormal frequency and deviation coefficient of the abnormal device;

[0097] The simplified processing unit is used to multiply the matrix to be input into the abnormal device with the degradation matrix. After the node device obtains the final data matrix, it is multiplied by the inverse matrix of the degradation matrix to restore the data of the final data matrix, thereby reducing the impact of the abnormal device on the final result before the abnormal device is repaired.

[0098] As shown in Figure 2, a method for intelligent supervision of abnormal data based on data monitoring includes the following steps:

[0099] S100. All node devices are numbered in the order of data processing flow, and the industrial control terminal assigns a unique characteristic matrix to each node device and calculates the inverse matrix of these characteristic matrices;

[0100] Step S100 includes:

[0101] Step S101. Number all node devices in the order of data processing, and record the numbering result as data set X, where X = {1, 2, ..., n}, where n is the number of node devices, n is an integer and n ≥ 2;

[0102] Step S102. The industrial control terminal generates n different characteristic matrices, wherein the characteristic matrices are reversible matrices, and the eigenvalues ​​of all characteristic matrices are different;

[0103] The feature matrix is ​​assigned to each node device, and the assignment result is recorded as data set Y, Y={A1, A2, ..., A n}, the dataset Y and dataset X form a mapping relationship, A1, A2 and A n They represent the feature matrices assigned to node devices numbered 1, 2, and n, respectively, and the mapping relationships of the remaining elements are similar;

[0104] Step S103. The industrial control terminal calculates the inverse matrix of each characteristic matrix, and the calculation results form the data set Z, Z={A1 -1 , A2 -1 ,…,A n -1}, the dataset Z and dataset Y form a mapping relationship, A1 -1 , A2 -1 , and A n -1 Represent the feature matrices A1, A2 and A respectively n The inverse matrix of , and the mapping relationship of the remaining elements is similar;

[0105] Step S104. Calculate the eigenvalues ​​of each characteristic matrix, and record the calculation results as data set R, R={r1, r2, ..., r n}, the dataset R and the dataset Y form a mapping relationship, r1, r2, and r n Represent the feature matrices A1, A2 and A respectively n The mapping relationships of the remaining elements are similar.

[0106] S200. After the industrial control probe obtains the initial data, it sends the matrix recording the initial data to the first node device according to the node device number. After the first node device receives the initial data matrix, it multiplies its own characteristic matrix by the initial data matrix to obtain a composite matrix. The composite matrix is ​​input as an input matrix to the device for processing and outputs a processed matrix.

[0107] Step S200 includes:

[0108] Step S201: The industrial control probe obtains initial data and sends the initial data in the form of a matrix to the node device numbered 1. The initial data matrix is ​​denoted as Q.

[0109] Step S202: After receiving the initial data matrix, node device 1 performs operation T1=Q·A1, where T1 is the composite matrix of node device 1. T1 is used as the input matrix and the operation is performed according to the operation program of node device 1 itself to obtain a processing matrix Q1.

[0110] S300. The first node device sends the first processing matrix to the second node device in order of number. The second node device uses the first processing matrix as the initial data matrix and repeats step S200 to obtain the second processing matrix. This process is repeated in this way until all node devices have completed the processing and obtain the final processing matrix.

[0111] Step S300 includes:

[0112] Step S301. Node device 1 sends the primary processing matrix to node device 2. Node device 2 performs the operation T2 = Q1·A2, where T2 is the composite matrix of node device 2. Node device 2 uses T2 as the input matrix and performs the operation according to its own operation program to obtain the secondary processing matrix Q2.

[0113] Step S302: Repeat the above steps until node device n completes matrix processing and obtains the final processing matrix Q. n , the matrix Q n Send to the industrial control terminal.

[0114] S400. The control terminal receives the final processing matrix, calculates the eigenvalue of the final processing matrix, and then multiplies the final processing matrix by the inverse matrix of the characteristic matrix of each node device in reverse order of the number to obtain multiple reduction matrices. All reduction matrices are arranged in ascending order of the number of reductions, and the ratio of the eigenvalue of each reduction matrix to the eigenvalue of the previous reduction matrix is ​​calculated;

[0115] Step S400 includes:

[0116] Step S401. The industrial control terminal receives Q n Then, calculate Q n The eigenvalue t n+1 , and then perform the following operations to calculate the restored matrix: ;

[0117] Among them, P m Represents the reduction matrix corresponding to the node device numbered m, where m∈X, Represents the ni-th element in the data set Z;

[0118] Step S402. Calculate the eigenvalues ​​of each restored matrix, and record the calculation results as data set T, T={t1,t2,…t m ,…,t n}, where t m Represents the restored matrix P m The corresponding eigenvalues;

[0119] Step S403: Verify the accuracy of the eigenvalue processing according to the following formula: ;

[0120] Among them, N m Represents the status parameters of node device m, the N m ∈{0,1}, IFS is a logical judgment function, is the judgment formula in the logical judgment function. If the judgment formula is established, then N m =1, if the judgment is not true, then N m =0;

[0121] Generate a verification sequence K in reverse order of numbering, where K=[n,n-1,…,1], and substitute the elements in sequence K into m in the above formula in turn to obtain the result state sequence L, where L=[N n ,N n-1 ,…,N1], the elements in the sequence L correspond one to one with the elements in the sequence K;

[0122] Step S404. Determine in order whether the elements in the sequence L are 1. If all the elements in the sequence L are 1, it is determined that there is no abnormality in the data, and the matrix P is restored. n It is sent as the final execution matrix to the execution terminal for execution;

[0123] When it is found that the value of an element in the sequence L is 0, it is determined that there is an abnormality in the data, and the position of this element in the sequence L is recorded as g, where g∈X. The node device numbered g is marked as an abnormal device, the maintenance personnel are notified to repair the device g, and an abnormal record of the node device g is added, and go to step S500.

[0124] S500. Determine whether the data is abnormal based on the ratio and the eigenvalue of the characteristic matrix of each node device. When the data is determined to be abnormal, give the number of the abnormal device and calculate the degradation coefficient of the abnormal device. According to the degradation coefficient, obtain the degradation matrix, and then use the degradation matrix to scale the data processing process of the abnormal device.

[0125] Step S500 includes:

[0126] Step S501. Calculate the deviation value e of the node device g, , where t g+1Represents the restored matrix eigenvalue corresponding to the node device numbered g+1, r g Represents the eigenvalue of the characteristic matrix of node device g;

[0127] Obtain the historical abnormality records of node device g. Within the preset time range T, the total number of records is recorded as b, and the abnormality frequency of node device g is V=b / T;

[0128] Step S502: Calculate the degradation coefficient J of node device g according to the following formula: ;

[0129] Among them, V0 is the system's preset fault-tolerant frequency;

[0130] Step S503: Calculate the degradation matrix W, where W = J·E·A g , where E is the identity matrix, A g is the characteristic matrix of node device g, and the restored degradation matrix W -1 =1 / J·A g -1 , where Ag -1 is the inverse matrix of the device g feature matrix;

[0131] Before the abnormal device g is repaired, the characteristic matrix of the node device g is replaced by W and the characteristic matrix of the node device g is replaced by W. -1 Replace the inverse matrix of the characteristic matrix of node device g, and in the subsequent execution of step S404, let N g =1;

[0132] The processed restored matrix P n It is sent as the final execution matrix to the execution terminal for execution.

[0133] Example:

[0134]

[0135] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0136] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for intelligent supervision of abnormal data based on data monitoring, characterized in that: The method comprises the following steps: S100. All node devices are numbered in the order of data processing flow, the industrial control terminal assigns a unique characteristic matrix to each node device, and calculates the inverse matrix of these characteristic matrices; S200. After the industrial control probe obtains the initial data, it sends the matrix recording the initial data to the first node device according to the number of the node device. After the first node device receives the initial data matrix, it multiplies its own characteristic matrix on the right after the initial data matrix to obtain a composite matrix, and inputs the composite matrix into the device as an input matrix for processing, and outputs a processed matrix; S300. The first node device sends the first processing matrix to the second node device in the order of numbers. The second node device uses the first processing matrix as the initial data matrix and repeats step S200 to obtain the second processing matrix. This is repeated by analogy. After all node devices have completed the processing, the final processing matrix is ​​obtained. S400. The control terminal receives the final processing matrix, calculates the eigenvalue of the final processing matrix, and then multiplies the final processing matrix by the inverse matrix of the characteristic matrix of each node device in reverse order of number to obtain multiple restored matrices, arranges all restored matrices from small to large according to the number of restorations, and calculates the ratio of the eigenvalue of each restored matrix to the eigenvalue of the previous restored matrix; S500. Determine whether the data is abnormal based on the ratio and the eigenvalue of the characteristic matrix of each node device. When the data is determined to be abnormal, give the number of the abnormal device, and calculate the degradation coefficient of the abnormal device based on the deviation value and historical records of the abnormal device. Get the degradation matrix based on the degradation coefficient, and then use the degradation matrix to scale the data processing process of the abnormal device.

2. The method for intelligent supervision of abnormal data based on data monitoring according to claim 1 is characterized in that: Step S100 includes: Step S101. All node devices are numbered in the order of data processing flow, and the numbering result is recorded as data set X, X={1,2,…,n}, where n is the number of node devices, n is an integer and n≥2; Step S102. The industrial control terminal generates n different characteristic matrices, wherein the characteristic matrices are reversible matrices, and the eigenvalues ​​of all characteristic matrices are different; The feature matrix is ​​assigned to each node device, and the assignment result is recorded as data set Y, Y = {A1, A2, ..., A n }, the data set Y and data set X form a mapping relationship, A1, A2 and A n They represent the characteristic matrices assigned to the node devices numbered 1, 2, and n, respectively, and the mapping relationships of the remaining elements are similar; Step S103. The industrial control terminal calculates the inverse matrix of each characteristic matrix, and the calculation results constitute the data set Z, Z={A1 -1 , A2 -1 , …, A n -1 }, the data set Z and the data set Y form a mapping relationship, A1 -1 , A2 -1 , and A n -1 Represent the feature matrices A1, A2 and A respectively n The inverse matrix of , and the mapping relationship of the remaining elements is similar; Step S104. Calculate the eigenvalues ​​of each feature matrix, and record the calculation results as data set R, R = {r1, r2, ..., r n }, the data set R and the data set Y form a mapping relationship, r1, r2, and r n Represent the feature matrices A1, A2 and A respectively n The mapping relationships of the remaining elements are similar.

3. The method for intelligent supervision of abnormal data based on data monitoring according to claim 2 is characterized in that: Step S200 includes: Step S201. The industrial control probe obtains initial data and sends the initial data in the form of a matrix to the node device numbered 1. The initial data matrix is ​​recorded as Q. Step S202. After receiving the initial data matrix, the node device 1 performs the operation T1=Q·A1, where T1 is the composite matrix of the node device 1. T1 is used as the input matrix and the operation is performed according to the operation program of the node device 1 itself to obtain a primary processing matrix Q1. Step S300 includes: Step S301. Node device 1 sends a primary processing matrix to node device 2, and node device 2 performs operation T2=Q1·A2, where T2 is a composite matrix of node device 2. T2 is used as an input matrix and is operated according to the operation program of node device 2 to obtain a secondary processing matrix Q2. Step S302: Repeat the above steps until node device n completes matrix processing and obtains the final processing matrix Q. n , the matrix Q n Sent to the industrial control terminal.

4. The method for intelligent supervision of abnormal data based on data monitoring according to claim 3 is characterized in that: Step S400 includes: Step S401. The industrial control terminal receives Q n Then, calculate Q n The eigenvalue t n+1 , and then perform the following operations to calculate the restored matrix: ; Among them, P m represents the reduction matrix corresponding to the node device numbered m, where m∈X, Represents the ni-th element in the data set Z; Step S402. Calculate the eigenvalues ​​of each restored matrix, and record the calculation results as data set T, T={t1,t2,…t m ,…,t n }, where t m Represents the restored matrix P m The corresponding eigenvalues; Step S403. Verify the accuracy of the eigenvalue processing according to the following formula: ; Among them, N m represents the state parameters of node device m, the N m ∈{0,1}, IFS is a logical judgment function, is the judgment formula in the logical judgment function. If the judgment formula is established, then N m =1, if the judgment is not true, then N m =0; Generate a verification sequence K in reverse order of numbering, where K=[n,n-1,…,1], and substitute the elements in sequence K into m in the above formula in turn to obtain the result state sequence L, where L=[N n ,N n-1 ,…,N1], the elements in the sequence L correspond one to one with the elements in the sequence K; Step S404. Determine in order whether the elements in the sequence L are 1. If all the elements in the sequence L are 1, it is determined that there is no abnormality in the data, and the matrix P is restored. n It is sent as the final execution matrix to the execution terminal for execution; When it is found that the value of an element in the sequence L is 0, it is determined that the data is abnormal, and the position of this element in the sequence L is recorded as g, where g∈X. The node device numbered g is marked as an abnormal device, the maintenance personnel are notified to repair the device g, and an abnormal record of the node device g is added, and go to step S500.

5. The method for intelligent supervision of abnormal data based on data monitoring according to claim 4 is characterized in that: Step S500 includes: Step S501. Calculate the deviation value e of the node device g, , where t g+1 represents the restored matrix eigenvalue corresponding to the node device numbered g+1, r g Represents the eigenvalue of the characteristic matrix of node device g; Obtain the historical abnormal records of node device g. Within the preset time range T, the total number of records is recorded as b, then the abnormal frequency of node device g is V=b / T; Step S502. Calculate the degradation coefficient J of the node device g according to the following formula: ; Among them, V0 is the fault-tolerant frequency preset by the system; Step S503: Calculate the degradation matrix W, where W = J·E·A g , where E is the identity matrix, A g is the characteristic matrix of node device g, and the restored degradation matrix W -1 =1 / J·A g -1 , where Ag -1 is the inverse matrix of the feature matrix of device g; Before the abnormal device g is repaired, the feature matrix of node device g is replaced by W and W is used to replace the feature matrix of node device g. -1 Replace the inverse matrix of the characteristic matrix of node device g, and in the subsequent execution of step S404, let N g =1; The processed restored matrix P n The final execution matrix is ​​sent to the execution terminal for execution.

6. An abnormal data intelligent supervision system based on data monitoring, characterized in that: The system includes the following modules: a data transceiver module, a matrix processing module, a feature analysis module, an abnormality location module and a component scaling module; The data transceiver module is used to extract matrix features, send and receive matrix data between node devices, and extract matrix row and column widths; The matrix processing module is used to perform a multiplication operation of the data matrix and the node device characteristic matrix after receiving the data matrix sent by the previous process to obtain a composite matrix, and complete the node device's own processing of the composite matrix; The feature analysis module is used to analyze the eigenvalues ​​and eigenvectors of the matrix after the industrial control terminal receives the final data matrix, and to perform step-by-step inverse operations on the data matrix according to the feature matrix of each node device to obtain a multi-dimensional transfer matrix, and then to determine whether there is a data processing error by calculating the eigenvalue of the transfer matrix. If there is no error, the final transfer matrix is ​​input to the execution terminal for execution; The abnormality positioning module is used to locate the node device where the error step occurs according to the number of inverse operations on the data matrix when a data processing error is found, calculate the deviation value of the device, add an abnormal record of the device, and notify the maintenance personnel of the number of the abnormal device; The component scaling module is used to calculate the degradation coefficient of the abnormal device according to the deviation value and the abnormal frequency of the abnormal device, and scale the processing matrix of the device according to the degradation coefficient to reduce the impact of the device abnormality on the final result.

7. The abnormal data intelligent supervision system based on data monitoring according to claim 6 is characterized by: The data transceiver module includes: a matrix extraction unit and a signal transceiver unit; The matrix extraction unit is used to obtain the initial matrix through the industrial control probe and extract the row and column widths of the matrix; The signal transceiver unit is used to receive the data matrix sent by the previous device and send the processed matrix information to the next device; The matrix processing module includes: a feature fitting unit and an engineering operation unit; The feature fitting unit is used to, after receiving the data matrix, multiply the device's own feature matrix with the data matrix to fit the device features into the data matrix to obtain a composite matrix; The engineering operation unit is used to take the composite matrix as an input matrix, input the node device to perform data processing, and output the processed data matrix.

8. The abnormal data intelligent supervision system based on data monitoring according to claim 7 is characterized by: The feature analysis module includes: an inversion feature unit, a solution fitting unit and an abnormality judgment unit; The inverse characteristic unit is used to calculate the inverse matrix of the characteristic matrix of all node devices, and calculate the eigenvalues ​​and eigenvectors of these inverse matrices; The solution fitting unit is used to calculate the eigenvalues ​​of the data matrix after the industrial control terminal receives the final data matrix, and multiply the eigenvalues ​​with the inverse matrices of the characteristic matrices of each node device one by one to obtain multiple restored matrices; The abnormality judgment unit is used to calculate the eigenvalue of the restored matrix each time a restored matrix is ​​fitted, and compare the eigenvalue of the restored matrix, the eigenvalue of the characteristic matrix and the eigenvalue of the previous restored matrix to judge whether there is an abnormality in the processing process.

9. The abnormal data intelligent supervision system based on data monitoring according to claim 8 is characterized by: The abnormality positioning module includes: an inversion analysis unit, a feature recovery unit and a deviation recording unit; The inverter analysis unit is used to determine the number of the abnormal device according to the number of restoration steps of the restoration matrix after the abnormality is found, and notify the maintenance personnel of the number of the abnormal device; The feature recovery unit is used to request the previous node device of the abnormal device to send a data matrix, continue to analyze the data matrix, and determine the abnormal conditions of other devices; The deviation recording unit is used to add an abnormal record of an abnormal device and calculate the deviation coefficient of the abnormal device according to the eigenvalue of the restoration matrix.

10. The abnormal data intelligent supervision system based on data monitoring according to claim 9 is characterized in that: The component scaling module includes: a degradation calculation unit and a simplification processing unit; The degradation calculation unit is used to calculate the degradation matrix of the device according to the abnormal frequency and deviation coefficient of the abnormal device; The simplified processing unit is used to multiply the matrix to be input into the abnormal device with the degradation matrix, and after the node device obtains the final data matrix, it is multiplied with the inverse matrix of the degradation matrix to restore the data of the final data matrix.

Images