Fault tracing method and device based on causal analysis, and electronic equipment

Abstract

The invention relates to a fault tracing method and device based on causal analysis, and electronic equipment, and the method comprises the steps: obtaining a to-be-processed data set which comprises a plurality of related variables corresponding to a target fault in a target system when the target fault occurs in the target system, and a data sequence of each related variable in a preset time; performing causal analysis on the to-be-processed data set, and determining a dependent variable having a causal relationship with the target fault in the plurality of related variables; based on a Bayesian network model, an intervention probability corresponding to each dependent variable is determined, and the intervention probability corresponding to each dependent variable is used for representing that the dependent variable is a probability value of a fault cause; and determining a fault reason corresponding to the target fault according to the intervention probability corresponding to each dependent variable. According to the embodiment of the invention, the influence of a hybrid factor and an intermediary variable can be eliminated, the accuracy of a fault tracing result is improved, and the instability of the fault tracing result caused by noise is avoided.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a fault tracing method and device based on causal analysis, and an electronic device. Background Art

[0002] Faults that occur during industrial production may spread rapidly within the industrial system, affecting production safety and product quality. Therefore, when a fault occurs, it is necessary to trace the fault source in a timely manner to determine the root cause node. However, due to the mutual influence between production variables, there may be multiple alarm paths from the root cause node to the fault detection node, and may contain false alarms, making it difficult to accurately trace the fault source. Summary of the invention

[0003] In view of this, the present disclosure proposes a fault tracing method and device based on causal analysis, and a technical solution for an electronic device.

[0004] According to one aspect of the present disclosure, a fault tracing method based on causal analysis is provided, comprising: obtaining a data set to be processed, wherein the data set to be processed includes a plurality of related variables corresponding to the target fault in the target system when a target fault occurs in the target system, and a data sequence of each of the related variables within a preset time; performing causal analysis on the data set to be processed, and determining a dependent variable that has a causal relationship with the target fault among the plurality of related variables; determining an intervention probability corresponding to each of the dependent variables based on a Bayesian network model, wherein the intervention probability corresponding to each of the dependent variables is used to represent the probability value that the dependent variable is the cause of the fault; and determining the fault cause corresponding to the target fault according to the intervention probability corresponding to each of the dependent variables.

[0005] In a possible implementation, the causal analysis is performed on the data set to be processed and the dependent variable is determined among the multiple related variables, including: grouping the data set to be processed to determine a processing data group and a control data group; determining the causal regression coefficient corresponding to each of the related variables based on the processing data group and the control data group and based on a stable learning algorithm, wherein the causal regression coefficient corresponding to each of the related variables is used to represent the correlation between the related variable and the target fault; and determining the dependent variable among the multiple related variables based on the causal regression coefficient corresponding to each of the related variables.

[0006] In a possible implementation, the grouping of the data set to be processed to determine the processing data group and the control data group includes: determining the indicator vector corresponding to each of the relevant variables according to the normal sample mean corresponding to each of the relevant variables, wherein the normal sample mean corresponding to each of the relevant variables is the average value of the relevant variable when the target system is in a normal state; and dividing the data set to be processed into the processing data group and the control data group according to the indicator vector corresponding to each of the relevant variables.

[0007] In a possible implementation, the causal regression coefficient corresponding to each of the related variables is determined based on the processed data group and the control data group based on a stable learning algorithm, including: determining a target optimization problem based on the processed data group and the control data group, wherein the target optimization problem is used to extract the dependent variable from the multiple related variables; solving the target optimization problem using the stable learning algorithm to determine the causal regression coefficient corresponding to each of the related variables.

[0008] In a possible implementation, determining the dependent variable from the multiple related variables based on the causal regression coefficient corresponding to each of the related variables includes: sorting the causal regression coefficient corresponding to each of the related variables to determine a sorted coefficient sequence, wherein the coefficient sequence includes the absolute value of the causal regression coefficient corresponding to each of the related variables, and the coefficients are sorted from large to small according to the absolute value; accumulating the absolute values ​​of the causal regression coefficient corresponding to each of the related variables included in the coefficient sequence in sequence, and comparing the cumulative percentage determined after each accumulation with a preset threshold; when the cumulative percentage corresponding to the first k related variables in the coefficient sequence is less than or equal to the preset threshold, and the cumulative percentage corresponding to the first k+1 related variables in the coefficient sequence is greater than the preset threshold, the first k related variables in the coefficient sequence are determined as the dependent variables.

[0009] In a possible implementation, the intervention probability corresponding to each of the dependent variables is determined based on a Bayesian network model, including: determining a causal graph corresponding to the target fault according to the dependent variable and a label data sequence corresponding to the target system, wherein the label data sequence is used to represent the working state of the target system within the preset time; discretizing each of the dependent variables according to the data sequence of each of the dependent variables within the preset time to obtain a discretized variable corresponding to each of the dependent variables; constructing a Bayesian network model according to the causal graph and the discretized variables corresponding to each of the dependent variables; determining a conditional probability table corresponding to the target fault and an edge probability corresponding to each node in the causal graph according to the Bayesian network model, wherein the conditional probability table is used to represent the probability of the target fault occurring when each of the dependent variables is abnormal, and the edge probability is used to represent the probability of each of the dependent variables being abnormal; determining the intervention probability corresponding to each of the dependent variables according to the conditional probability table and the edge probability.

[0010] In a possible implementation manner, the method further includes: when the target failure occurs in the target system, acquiring an original data set corresponding to the target system; performing data standardization processing on the original data set to determine the data set to be processed.

[0011] According to another aspect of the present disclosure, a fault tracing device based on causal analysis is provided, including: a data acquisition module, which acquires a data set to be processed, wherein the data set to be processed includes a plurality of related variables corresponding to the target fault in the target system when a target fault occurs in the target system, and a data sequence of each of the related variables within a preset time; a dependent variable extraction module, which is used to perform causal analysis on the data set to be processed, and determine the dependent variable that has a causal relationship with the target fault among the plurality of related variables; an intervention probability determination module, which is used to determine the intervention probability corresponding to each of the dependent variables based on a Bayesian network model, wherein the intervention probability corresponding to each of the dependent variables is used to represent the probability value that the dependent variable is the cause of the fault; and a fault source determination module, which is used to determine the fault cause corresponding to the target fault according to the intervention probability corresponding to each of the dependent variables.

[0012] According to another aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0013] According to another aspect of the present disclosure, a non-volatile computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions implement the above method when executed by a processor.

[0014] According to another aspect of the present disclosure, a computer program product is provided, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0015] In an embodiment of the present disclosure, a data set to be processed is obtained, and the data set to be processed includes multiple related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and the data sequence of each related variable within a preset time. A causal analysis is performed on the data set to be processed, and the dependent variable that has a causal relationship with the target fault is determined among the multiple related variables, so that the key variables related to the target fault can be determined, and the accuracy and efficiency of locating the cause of the fault can be improved. Based on the Bayesian network model, the intervention probability corresponding to each dependent variable is determined. According to the intervention probability corresponding to each dependent variable, the fault cause corresponding to the target fault can be directly determined. Compared with the reasoning method based on conditional probability in traditional fault tracing, it can eliminate the influence of confounding factors and mediating variables, improve the accuracy of the fault tracing results, and avoid the instability of the fault tracing results caused by noise.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only and do not limit the present disclosure. Other features and aspects of the present disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0018] Figure 1 A causal diagram between a processing variable X and an experimental result Y in the related art is shown;

[0019] Figure 2 A causal diagram between a processing variable X and an experimental result Y in the related art is shown;

[0020] Figure 3 A causal diagram between a processing variable X and an experimental result Y in the related art is shown;

[0021] Figure 4 A flow chart of a fault tracing method based on causal analysis according to an embodiment of the present disclosure is shown;

[0022] Figure 5 A histogram showing a causal regression coefficient of a condenser cooling water inlet temperature fault according to an embodiment of the present disclosure;

[0023] Figure 6 A histogram showing a square prediction error of a condenser cooling water inlet temperature fault in the related art;

[0024] Figure 7 A cause-effect diagram corresponding to a feed ratio change fault according to an embodiment of the present disclosure is shown;

[0025] Figure 8 A cause-effect diagram corresponding to a feed temperature fault according to an embodiment of the present disclosure is shown;

[0026] Fig. 9 A histogram showing a causal regression coefficient corresponding to a failure of a stand roll gap control actuator according to an embodiment of the present disclosure;

[0027] Fig.10 A histogram showing causal regression coefficients corresponding to a cooling water valve actuator fault between racks according to an embodiment of the present disclosure;

[0028] Fig.11 A cause-effect diagram corresponding to a cooling water valve actuator fault between racks according to an embodiment of the present disclosure is shown;

[0029] Fig.12 A block diagram of a fault tracing device based on causal analysis according to an embodiment of the present disclosure is shown;

[0030] Fig.13 A block diagram of an electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0031] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0032] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0033] The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there may 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. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can represent including any one or more elements selected from the set consisting of A, B, and C.

[0034] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present disclosure.

[0035] When an industrial system fails during the production process, it may spread rapidly throughout the entire industrial system, causing adverse effects on production safety and product quality. Therefore, when a failure occurs, it is necessary to trace the fault in a timely manner to determine the root cause node. However, due to the mutual influence between production variables, there may be multiple different alarm paths from the root cause node to the fault detection node, which may even include false alarms, making fault tracing difficult.

[0036] In the prior art, the Bayesian network-based reasoning method is usually used to analyze the fault tracing problem, and the fault propagation path and root cause are determined by reasoning conditional probability. However, due to the large number of components in complex industrial systems and the difficulty in obtaining relevant knowledge, it is more difficult to establish a graphical model based on prior knowledge, and most methods that rely on data to establish a graphical model have a large amount of calculation, which makes it difficult to build a Bayesian network model.

[0037] In the prior art, a confounding factor refers to a variable that affects both the treatment variable and the experimental results of the causal analysis when performing causal analysis.

[0038] Figure 1 FIG. 1 shows a causal diagram between a processing variable X and an experimental result Y in the related art. Figure 1 As shown in the figure, there is no pointing path between the treatment variable X and the experimental result Y, but the treatment result X and the experimental result Y are both affected by the variable Z. Therefore, when performing causal analysis based on relevant data to analyze the causal relationship between the treatment variable X and the experimental result Y, due to the existence of variable Z, the conditional probability P(Y|X)≠0 can be obtained. At this time, if the conditional probability P(Y|X)≠0 is used as the result of causal analysis, it will lead to a false causal relationship and pseudo-correlation between the treatment variable X and the experimental result Y. In fact, the conditional probability P(Y|X)≠0 here only means that there is a correlation between the treatment variable X and the experimental result Y, not a causal relationship. In this analysis process, variable Z is a confounding factor.

[0039] Figure 2 FIG. 1 shows a causal diagram between a processing variable X and an experimental result Y in the related art. Figure 2As shown in the figure, the treatment variable X points to the experimental result Y, and the variable Z points to both the treatment variable X and the experimental result Z. At this time, in the process of causal analysis of the treatment variable X and the experimental result Y, due to the existence of the confounding factor Z, the conditional probability P(Y|X) obtained based on the relevant data cannot reflect the true causal relationship between the treatment variable X and the experimental result Y.

[0040] The traditional Bayesian network-based reasoning method is based on the conditional probability between variables. As can be seen from the above, the conditional probability in the Bayesian network actually represents the correlation between two variables. According to the relevant theory of causal inference, the variable relationship obtained based on correlation analysis may be affected by confounding factors, thereby affecting the performance of the fault diagnosis system and making the fault tracing results lack of interpretability. Furthermore, due to the influence of confounding factors, the causal graph constructed based on variable data may deviate from the real causal graph, and there may be falsely associated connection edges, resulting in inaccurate fault tracing results.

[0041] Figure 3 FIG. 1 shows a causal diagram between a processing variable X and an experimental result Y in the related art. Figure 2 As shown in the figure, there are two paths between the treatment variable X and the experimental result Y, namely X→Y and X→Z→Y, and there is also the influence of the confounding factor C. In this case, variable Z is called a mediating variable. The first path X→Y between the treatment variable X and the experimental result Y is a direct causal path.

[0042] After a fault occurs, the propagation process from the fault root cause node to the fault detection node may be affected by the propagation from the mediating variable to the fault detection node. Therefore, in order to accurately determine the direct causal path from the fault root cause node to the fault detection node, the influence of the mediating variable needs to be considered, which is not considered in the existing Bayesian network inference method.

[0043] The embodiment of the present disclosure provides a fault tracing method based on causal analysis, which can eliminate the influence of confounding factors and mediating variables on fault tracing based on the principle of causal analysis and directly determine the real cause of the fault. The fault tracing method based on causal analysis provided by the embodiment of the present disclosure is described in detail below.

[0044] Figure 4A flow chart of a fault tracing method based on causal analysis according to an embodiment of the present disclosure is shown. The fault tracing method based on causal analysis can be executed by an electronic device such as a terminal device or a server. The terminal device can be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. The fault tracing method based on causal analysis can be implemented by a processor calling computer-readable instructions stored in a memory. Alternatively, the fault tracing method based on causal analysis can be executed by a server. Figure 4 As shown, the fault tracing method based on causal analysis includes:

[0045] In step S41, a data set to be processed is obtained, wherein the data set to be processed includes a plurality of relevant variables corresponding to the target fault in the target system when the target fault occurs in the target system, and a data sequence of each relevant variable within a preset time.

[0046] The target system here can be an industrial system or other system with multiple components. The target fault can be an abnormal phenomenon in which the characteristics or parameters of the target system deviate greatly beyond the acceptable range during the operation and / or production process of the target system, thereby causing the performance of the target system to deteriorate. The relevant variables corresponding to the target fault are variables in the target system that may be associated with the target fault. For example, when the target system A has a target fault, the temperature, humidity, pressure, etc. corresponding to the component a of the target system A are all relevant variables corresponding to the target fault.

[0047] The data set to be processed includes multiple related variables corresponding to the target fault in the target system, and the data sequence of each related variable within the preset time. Among them, the data sequence of any related variable within the preset time includes the data corresponding to the related variable at each moment within the preset time. For example, the temperature corresponding to the component a of the target system A is the related variable corresponding to the target fault occurring in the target system A. The preset time includes time T1 to time T3. The data corresponding to the related variable at time T1 is 35°C, the data corresponding to the related variable at time T2 is 39°C, and the data corresponding to the related variable at time T3 is 32°C. It can be determined that the data sequence corresponding to the related variable within the preset time is Y=[35, 39, 32]. The preset time is a time period including the time when the target fault occurs, which can be set according to actual usage requirements, and the present disclosure does not make specific restrictions on this. The specific form of the data sequence of each related variable within the preset time can be set according to actual usage requirements, and the present disclosure does not make specific restrictions on this.

[0048] For the specific method of obtaining the data set to be processed, reference may be made to the implementation methods in the related art, and the embodiments of the present disclosure do not specifically limit this.

[0049] In step S42, a causal analysis is performed on the data set to be processed, and a dependent variable having a causal relationship with the target fault is determined from a plurality of related variables.

[0050] By performing causal analysis on the data set to be processed, the dependent variable that has a causal relationship with the target fault can be determined from the multiple related variables corresponding to the target fault. Among them, the dependent variable can be used to represent the variable corresponding to the fault cause that may cause the target fault in the target system. Fault tracing based on the dependent variable can narrow the candidate set of fault causes and improve the accuracy and efficiency of fault cause location.

[0051] The process of causal analysis and determining a dependent variable having a causal relationship with a target fault among a plurality of related variables will be described in detail later in conjunction with possible implementations of the present disclosure, and will not be elaborated here.

[0052] In step S43, based on the Bayesian network model, the intervention probability corresponding to each dependent variable is determined, wherein the intervention probability corresponding to each dependent variable is used to represent the probability value that the dependent variable is the cause of the fault.

[0053] After determining the dependent variables corresponding to the target fault based on causal analysis, a Bayesian network model including the dependent variables can be constructed, and the intervention probability corresponding to each dependent variable can be determined based on the Bayesian network model.

[0054] Compared with the traditional Bayesian network model in the prior art, the Bayesian network model including dependent variables determined in the embodiment of the present disclosure takes into account more the relationship between the dependent variable and the target fault, as well as the relationship between different dependent variables, and can eliminate the influence of confounding factors and mediating variables.

[0055] The process of determining the intervention probability corresponding to each dependent variable based on the Bayesian network model will be described in detail later in conjunction with possible implementation methods of the present disclosure, and will not be elaborated here.

[0056] In step S44, the fault cause corresponding to the target fault is determined according to the intervention probability corresponding to each dependent variable.

[0057] After determining the intervention probability corresponding to each dependent variable respectively, the dependent variable that causes the target fault can be directly determined according to the intervention probability, thereby determining the fault cause corresponding to the target fault.

[0058] In an embodiment of the present disclosure, a data set to be processed is obtained, and the data set to be processed includes multiple related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and the data sequence of each related variable within a preset time. A causal analysis is performed on the data set to be processed, and the dependent variable that has a causal relationship with the target fault is determined among the multiple related variables, so that the key variables related to the target fault can be determined, and the accuracy and efficiency of locating the cause of the fault can be improved. Based on the Bayesian network model, the intervention probability corresponding to each dependent variable is determined. According to the intervention probability corresponding to each dependent variable, the fault cause corresponding to the target fault can be directly determined. Compared with the reasoning method based on conditional probability in traditional fault tracing, it can eliminate the influence of confounding factors and mediating variables, improve the accuracy of the fault tracing results, and avoid the instability of the fault tracing results caused by noise.

[0059] In a possible implementation, the method further includes: when a target failure occurs in the target system, acquiring an original data set corresponding to the target system; performing data standardization processing on the original data set to determine a data set to be processed.

[0060] When a target fault occurs in a target system, an original data set of the target system within a preset time can be obtained, and the original data set can include multiple relevant variables corresponding to the target fault in the target system, and an initial data sequence of each relevant variable within the preset time. The initial data sequence corresponding to each relevant variable can include unprocessed initial data collected at each moment of the relevant variable within the preset time.

[0061] Since the value ranges and units of the initial data corresponding to different related variables in the original data set may be different, after obtaining the original data set, the original data set can be subjected to data standardization processing to determine the corresponding data set to be processed, so as to facilitate subsequent fault tracing. The specific method of data standardization can refer to the implementation methods in the relevant technology, and this disclosure does not make specific limitations on this.

[0062] In one example, the data set to be processed can be expressed as formula (1):

[0063]

[0064] Among them, X′ represents the data set to be processed; X represents the original data set, and X∈R n×p , n represents the number of data sequences of relevant variables in the original data set within the preset time, p represents the number of relevant variables in the original data set; M represents the normal sample mean corresponding to each relevant variable, and M∈R 1×p ;e i Represents the standard deviation corresponding to the i-th related variable in the original data set X, i = 1, 2, ..., p.

[0065] In one possible implementation, a causal analysis is performed on a data set to be processed, and a dependent variable is determined among multiple related variables, including: grouping the data set to be processed to determine a processed data group and a control data group; determining a causal regression coefficient corresponding to each related variable based on the processed data group and the control data group based on a stable learning algorithm, wherein the causal regression coefficient corresponding to each related variable is used to represent the correlation between the related variable and the target fault; and determining a dependent variable among multiple related variables based on the causal regression coefficient corresponding to each related variable.

[0066] In the prior art, traditional machine learning usually assumes that the training data set and the test data set have the same distribution. When the data set does not meet this assumption in actual applications, the prediction performance of traditional machine learning methods will be unstable. Stable learning is a causal analysis method based on the potential outcome model framework. When the actual causal structure is unknown, it can improve the stability of prediction and the performance of traditional machine learning by balancing the influence of confounding factors.

[0067] On this basis, after obtaining the data set to be processed, the data set to be processed can be grouped to determine the processing data group and the control data group. Based on the stable learning algorithm, causal analysis is performed on the processing data group and the control data group to determine the causal regression coefficient corresponding to each relevant variable.

[0068] Since the causal regression coefficient corresponding to each relevant variable can represent the correlation between the relevant variable and the target fault, the dependent variable corresponding to the target fault can be determined from multiple relevant variables according to the causal regression coefficient corresponding to each relevant variable.

[0069] In one possible implementation, the data set to be processed is grouped to determine a processing data group and a control data group, including: determining an indicator vector corresponding to each relevant variable according to a normal sample mean corresponding to each relevant variable, wherein the normal sample mean corresponding to each relevant variable is the average value of the relevant variable when the target system is in a normal state; and dividing the data set to be processed into a processing data group and a control data group according to the indicator vector corresponding to each relevant variable.

[0070] According to the data of each relevant variable in the data set to be processed when the target system is in a normal state, the normal sample mean corresponding to each relevant variable can be determined, wherein the normal sample mean corresponding to each relevant variable can represent the average value of the relevant variable when the target system is in a normal state.

[0071] By using the normal sample mean corresponding to each relevant variable, the indicator vector corresponding to each relevant variable can be determined. Among them, the indicator vector corresponding to each relevant variable can be used to indicate whether the data corresponding to the relevant variable exceeds the normal sample mean. The specific method of determining the indicator vector corresponding to each relevant variable can refer to the implementation method in the relevant technology, and the present disclosure does not make specific limitations on this.

[0072] In one example, for a data set to be processed that is determined by data standardization using formula (1), the indicator vector corresponding to each relevant variable in the data set to be processed can be expressed as formula (2):

[0073]

[0074] Where X′(i, j) represents the i-th data of the j-th related variable of the data set X′ to be processed in the data sequence; j (i) represents the index vector I j The i-th data, i = 1, 2, 3, ... n; j = 1, 2, 3 ... p.

[0075] For the data set to be processed determined by data standardization through formula (1), when X′(i, j)≥0, it means that the corresponding data in the original data set exceeds the corresponding normal sample mean; when X′(i, j)<0, it means that the corresponding data in the original data set does not exceed the corresponding normal sample mean. Therefore, the indicator vector determined by formula (2) can distinguish the data that exceeds the normal sample mean and the data that does not exceed the normal sample mean in all the data of the jth variable of the data set to be processed, thereby dividing the data set to be processed into a processing data group and a control data group.

[0076] In one possible implementation, the causal regression coefficient corresponding to each relevant variable is determined based on a stable learning algorithm according to the processed data group and the control data group, including: determining a target optimization problem according to the processed data group and the control data group, wherein the target optimization problem is used to extract dependent variables from multiple relevant variables; solving the target optimization problem using a stable learning algorithm to determine the causal regression coefficient corresponding to each relevant variable.

[0077] Based on the stable learning algorithm of Causally Regularized Logistic Regression (CRLR), each relevant variable in the data set to be processed can be set as a processing variable in turn, and other relevant variables can be set as confounding factors, and then the causal influence of the relevant variables set as processing variables on the target fault can be estimated. In this process, the CRLR algorithm can balance the moments of the confounding factors of the processing data group and the control data group by learning the sample weighting matrix, that is, the sample weighting matrix is ​​used to ensure that all input relevant variables set as processing variables are independent of each other, thereby minimizing the influence of confounding factors.

[0078] Logistic regression analysis is used to study the influence relationship, which can be described as the influence of variable A on variable B, where variable A is categorical data and variable B can be either quantitative data or categorical data. The CRLR algorithm can add a sample weight matrix to the loss function of the logistic regression model, and then determine the sample weight matrix and causal regression coefficients through iterative optimization.

[0079] According to the above principle, the target optimization problem can be expressed as formula (3) to formula (6):

[0080]

[0081]

[0082]

[0083]

[0084] Among them, formula (3) represents the weighted logistic regression loss function determined based on the CRLR algorithm; formula (4) can be used to balance the impact of confounding factors; formula (5) can be used to avoid model overfitting; formula (6) can be used to ensure that the sample weight matrix W is non-negative and not all zero. W represents the sample weight matrix; Y represents the label data sequence, and Y∈R n×1 , the label data sequence can be used to represent the working state of the target system within a preset time; β represents the causal regression coefficient sequence; It represents the result after setting the j-th column of the data set X′ to be processed to 0 (i.e. removing the j-th variable of X′); γ1, γ2, γ3 and γ4 respectively represent different preset thresholds, which can be set according to actual usage requirements, and the present disclosure does not make specific limitations on this.

[0085] The CRLR algorithm can be used to iteratively optimize the above target optimization problem, determine the optimization hyperparameters, and determine the final causal regression coefficient based on the accuracy. Sample weighting matrix And hyperparameters and

[0086] In addition to the above-mentioned method, other stable learning algorithms may also be used to construct and solve the corresponding target optimization problem to determine the causal regression coefficient corresponding to each relevant variable. The present disclosure does not make specific limitations on this.

[0087] The causal regression coefficient corresponding to any relevant variable can reflect the correlation between the relevant variable and the target fault.

[0088] In one example, the target system is a condenser, and the target fault is a condenser cooling water inlet temperature fault. After the fault occurs, in order to maintain the system temperature unchanged, the condenser cooling water flow rate is changed through closed-loop control. Obtain a data set to be processed corresponding to the target fault, wherein the data set to be processed includes 52 related variables, namely related variables 1 to related variables 52, and the data sequence of each related variable within a preset time. Based on the above process, the causal regression coefficient corresponding to each related variable in the data set to be processed can be determined. As a comparison, the squared prediction error (SPE) corresponding to each related variable in the data set to be processed can also be determined based on the principal component analysis method in the prior art.

[0089] Figure 5 A histogram showing causal regression coefficients of a condenser cooling water inlet temperature fault according to an embodiment of the present disclosure. Figure 6 A histogram showing a square prediction error of a condenser cooling water inlet temperature fault in the related art.

[0090] like Figure 5 As shown in the figure, through the histogram of causal regression coefficients, it can be identified that the proportion of relevant variable 52 exceeds 50%, indicating that there is a causal relationship between relevant variable 52 and the target fault. Based on the prior knowledge of the target system, it can be determined that relevant variable 52 represents the condenser cooling water flow rate, which is consistent with the actual situation that in order to maintain the system temperature unchanged after the aforementioned fault occurs, the condenser cooling water flow rate changes through closed-loop control. Figure 6As shown, based on the histogram of the squared prediction error determined by the principal component analysis of the condenser, it is identified that the relevant variable that contributes the most to the target fault is the relevant variable 38, while the contribution of the relevant variable 52 is relatively small. By comparing the two, it can be seen that, compared with the method of judging the correlation between each relevant variable and the target fault by the principal component analysis method in the prior art, the embodiment of the present disclosure performs causal analysis on multiple relevant variables corresponding to the target fault by a stable learning algorithm, and determines the causal regression coefficient corresponding to each relevant variable, which can reveal the causal relationship between the relevant variables and the target fault, eliminate the influence of confounding factors, and more accurately represent the correlation between the relevant variables and the target fault, thereby improving the accuracy and efficiency of fault cause location.

[0091] In a possible implementation, a dependent variable is determined among multiple related variables according to the causal regression coefficient corresponding to each related variable, including: sorting the causal regression coefficient corresponding to each related variable to determine a sorted coefficient sequence, wherein the coefficient sequence includes the absolute value of the causal regression coefficient corresponding to each related variable, and the coefficients are sorted from large to small according to the absolute value; the absolute values ​​of the causal regression coefficient corresponding to each related variable included in the coefficient sequence are sequentially accumulated, and the cumulative percentage determined after each accumulation is compared with a preset threshold; when the cumulative percentage corresponding to the first k related variables in the coefficient sequence is less than or equal to the preset threshold, and the cumulative percentage corresponding to the first k+1 related variables in the coefficient sequence is greater than the preset threshold, the first k related variables in the coefficient sequence are determined as dependent variables.

[0092] Specifically, the causal regression coefficient sequence obtained can be The causal regression coefficients corresponding to each relevant variable in are sorted from small to large according to their absolute values ​​to obtain the sorted coefficient sequence

[0093] The absolute value of the causal regression coefficient corresponding to each relevant variable included in the coefficient sequence is accumulated in turn, and after each accumulation, the corresponding cumulative percentage can be obtained.

[0094] In one example, the cumulative percentage can be expressed as formula (7):

[0095]

[0096] Among them, η represents the cumulative percentage; k represents the current cumulative number of times.

[0097] The preset threshold can be set according to actual usage requirements. When the cumulative percentage corresponding to the first k related variables in the coefficient sequence is less than or equal to the preset threshold, and the cumulative percentage corresponding to the first k+1 related variables in the coefficient sequence is greater than the preset threshold, the first k related variables in the coefficient sequence are determined as dependent variables.

[0098] In one example, the preset threshold is 30%, the cumulative percentage corresponding to the first two related variables in the coefficient sequence is 20%, and the cumulative percentage corresponding to the first three related variables in the coefficient sequence is 80%. At this time, it can be determined that the number of dependent variables is 2, and the first two related variables in the coefficient sequence are dependent variables.

[0099] In particular, when the absolute value of the causal regression coefficient corresponding to the first related variable in the coefficient sequence is greater than a preset threshold, the first related variable in the coefficient sequence can be directly determined as the dependent variable.

[0100] In one example, the preset threshold is 30%, and the absolute value of the causal regression coefficient corresponding to the first related variable in the coefficient sequence is 50%. At this time, the first related variable in the coefficient sequence can be directly determined as the dependent variable.

[0101] In a possible implementation, based on a Bayesian network model, the intervention probability corresponding to each dependent variable is determined, including: determining a causal graph corresponding to a target fault according to the dependent variable and a label data sequence corresponding to the target system, wherein the label data sequence is used to represent the working state of the target system within a preset time; discretizing each dependent variable according to the data sequence of each dependent variable within the preset time to obtain a discretized variable corresponding to each dependent variable; constructing a Bayesian network model according to the causal graph and the discretized variables corresponding to each dependent variable; determining a conditional probability table corresponding to the target fault and an edge probability corresponding to each node in the causal graph according to the Bayesian network model, wherein the conditional probability table is used to represent the probability of a target fault occurring when each dependent variable is abnormal, and the edge probability is used to represent the probability of an abnormality occurring in each dependent variable; determining the intervention probability corresponding to each dependent variable according to the conditional probability table and the edge probability.

[0102] When a target failure occurs in the target system, a tag data sequence of the target system within a preset time can be obtained. The tag data sequence can be used to indicate the working state of the target system within the preset time. The specific form of the tag data sequence can refer to the implementation in the relevant technology, and this disclosure does not specifically limit this.

[0103] In one example, the preset time includes time T1 to time T4, the target system is in a normal state at time T1 and time T2, and is in a target fault state at time T3 and time T4. When the target system is in a normal state, the label data corresponding to the target system can be determined to be 0, and when the target system is in a target fault state, the label data corresponding to the target system can be determined to be 1. Therefore, the label data sequence of the target system from time T1 to time T4 can be determined as Y=[0,0,1,1].

[0104] The specific method of obtaining the tag data sequence can refer to the implementation method in the relevant technology, and the present disclosure does not make any specific limitation on this.

[0105] According to the label data sequence corresponding to the dependent variable and the target system, the causal graph corresponding to the target fault can be determined. The specific method of determining the causal graph can refer to the implementation method in the relevant technology. For example, k dependent variables and the label data sequence corresponding to the target system can be formed into a matrix, and the DAG-NOTEARS structure learning algorithm can be used to perform Bayesian network structure learning to obtain the causal graph, etc. The embodiments of the present disclosure do not specifically limit this.

[0106] According to the data sequence of each dependent variable within the preset time and the normal sample mean corresponding to each dependent variable, each dependent variable can be discretized to obtain a discretized variable corresponding to each dependent variable. The specific method of discretization can refer to the implementation method in the relevant technology, and this disclosure does not make specific limitations on this.

[0107] In one example, the discretized variable can be expressed as formula (8):

[0108]

[0109] Among them, "1" indicates that the target system has a target fault, and "0" indicates that the target system is in a normal state; t represents the preset discretization threshold. Using formula (8), each dependent variable can be discretized into "0" and "1".

[0110] Based on the causal graph and the discretized variables corresponding to each dependent variable, a Bayesian network model containing the dependent variable can be constructed. Compared with directly constructing a Bayesian network model based on all relevant variables, the amount of calculation required for model construction is reduced and the efficiency of model construction is improved.

[0111] Using the Bayesian estimation method, the network parameters of the Bayesian network model can be learned to determine the conditional probability table corresponding to the target fault and the edge probability corresponding to each node in the causal graph. The conditional probability table is used to represent the probability of the target fault occurring when each dependent variable is abnormal, and the edge probability is used to represent the probability of each dependent variable being abnormal; based on the conditional probability table and the edge probability, the intervention probability corresponding to each dependent variable can be determined.

[0112] The specific method of determining the intervention probability corresponding to each dependent variable according to the conditional probability table and the marginal probability can refer to the implementation methods in the relevant technology, for example, using the backdoor adjustment formula to calculate the intervention probability, etc., and the present disclosure does not make specific limitations on this.

[0113] According to the intervention probability corresponding to each dependent variable, the fault cause corresponding to the target fault can be directly determined. Compared with the reasoning method of inferring the fault cause based on conditional probability in traditional fault tracing, it can eliminate the influence of confounding factors and mediating variables, improve the accuracy of fault tracing results, and avoid the instability of fault tracing results caused by noise.

[0114] In one example, the target system is a feeding system, and the target fault is a feed ratio change fault during the feeding process, which is a step fault. The dependent variables corresponding to the feed ratio change fault during the feeding process extracted by the stable learning algorithm are related variables 1, related variables 17, related variables 31, related variables 42, and related variables 44. Using the above-mentioned dependent variables and the label data sequence corresponding to the target system, the causal graph corresponding to the feed ratio change fault can be obtained.

[0115] Figure 7 A cause-effect diagram corresponding to a feed ratio change fault according to an embodiment of the present disclosure is shown. Figure 7 As shown, the target fault does not point to dependent variable 17, so there is no real causal relationship between dependent variable 17 and the target fault. In the subsequent fault tracing process, the influence of dependent variable 17 on the target fault can no longer be considered. Based on the above method, the intervention probabilities corresponding to dependent variable 1, dependent variable 31, dependent variable 42 and dependent variable 44 can be determined respectively.

[0116] Referring to Table 1, for the feed ratio change fault, the corresponding conditional probability is determined for each dependent variable based on the traditional Bayesian network fault tracing method, and the corresponding intervention probability is determined for each dependent variable based on the causal analysis-based fault tracing method of the embodiment of the present disclosure.

[0117] Table 1

[0118]

[0119] According to Table 1, the conditional probability corresponding to dependent variable 1 is 95.1%, and the conditional probability corresponding to dependent variable 44 is 95.2%. The conditional probabilities corresponding to the two are very close, so the cause of the fault cannot be directly determined based on the conditional probability. The intervention probability corresponding to dependent variable 1 is 10.0%, and the intervention probability corresponding to dependent variable 44 is 95.5%. Therefore, the cause of the fault can be directly determined to be dependent variable 44 based on the intervention probability.

[0120] According to the known information of the target system, it can be determined that dependent variable 44 is an operating variable, indicating the feed amount; dependent variable 1 is a process variable, indicating the feed flow rate. Since in the actual production process, the operating variable affects the process variable, the dependent variable 44 is the cause of the target fault. The tracing result determined by the fault tracing method based on causal analysis disclosed in the present invention is consistent with the actual situation.

[0121] In one example, the target system is a feed system, and the target fault is a feed temperature fault, which is a random fault. The dependent variables corresponding to the feed temperature fault are extracted using a stable learning algorithm as related variables 7, 11, 13, and 18. Using the above-mentioned dependent variables and the label data sequence corresponding to the target system, the causal graph corresponding to the feed temperature fault can be obtained.

[0122] Figure 8 FIG. 2 shows a cause-effect diagram corresponding to a feed temperature fault according to an embodiment of the present disclosure. Figure 8 As shown, the target fault points to dependent variable 7, dependent variable 11, dependent variable 13 and dependent variable 18. Based on the above method, the intervention probabilities corresponding to dependent variable 7, dependent variable 11, dependent variable 13 and dependent variable 18 can be determined respectively.

[0123] Referring to Table 2, for the feed temperature fault, the corresponding conditional probability is determined for each dependent variable based on the traditional Bayesian network fault tracing method, and the corresponding intervention probability is determined for each dependent variable based on the causal analysis-based fault tracing method of the embodiment of the present disclosure.

[0124] Table 2

[0125]

[0126] According to Table 2, the conditional probability corresponding to dependent variable 7 is 53.2%, the conditional probability corresponding to dependent variable 13 is 52.3%, and the conditional probability corresponding to dependent variable 18 is 54.7%. The conditional probabilities corresponding to the three are almost equal. Therefore, it is impossible to directly determine the cause of the fault based on the conditional probability. The intervention probability corresponding to dependent variable 7 is 2.2%, the intervention probability corresponding to dependent variable 13 is 32.3%, and the intervention probability corresponding to dependent variable 18 is 76.4%. Therefore, the cause of the fault can be directly determined to be dependent variable 18 based on the intervention probability.

[0127] According to the known information of the target system, it can be determined that the dependent variable 18 represents the temperature of the stripper system. Since in the actual production process, the raw materials first enter the stripper system, the abnormal sampling value of the stripper system temperature will cause the target fault, so the dependent variable 18 is the fault cause causing the target fault. The tracing result determined by the fault tracing method based on causal analysis disclosed in the present invention is consistent with the actual situation.

[0128] Therefore, for a target system that needs to perform fault tracing, no matter whether the target fault is a step fault or a random fault, the fault tracing method based on causal analysis disclosed in the present invention can effectively improve the accuracy of the fault tracing result.

[0129] In one example, the target system is a complex industrial system for hot rolling process, including multiple components and involving multiple different production processes. Specifically, the hot rolling process is the process of processing slab into steel plate. The whole process can be divided into heating, rough rolling, flying shear, finishing rolling, laminar cooling and coiling, among which the finishing rolling process is the core part of the whole process.

[0130] Referring to Table 3, two faults in the finishing rolling process are taken as target faults, wherein target fault 1 is a fault of the stand roll gap control actuator, and target fault 2 is a fault of the cooling water valve actuator between the stands.

[0131] Table 3

[0132]

[0133] The relevant variables of the data set to be processed corresponding to the target fault are the roll gap, rolling force and bending roll force of each stand, where the bending roll force is sampled from the second stand and includes data when the target system is in a normal state and when the target system is in a target fault state.

[0134] Through the above method, the causal regression coefficient corresponding to each related variable corresponding to the target fault 1 and the dependent variable corresponding to the target fault 1 can be determined.

[0135] Fig. 9 A histogram of causal regression coefficients corresponding to a failure of a stand roll gap control actuator according to an embodiment of the present disclosure is shown. Fig. 9 As shown, the dependent variable corresponding to target fault 1 is only related variable 4. Based on the prior knowledge of the target system, it can be known that dependent variable 4 represents the roll gap of the frame. Therefore, it can be determined that when the frame roll gap control actuator fails, the sampling value of the roll gap of the frame becomes abnormal, which is consistent with the actual situation in the production process.

[0136] Through the above method, the causal regression coefficient corresponding to each related variable corresponding to the target fault 2 and the dependent variable corresponding to the target fault 2 can be determined.

[0137] Fig.10 A histogram of causal regression coefficients corresponding to a cooling water valve actuator fault between racks according to an embodiment of the present disclosure is shown. Fig.10 As shown, the dependent variables corresponding to the target fault 2 are the related variables 3 and the related variables 10. Through the above method, using the dependent variables and the label data sequence corresponding to the target system, a causal graph of the target fault 2 and the dependent variables 3 and 10 can be drawn.

[0138] Fig.11 FIG. 1 shows a cause-effect diagram corresponding to a cooling water valve actuator failure between racks according to an embodiment of the present disclosure. Fig.11As shown, target fault 2 points to dependent variable 3 and dependent variable 10, and both dependent variable 3 and dependent variable 10 have a causal relationship with target fault 2.

[0139] Referring to Table 4, for target fault 2, the corresponding conditional probability is determined for each dependent variable based on the traditional Bayesian network fault tracing method, and the corresponding intervention probability is determined for each dependent variable based on the causal analysis-based fault tracing method of the embodiment of the present disclosure.

[0140] Table 4

[0141]

[0142] According to Table 4, the conditional probability corresponding to dependent variable 3 is 96.6%, and the conditional probability corresponding to dependent variable 10 is 99.8%. The conditional probabilities corresponding to the two are almost equal. Therefore, the cause of the fault cannot be directly determined based on the conditional probability. The intervention probability corresponding to dependent variable 3 is 50%, and the intervention probability corresponding to dependent variable 10 is 99.8%. Therefore, the cause of the fault can be directly determined to be dependent variable 10 based on the intervention probability.

[0143] Based on the prior knowledge of the target system, it can be known that the dependent variable 10 represents the rolling force of the stand. In the actual production process, when a fault occurs, the rolling force of the stand is first abnormally changed, which in turn affects the roll gap of the stand, which is consistent with the actual situation. Therefore, the dependent variable 10 is the cause of the fault that causes the target fault. The tracing result determined by the fault tracing method based on causal analysis disclosed in the present invention is consistent with the actual situation.

[0144] Therefore, for a complex system with many components and production processes, the fault tracing method based on causal analysis disclosed in the present invention can effectively improve the accuracy and efficiency of the fault tracing results.

[0145] In an embodiment of the present disclosure, a data set to be processed is obtained, and the data set to be processed includes multiple related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and the data sequence of each related variable within a preset time. Based on the stable learning algorithm, a causal analysis is performed on the data set to be processed, and the dependent variable that has a causal relationship with the target fault is determined among the multiple related variables, so that the key variables related to the target fault can be determined, the number of candidate related variables is reduced, and the accuracy and efficiency of fault cause location are improved. Based on the Bayesian network model, the intervention probability corresponding to each dependent variable is determined. According to the intervention probability corresponding to each dependent variable, the fault cause corresponding to the target fault can be directly determined. Compared with the reasoning method based on conditional probability in traditional fault tracing, it can eliminate the influence of confounding factors and mediating variables, improve the accuracy of the fault tracing results, and avoid the instability of the fault tracing results caused by noise.

[0146] It can be understood that the above-mentioned various method embodiments mentioned in the present disclosure can be combined with each other to form a combined embodiment without violating the principle logic. Due to space limitations, the present disclosure will not repeat them. It can be understood by those skilled in the art that in the above-mentioned method of the specific implementation method, the specific execution order of each step should be determined according to its function and possible internal logic.

[0147] In addition, the present disclosure also provides a fault tracing device, electronic device, computer-readable storage medium, and program based on causal analysis, all of which can be used to implement any fault tracing method based on causal analysis provided by the present disclosure. The corresponding technical solutions and descriptions are referred to the corresponding records in the method part and will not be repeated here.

[0148] Fig.12 FIG. 2 is a block diagram of a fault tracing device based on causal analysis according to an embodiment of the present disclosure. Fig.12 As shown, the apparatus 1200 includes:

[0149] The data acquisition module 1201 acquires a data set to be processed, wherein the data set to be processed includes a plurality of related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and a data sequence of each related variable within a preset time;

[0150] A dependent variable extraction module 1202 is used to perform causal analysis on the data set to be processed, and determine the dependent variable that has a causal relationship with the target fault among multiple related variables;

[0151] An intervention probability determination module 1203 is used to determine the intervention probability corresponding to each dependent variable based on a Bayesian network model, wherein the intervention probability corresponding to each dependent variable is used to represent the probability value that the dependent variable is the cause of the fault;

[0152] The fault source determination module 1204 is used to determine the fault cause corresponding to the target fault according to the intervention probability corresponding to each dependent variable.

[0153] In a possible implementation, the data acquisition module 1201 is further used to: when a target failure occurs in the target system, acquire an original data set corresponding to the target system; perform data standardization processing on the original data set to determine a data set to be processed.

[0154] In one possible implementation, the dependent variable extraction module 1202 is also used to: group the data sets to be processed to determine a processed data group and a control data group; determine the causal regression coefficient corresponding to each related variable based on the processed data group and the control data group based on a stable learning algorithm, wherein the causal regression coefficient corresponding to each related variable is used to represent the correlation between the related variable and the target fault; determine the dependent variable among multiple related variables based on the causal regression coefficient corresponding to each related variable.

[0155] In a possible implementation, the dependent variable extraction module 1202 is also used to: determine the indicator vector corresponding to each relevant variable based on the normal sample mean corresponding to each relevant variable, wherein the normal sample mean corresponding to each relevant variable is the average value of the relevant variable when the target system is in a normal state; and divide the data set to be processed into a processing data group and a control data group based on the indicator vector corresponding to each relevant variable.

[0156] In one possible implementation, the dependent variable extraction module 1202 is also used to: determine a target optimization problem based on a processed data group and a control data group, wherein the target optimization problem is used to extract dependent variables from a plurality of related variables; solve the target optimization problem using a stable learning algorithm to determine the causal regression coefficient corresponding to each related variable.

[0157] In one possible implementation, the dependent variable extraction module 1202 is also used to: sort the causal regression coefficient corresponding to each related variable to determine a sorted coefficient sequence, wherein the coefficient sequence includes the absolute value of the causal regression coefficient corresponding to each related variable, and is sorted from large to small according to the absolute value; sequentially accumulate the absolute value of the causal regression coefficient corresponding to each related variable included in the coefficient sequence, and compare the cumulative percentage determined after each accumulation with a preset threshold; when the cumulative percentage corresponding to the first k related variables in the coefficient sequence is less than or equal to the preset threshold, and the cumulative percentage corresponding to the first k+1 related variables in the coefficient sequence is greater than the preset threshold, the first k related variables in the coefficient sequence are determined as dependent variables.

[0158] In a possible implementation, the intervention probability determination module 1203 is also used to: determine the causal graph corresponding to the target fault based on the dependent variable and the label data sequence corresponding to the target system, wherein the label data sequence is used to represent the working state of the target system within a preset time; discretize each dependent variable according to the data sequence of each dependent variable within the preset time to obtain the discretized variable corresponding to each dependent variable; construct a Bayesian network model based on the causal graph and the discretized variables corresponding to each dependent variable; determine the conditional probability table corresponding to the target fault and the edge probability corresponding to each node in the causal graph based on the Bayesian network model, wherein the conditional probability table is used to represent the probability of the target fault occurring when each dependent variable is abnormal, and the edge probability is used to represent the probability of each dependent variable being abnormal; determine the intervention probability corresponding to each dependent variable based on the conditional probability table and the edge probability.

[0159] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0160] The embodiment of the present disclosure also provides a computer-readable storage medium on which computer program instructions are stored, and the computer program instructions implement the above method when executed by a processor. The computer-readable storage medium can be a volatile or non-volatile computer-readable storage medium.

[0161] An embodiment of the present disclosure further proposes an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0162] The embodiments of the present disclosure also provide a computer program product, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0163] Fig.13 1 shows a block diagram of an electronic device according to an embodiment of the present disclosure. For example, the electronic device 1900 may be provided as a server or a terminal device. Fig.13 , the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions executable by the processing component 1922, such as an application. The application stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute instructions to perform the above method.

[0164] The device 1900 may also include a power supply component 1926 configured to perform power management of the device 1900, a wired or wireless network interface 1950 configured to connect the device 1900 to a network, and an input / output (I / O) interface 1958. The device 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server™, MacOS X™, Unix™, Linux™, FreeBSD™, or the like.

[0165] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the device 1900 to perform the above method.

[0166] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0167] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0168] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0169] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0170] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0171] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0172] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0173] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0174] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A fault tracing method based on causal analysis, characterized in that: include: Acquire a data set to be processed, wherein the data set to be processed includes a plurality of related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and a data sequence of each of the related variables within a preset time; Performing causal analysis on the data set to be processed, and determining a dependent variable having a causal relationship with the target fault from among the multiple related variables; Based on the Bayesian network model, determining the intervention probability corresponding to each of the dependent variables, wherein the intervention probability corresponding to each of the dependent variables is used to represent the probability value that the dependent variable is the cause of the fault; According to the intervention probability corresponding to each of the dependent variables, the fault cause corresponding to the target fault is determined.

2. The method according to claim 1, characterized in that The performing causal analysis on the data set to be processed and determining a dependent variable from the multiple related variables includes: Grouping the data set to be processed to determine a processing data group and a control data group; According to the processed data group and the control data group, based on a stable learning algorithm, determining a causal regression coefficient corresponding to each of the related variables, wherein the causal regression coefficient corresponding to each of the related variables is used to represent the correlation between the related variable and the target fault; The dependent variable is determined among the multiple related variables according to the causal regression coefficient corresponding to each of the related variables.

3. The method according to claim 2, characterized in that The step of grouping the data set to be processed to determine a processing data group and a control data group includes: Determine an indicator vector corresponding to each of the related variables according to the normal sample mean corresponding to each of the related variables, wherein the normal sample mean corresponding to each of the related variables is the average value of the related variable when the target system is in a normal state; According to the indicator vector corresponding to each of the relevant variables, the data set to be processed is divided into the processing data group and the control data group.

4. The method according to claim 2 or 3, characterized in that: Determining the causal regression coefficient corresponding to each of the related variables based on the processing data group and the control data group and based on a stable learning algorithm includes: Determining a target optimization problem based on the processed data set and the control data set, wherein the target optimization problem is used to extract the dependent variable from the multiple related variables; The target optimization problem is solved by using the stable learning algorithm to determine the causal regression coefficient corresponding to each of the related variables.

5. The method according to claim 2 or 3, characterized in that: Determining the dependent variable from the plurality of related variables according to the causal regression coefficient corresponding to each of the related variables includes: Sorting the causal regression coefficients corresponding to each of the related variables to determine a sorted coefficient sequence, wherein the coefficient sequence includes the absolute value of the causal regression coefficients corresponding to each of the related variables, and the coefficients are sorted from large to small according to the absolute value; The absolute value of the causal regression coefficient corresponding to each of the related variables included in the coefficient sequence is sequentially accumulated, and the cumulative percentage determined after each accumulation is compared with a preset threshold value; When the cumulative percentages corresponding to the first k related variables in the coefficient sequence are less than or equal to the preset threshold, and the cumulative percentages corresponding to the first k+1 related variables in the coefficient sequence are greater than the preset threshold, the first k related variables in the coefficient sequence are determined as the dependent variables.

6. The method according to claim 1, characterized in that Determining the intervention probability corresponding to each dependent variable based on the Bayesian network model includes: Determine a cause-effect graph corresponding to the target fault according to the dependent variable and the label data sequence corresponding to the target system, wherein the label data sequence is used to represent the working state of the target system within the preset time; According to the data sequence of each dependent variable within a preset time, each dependent variable is discretized to obtain a discretized variable corresponding to each dependent variable; Constructing a Bayesian network model according to the causal graph and the discretized variables corresponding to each of the dependent variables; According to the Bayesian network model, determine the conditional probability table corresponding to the target fault and the edge probability corresponding to each node in the causal graph, wherein the conditional probability table is used to represent the probability of the target fault occurring when each of the dependent variables is abnormal, and the edge probability is used to represent the probability of each of the dependent variables being abnormal; The intervention probability corresponding to each of the dependent variables is determined according to the conditional probability table and the marginal probability.

7. The method according to claim 1, characterized in that The method further comprises: When the target failure occurs in the target system, obtaining an original data set corresponding to the target system; The original data set is subjected to data standardization processing to determine the data set to be processed.

8. A fault tracing device based on causal analysis, characterized in that: include: A data acquisition module, which acquires a data set to be processed, wherein the data set to be processed includes a plurality of related variables corresponding to the target fault in the target system when the target fault occurs in the target system, and a data sequence of each of the related variables within a preset time; A dependent variable extraction module, used to perform causal analysis on the data set to be processed, and determine a dependent variable having a causal relationship with the target fault from among the multiple related variables; An intervention probability determination module, used to determine the intervention probability corresponding to each of the dependent variables based on a Bayesian network model, wherein the intervention probability corresponding to each of the dependent variables is used to represent the probability value that the dependent variable is the cause of the fault; The fault source determination module is used to determine the fault cause corresponding to the target fault according to the intervention probability corresponding to each of the dependent variables.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images