Hardware defect risk distribution prediction method, device, equipment, storage medium and program product

Abstract

The invention discloses a hardware defect risk distribution prediction method, device and equipment, a storage medium and a program product, and the method comprises the steps: carrying out the preprocessing of historical monitoring data of a target line, and obtaining an event set and a feature set, and constructing a mapping matrix between the fault events and defect related elements based on a mapping relationship between the event set and the feature set, and constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship. And constructing a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph, inputting the real-time monitoring data of the target line into the hardware defect prediction model for risk prediction, and obtaining hardware defect risk distribution information of the target line, thereby accurately mining the relevance between a fault event and a defect related element, and improving the accuracy of hardware defect prediction. And the prediction accuracy of the hardware defect is improved, so that the stability of a power system is improved, and hardware maintenance resources are reasonably allocated to different areas.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method, device, equipment, storage medium and program product for predicting risk distribution of hardware defects. Background Art

[0002] In recent years, the scale of photovoltaic and wind power resource development has become increasingly large. As the main transportation channel for electric energy in the power transmission strategy, the safe and stable operation of medium- and ultra-high-voltage transmission lines has become an important part of the construction goals of modern new power systems. As a key defect-related element for connecting and combining various types of cables, power fittings play a role in supporting, protecting, connecting and stabilizing various components of the transmission line, and are of great significance to maintaining the overall safe and stable operation of the line.

[0003] The existing maintenance of power fittings is mainly completed manually, and manual inspections generally take the form of pole climbing inspections and visual observations along the transmission lines. Pole climbing inspections mainly include checking whether power fittings (wires, connectors, fasteners, lightning arresters, etc.) are loose, damaged or rusted, and using insulating tools to test the electrical insulation performance of the fittings; visual observations mainly include checking the working status of power fittings and checking whether insulators, connectors, etc. are broken, deformed, or dusty. However, manual inspections inevitably require maintenance personnel to travel through mountains, hills, jungles and other environments, which is inefficient and has high costs and safety risks; visual observations are extremely dependent on the work experience of maintenance personnel. For some hidden fault risks, it is difficult to recognize their severity by naked eye observation alone, and the maintenance effect is not good. However, the use of helicopters and other equipment to detect hardware risks has the problems of high equipment operating costs and a small inspection range. Summary of the invention

[0004] The main purpose of the present invention is to provide a method, device, equipment, storage medium and program product for predicting the risk distribution of hardware defects, aiming to solve the technical problems of low accuracy and efficiency in the prior art of hardware fault risk detection and inability to detect hidden fault risks.

[0005] To achieve the above object, the present invention provides a method for predicting the risk distribution of hardware defects, the method comprising the following steps:

[0006] Preprocessing the historical monitoring data of the target line to obtain an event set and a feature set, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data;

[0007] Constructing a mapping matrix between fault events and defect-related elements based on a mapping relationship between the event set and the feature set;

[0008] Constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship;

[0009] Constructing a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph;

[0010] The real-time monitoring data of the target line is input into the hardware defect prediction model for risk prediction to obtain the hardware defect risk distribution information of the target line.

[0011] Optionally, constructing a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph includes:

[0012] Dividing the target line into hardware defect intervals according to the hardware defect risk knowledge graph to obtain multiple initial intervals;

[0013] Performing range correction on the initial interval based on the historical monitoring data to obtain a target interval;

[0014] Obtaining defect assessment indicators, and generating fuzzy rule sets corresponding to each target interval based on the defect assessment indicators;

[0015] generating a defect assessment structure based on the target interval and the fuzzy rule set, wherein the defect assessment structure includes a fuzzy risk level of each target interval, defect-related elements, and a hazard weight corresponding to each defect-related element;

[0016] According to the hardware defect risk knowledge graph, the defect assessment structure is weighted to obtain a target rule structure;

[0017] A hardware defect prediction model is constructed based on the target rule structure and the mapping matrix.

[0018] Optionally, the hardware defect prediction model includes:

[0019]

[0020] Among them, t i represents the i-th fault event in the mapping matrix M, m and n are the total number of records in M ​​and the total number of defect-related elements occurring in X, respectively. i ∈M(i,1)| is the cardinality of all internal standard simultaneously qualified records, x i,j Indicates t i Medium feature f j A defect-related element of the data, t i(x i,j ) indicates t i Medium i,j The specific value of Represents defect related elements c k The importance of represents a threshold coefficient, which is determined based on the fuzzified risk level of the fuzzy rule set and the hazard weight of each defect-related element in the defect assessment structure.

[0021] Optionally, constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship includes:

[0022] Fault classification is performed on each fault event in the event set to obtain a fault type corresponding to each fault event;

[0023] Generate fault cause information of each fault type according to the defect-related elements of each feature data in the feature set;

[0024] Determine the disaster impact level, time inducing factors and space inducing factors of each fault type based on the fault cause information;

[0025] Determine the hardware defect weight of each fault event according to the disaster impact level, the time inducing factor and the space inducing factor;

[0026] A hardware defect risk knowledge graph is constructed based on the hardware defect weights and the mapping relationship.

[0027] Optionally, the preprocessing of the historical monitoring data of the target line to obtain an event set and a feature set includes:

[0028] Identify abnormal data from the historical monitoring data of the target line:

[0029]

[0030] Wherein, X is the data point to be detected in the historical monitoring data, μ is the average value of the historical monitoring data, σ is the standard deviation of the historical monitoring data, Z is the standard deviation of the value of the data point to be detected in the historical monitoring data from the average value, and the data point whose standard deviation exceeds the preset range is determined as abnormal data;

[0031] Eliminating abnormal data from the historical monitoring data based on the identification result;

[0032] Fill the missing data in the historical monitoring data to obtain candidate data:

[0033]

[0034] Among them, Xi represents the load sequence after filling the missing data, i is the time node, β1 and β2 represent the data weighting coefficients of the two time nodes before and after the two days;

[0035] Perform feature analysis on the candidate data to obtain an event set and a feature set.

[0036] Optionally, performing feature analysis on the candidate data to obtain an event set and a feature set includes:

[0037] Acquire time series information of each modal data in the candidate data;

[0038] Determine a sequence sampling frequency of each modal data based on the timing information;

[0039] Determine a sampling priority of each modality data according to the sequence sampling frequency, and determine the data to be synchronized among the candidate data based on the sampling priority;

[0040] Performing time synchronization on the candidate data based on the data to be synchronized:

[0041]

[0042] Wherein, x(t) is the predicted value at time t, x1 and x2 are the recorded values ​​at time points t1 and t2, and two known data points (t1, x1) and (t2, x2) before and after time point t are connected. The predicted value is used to interpolate the candidate data to synchronize the candidate data in time;

[0043] Perform feature analysis on the candidate data after time synchronization to obtain event sets and feature sets.

[0044] In addition, to achieve the above-mentioned purpose, the present invention also proposes a hardware defect risk distribution prediction device, the hardware defect risk distribution prediction device comprising:

[0045] A data processing module, used to pre-process the historical monitoring data of the target line, and obtain an event set and a feature set, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, and the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data;

[0046] A matrix construction module, used to construct a mapping matrix between fault events and the defect-related elements based on the mapping relationship between the event set and the feature set;

[0047] A knowledge graph construction module, used to construct a knowledge graph of hardware defect risks according to the hardware defect weight of each fault event in the event set and the mapping relationship;

[0048] A prediction model building module, used to build a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph;

[0049] The risk prediction module is used to input the real-time monitoring data of the target line into the hardware defect prediction model for risk prediction, so as to obtain the hardware defect risk distribution information of the target line.

[0050] In addition, to achieve the above-mentioned purpose, the present application also proposes a hardware defect risk distribution prediction device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the hardware defect risk distribution prediction method as described above.

[0051] In addition, to achieve the above-mentioned purpose, the present application also proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the hardware defect risk distribution prediction method as described above are implemented.

[0052] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the hardware defect risk distribution prediction method as described above.

[0053] The present invention obtains an event set and a feature set by preprocessing the historical monitoring data of the target line, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data; a mapping matrix between the fault event and the defect-related elements is constructed based on the mapping relationship between the event set and the feature set, a knowledge graph of hardware defect risks is constructed according to the hardware defect weights of each fault event in the event set and the mapping relationship, and a hardware defect prediction model is constructed based on the mapping matrix and the knowledge graph of hardware defect risks , the real-time monitoring data of the target line is input into the hardware defect prediction model for risk prediction, and the hardware defect risk distribution information of the target line is obtained; since the present invention constructs a mapping matrix between fault events and defect-related elements based on the mapping relationship between event sets and feature sets, it can effectively mine the relevant conditions and factors that lead to fault events, realize the mining of potential risks of hardware, build a prediction model based on the mapping matrix and the hardware defect risk knowledge graph, and perform risk prediction based on the prediction model, thereby accurately predicting the hardware defect risks in different regions and environments, effectively improving the accuracy and efficiency of hardware defect risk prediction, and realizing the reasonable allocation of hardware maintenance resources to different regions. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings herein are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the description, are used to explain the principles of the present application.

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0056] Figure 1 It is a structural schematic diagram of a hardware defect risk distribution prediction device in a hardware operating environment involved in an embodiment of the present invention;

[0057] Figure 2 It is a schematic diagram of the flow chart of the first embodiment of the method for predicting the risk distribution of hardware defects of the present invention;

[0058] Figure 3 It is a knowledge graph of hardware defect risk in an embodiment of a method for predicting hardware defect risk distribution of the present invention;

[0059] Figure 4 It is a structural block diagram of the first embodiment of the hardware defect risk distribution prediction device of the present invention.

[0060] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0061] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0062] Reference Figure 1 , Figure 1 It is a schematic diagram of the structure of a hardware defect risk distribution prediction device in a hardware operating environment involved in an embodiment of the present invention.

[0063] like Figure 1 As shown, the hardware defect risk distribution prediction device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these defect-related elements. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0064] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the hardware defect risk distribution prediction device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0065] like Figure 1 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a hardware defect risk distribution prediction program.

[0066] exist Figure 1In the fitting defect risk distribution prediction device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the fitting defect risk distribution prediction device of the present invention can be set in the fitting defect risk distribution prediction device, and the fitting defect risk distribution prediction device calls the fitting defect risk distribution prediction program stored in the memory 1005 through the processor 1001, and executes the fitting defect risk distribution prediction method provided by the embodiment of the present invention.

[0067] The embodiment of the present invention provides a method for predicting the risk distribution of hardware defects, referring to Figure 2 , Figure 2 It is a schematic flow chart of the first embodiment of the method for predicting the risk distribution of hardware defects of the present invention.

[0068] In this embodiment, the method for predicting the risk distribution of hardware defects includes the following steps:

[0069] Step S10: pre-process the historical monitoring data of the target line to obtain an event set and a feature set.

[0070] It should be understood that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a computer, a server, etc., or a terminal electronic device capable of realizing the above functions, etc. The following takes the hardware defect risk distribution prediction device (referred to as the prediction device) as an example to illustrate this embodiment and the following embodiments.

[0071] It should be noted that the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data.

[0072] It should be noted that the historical monitoring data may be hardware failure data collected by multiple data collection modules, and the historical monitoring data may be multimodal data, for example, the historical monitoring data may include image data, text data, etc.

[0073] It is understandable that preprocessing may include data cleaning, abnormal data removal, and missing data filling of historical monitoring data.

[0074] In a specific implementation, the prediction device can combine historical monitoring data, centrally process and classify these image and text data sets, and extract fault events in the historical monitoring data and the feature data corresponding to the fault events.

[0075] Furthermore, in order to eliminate the influence of abnormal data and improve prediction accuracy, the above step S10 may include:

[0076] Step S11: Identify abnormal data on the historical monitoring data of the target line;

[0077] Step S12: removing abnormal data from the historical monitoring data based on the recognition result;

[0078] Step S13: Filling the missing data in the historical monitoring data to obtain candidate data;

[0079] Step S14: Perform feature analysis on the candidate data to obtain an event set and a feature set.

[0080] It should be noted that this embodiment can identify abnormal data based on the following formula:

[0081]

[0082] Wherein, X is the data point to be detected in the historical monitoring data, μ is the average value of the historical monitoring data, σ is the standard deviation of the historical monitoring data, and Z is the standard deviation of the numerical value of the data point to be detected in the historical monitoring data from the average value. The data point whose standard deviation exceeds the preset range is determined to be abnormal data, that is, if the absolute value of the standard deviation exceeds the preset range, it is considered to be abnormal data.

[0083] Fill in the removed abnormal data by taking the weighted average of the data at the same time two days before and after the missing data, and fill in the missing values ​​in combination with the load change rate. Fill in the missing data after removing the abnormal data based on the following formula:

[0084]

[0085] Among them, X i represents the load sequence after filling the missing data, i is the time node, β1 and β2 represent the data weighting coefficients of the two time nodes before and after two days

[0086] Furthermore, in order to ensure that the data collected by the air-based sensing module and the ground-based sensing module can be aligned at a given time point, the above step S14 may include:

[0087] Step S141: obtaining the time series information of each modal data in the candidate data;

[0088] Step S142: determining the sequence sampling frequency of each modal data based on the time series information;

[0089] Step S143: determining the sampling priority of each modality data according to the sequence sampling frequency, and determining the data to be synchronized in the candidate data based on the sampling priority;

[0090] Step S144: performing time synchronization on the candidate data based on the data to be synchronized;

[0091] Step S145: Perform feature analysis on the candidate data after time synchronization to obtain an event set and a feature set.

[0092] It should be noted that this embodiment can timestamp and synchronize all data to ensure that the data collected by the air-based perception module and the ground-based perception module can be aligned at a given time point, ensuring that subsequent analysis and decision-making have complete time series information support, and ensuring the accuracy and availability of data analysis. The timestamp synchronization method is specifically: if the sampling frequency of two time series data sequences A and B is higher than that of sequence B, in order to fill the data value of sequence B at a certain time point of A, the two data values ​​in sequence B closest to time point A are interpolated. The specific formula is:

[0093]

[0094] Where x(t) is the predicted value at time t, x1 and x2 are the recorded values ​​at time points t1 and t2, and the two known data points (t1, x1) and (t2, x2) before and after time point t are connected. The predicted value is used to interpolate the candidate data to synchronize the candidate data in time.

[0095] Step S20: constructing a mapping matrix between fault events and the defect-related elements based on the mapping relationship between the event set and the feature set.

[0096] It should be noted that the mapping matrix may include fault events, feature types corresponding to the fault events, and defect-related elements contained in the feature types, and the mapping matrix may reflect the relevant factors and conditions that lead to the occurrence of each fault event.

[0097] In some embodiments, the prediction device can identify factors that are strongly correlated with hardware defects and pre-process the historical data monitored by all drones, robots, and ground monitors (freezing, forest fires, pollution flashovers, lightning strikes, strong winds, and line dancing) to form a unified mapping interval.

[0098] For example, let D = {t1, t2, ..., t i ,……t m} is an event set (fault event record), where i = 1, 2, ..., m represents an event. Assume F = {f1, f2, ..., f j ,…,f n ,fY} is the feature set in D, where i = 1, 2, ..., n represents one of the total n features, f Y Represents the prediction target feature. Each feature consists of a set of defect-related elements, f j ={c j,1 ,c j,2 ,…,c j,k ,…,c j,l}∈F where i=1,2,...,l represents f j One of all l defect-related elements in .

[0099] In an embodiment, a defect-related element c j,k is considered as a variable x u,j , and is contained in the variable set X = {x i,1 ,x i,2 ,…,x i,j ,…,x i,n}, that is, the fault event t i All relevant conditional factors in . Assume that Y = {y1,y2,…,y i ,……y m} is a set of target defect related elements, each target defect related element y i is the fault event t i The final result.

[0100] The preprocessed data mapping space can be represented as a mapping matrix M, including submatrices D = [t1, t2, ..., t i ,……t m ] T ,F=[f1,f2,…,f j ,…,f n ,f Y ], X m×n ,Y=[y1,y2,…,y m ] T , for the distribution analysis of discrete features and continuous features, the submatrix x is divided into two parts: d j Represents the elements and c in the discrete features o Represents the elements in the continuous feature. The mapping matrix M can be expressed as:

[0101]

[0102] Each row of the mapping matrix represents a fault event record, and the left part f1,…,f d is a discrete feature, the right part f d+1 ,…,f n represents a continuous feature. i,j Representative record iMedium feature f j A defect related element c j,k Therefore, in a record t i In, x i1 ,…,x in is the corresponding defect-related element of each feature, y i is the final result.

[0103] Step S30: constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship.

[0104] It should be noted that the mapping relationship can represent the degree of association between the hardware defects and defect-related elements in each fault event, and the hardware defect weight can reflect the importance of each defect-related element in causing the hardware defect.

[0105] It should be understood that this embodiment can construct a power fitting defect risk map by analyzing the specific degree of correlation between the strong correlation factors and the fitting defects, wherein the entities in the fitting defect risk knowledge map can be the fault event type and the fitting defect type, and the fitting defect weight can define the importance of the entity. Figure 3 , Figure 3 The figure is a schematic diagram of the structure of the knowledge graph of hardware defect risk in one embodiment, wherein fault events may include wildfire, ice cover, line dancing, bird dirt, etc. The importance of each entity is defined based on the hardware defect weight of each fault event, and the importance may include Vital, Crucial, Important, Staple and Minor in the order of important to minor.

[0106] Furthermore, in order to accurately construct a knowledge graph of hardware defect risks, the above step S30 may include:

[0107] Step S301: classifying each fault event in the event set to obtain a fault type corresponding to each fault event;

[0108] Step S302: generating fault cause information of each fault type according to the defect-related elements of each feature data in the feature set;

[0109] Step S303: determining the disaster impact level, time inducing factors and space inducing factors of each fault type based on the fault inducing information;

[0110] Step S304: determining the hardware defect weight of each fault event according to the disaster impact level, the time inducing factor and the space inducing factor;

[0111] Step S305: constructing a hardware defect risk knowledge graph based on the hardware defect weights and the mapping relationship.

[0112] In some embodiments, the prediction device may select a set of multimodal data such as images, videos, and texts collected by the spatiotemporal perception module and the ground-based perception module when the line accident occurs based on historical record data, and classify them according to fault types such as wildfires, ice cover, and line dancing; use the models of each data set to extract the key features in each modal data, and further analyze the perception capabilities of the multimodal data features to the target.

[0113] In some embodiments, the prediction device can design a mechanism for identifying the main causes of common disasters (wildfires, ice cover, strong winds, line dancing, etc.) based on the key features of the fault data in each module, explore the internal connection mechanism of high-risk factors and inducing factors in both time and space dimensions, classify the importance of types of hardware defects that cause disasters, and build a knowledge graph for the risk classification of hardware defects in new power systems under air-ground multi-source data.

[0114] Step S40: constructing a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph.

[0115] It should be noted that the hardware defect prediction model can be a deep learning model or a mathematical model for predicting hardware defects.

[0116] In some embodiments, the prediction device can rely on the power fittings defect risk map to construct a prediction method implementation path and improve the ability to predict the risk distribution of fittings defects.

[0117] In some embodiments, the prediction device may construct a data set based on the mapping matrix and the hardware defect risk knowledge graph. The data set may include fault event data, defect-related element data corresponding to the fault event data, and weight data corresponding to the defect-related elements, and a hardware defect prediction model is constructed based on the data set.

[0118] Furthermore, in order to construct a high-performance hardware defect prediction model, the above step S40 may include:

[0119] Step S401: dividing the target line into hardware defect intervals according to the hardware defect risk knowledge graph to obtain a plurality of initial intervals;

[0120] Step S402: performing range correction on the initial interval based on the historical monitoring data to obtain a target interval;

[0121] Step S403: obtaining defect assessment indicators, and generating fuzzy rule sets corresponding to each target interval based on the defect assessment indicators;

[0122] Step S404: generating a defect assessment structure based on the target interval and the fuzzy rule set;

[0123] Step S405: weighting the defect assessment structure according to the hardware defect risk knowledge graph to obtain a target rule structure;

[0124] Step S406: constructing a hardware defect prediction model based on the target rule structure and the mapping matrix.

[0125] It should be noted that the defect assessment structure includes the fuzzy risk level of each target interval, defect-related elements, and the hazard weight corresponding to each defect-related element.

[0126] In some embodiments, the prediction device can divide the hardware defect interval into four intervals: less important, less important, important, and more important according to the hardware defect knowledge graph under the air-ground multi-source data. The boundaries of two adjacent sets are divided according to importance, and the overlapping range is determined according to historical statistical data to soften the clear boundary, thereby establishing a membership function for common hardware defects that cause line failures.

[0127] In some embodiments, the prediction device can be based on the level of the fault event in the energy system, according to the impact range, economic loss, power outage duration and other indicators divided into four levels: low, medium, high, and very high; build a fuzzy rule set, split the multi-level fuzzy structure into multiple subsets, and use the output of the previous subset as the input result of the next subset to solve, gradually solve the risk level, and divide the weight according to the knowledge graph, as shown in the following table:

[0128]

[0129] In some embodiments, since large-scale high-level events require more attention, higher risk weights should be assigned. The corresponding weights for low, medium, high, and very high are 0.16, 0.27, 0.42, and 0.75 respectively.

[0130] In some embodiments, for various features of hardware defects, each feature and its fuzzy set in the hardware defect state feature set are expanded. A fuzzy utility measurement mechanism under significant utility is designed to identify a high-utility model of hardware distribution with defects, and enrich the knowledge graph of hardware defect risk division under the new power system.

[0131] Furthermore, in order to improve the defect prediction performance, the hardware defect prediction model includes:

[0132]

[0133] Among them, t i represents the i-th fault event in the mapping matrix M, m and n are the total number of records in M ​​and the total number of defect-related elements occurring in X, respectively. i ∈M(i,1)| is the cardinality of all internal standard simultaneously qualified records, xi,j Indicates t i Medium feature f j A defect-related element of the data, t i (x i,j ) indicates t i Medium i,j The specific value of Represents defect related elements c k The importance of represents a threshold coefficient, which is determined based on the fuzzified risk level of the fuzzy rule set and the hazard weight of each defect-related element in the defect assessment structure.

[0134] It should be noted that a high-utility pattern X→Y can be further expanded to:<X,O> →

[0135] <Y→P> , where X and Y represent a set of factors and the target factor, respectively;

[0136] O={o 1,1 ,o 1,2 ,…,o i,j ,…,o m,n} and P = {p1,p2,…,p i ,…p m} respectively describe the corresponding fuzzy sets, which means: if X and O are associated, then it can be confirmed that Y and P are associated.<X,O> The mathematical expression can be expressed as:

[0137]

[0138] Among them, t i represents the i-th fault event in the mapping matrix M, m and n are the total number of records in M ​​and the total number of defect-related elements occurring in X, respectively. i ∈M(i,1)| is the cardinality of all internal standard simultaneously qualified records, x i,j Indicates t i Medium feature f j A defect-related element of the data, t i (x i,j ) indicates t i Medium i,j The specific value of Represents defect related elements c k The importance of represents a threshold coefficient, which is determined based on the fuzzified risk level of the fuzzy rule set and the hazard weight of each defect-related element in the defect assessment structure.

[0139] It should be noted that the threshold coefficient It can be calculated based on the following formula:

[0140]

[0141] in, Indicates that i (x i,j ) related fuzzy sets o i,j The fuzzy risk level, Indicates the corresponding hazard weight in the table, tu(t i ) is the fault event t i Utility, s is a significant parameter.

[0142] Assume two feature-fuzzy set pairs<X,O> and<Y,P> is a larger set of fault-related elements<Z,Q> A subset of And Z = X ∪ Y, And Q=O∪P. Let Z={z1,z2,...,z i ,...,z m} and Q={q1,q2,...,q i ,...,q m} is the fuzzy set corresponding to Z. Then these two feature-fuzzy set pairs <<X,O> ,<Y,P> >Confidence utility U c It is expressed as:

[0143]

[0144] Among them, t i is the i-th record in M, m and n are the total number of records and fault-related elements, x i,j Indicates t i Medium feature f i A fault-related element, z i is the corresponding fault-related element in Z, t i (z i ) Display t i Middle i The value of Y(r) represents a final result of the fault event, so It will take different forms when detecting different target fault-related elements. It shows the importance of fault-related elements, Φ qi [t i (z i )]and Two threshold parameters, Φ qi [t i (z i )]and It can be obtained by:

[0145]

[0146] in, or Describes the i (z i ) related fuzzy set q i The fuzzy risk level, or Represents and t i (x i,j ) Related i,j The risk level, and is the relevant risk weight, which can be calculated according to the above table, tu(t i ) is the fault event t i The utility of c is the confidence coefficient.

[0147] Step S50: inputting the real-time monitoring data of the target line into the hardware defect prediction model for risk prediction to obtain hardware defect risk distribution information of the target line.

[0148] It should be noted that the real-time monitoring data may be monitoring data of the target line collected by the prediction device through a data acquisition system. For example, the data acquisition system may be a sky-to-ground system.

[0149] In some embodiments, the prediction device can collect monitoring data of the target line through the space-based collection module, the air-based collection module and the ground-based collection module in the sky-ground system. The monitoring data may include image data, video data, text data, etc.

[0150] This embodiment obtains an event set and a feature set by preprocessing the historical monitoring data of the target line, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data; a mapping matrix between fault events and the defect-related elements is constructed based on the mapping relationship between the event set and the feature set, a knowledge graph of hardware defect risks is constructed according to the hardware defect weights of each fault event in the event set and the mapping relationship, and a hardware defect prediction model is constructed based on the mapping matrix and the knowledge graph of hardware defect risks. , the real-time monitoring data of the target line is input into the hardware defect prediction model for risk prediction, and the hardware defect risk distribution information of the target line is obtained; since this embodiment constructs a mapping matrix between fault events and defect-related elements based on the mapping relationship between event sets and feature sets, it effectively mines the relevant conditions and factors that lead to fault events, realizes the mining of potential risks of hardware, and constructs a prediction model based on the mapping matrix and the hardware defect risk knowledge graph, and performs risk prediction based on the prediction model, thereby accurately predicting the hardware defect risks in different regions and environments, effectively improving the accuracy and efficiency of hardware defect risk prediction, and thus realizing the reasonable allocation of hardware maintenance resources to different regions.

[0151] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which a hardware defect risk distribution prediction program is stored. When the hardware defect risk distribution prediction program is executed by a processor, the steps of the hardware defect risk distribution prediction method described above are implemented.

[0152] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0153] The above-mentioned computer-readable storage medium may be included in the hardware defect risk distribution prediction device; or it may exist independently without being assembled into the hardware defect risk distribution prediction device.

[0154] In addition, an embodiment of the present invention further proposes a computer program product, including a hardware defect risk distribution prediction program, which implements the steps of the hardware defect risk distribution prediction method as described above when executed by a processor.

[0155] The specific implementation methods of the computer program product of the present invention are basically the same as the embodiments of the above-mentioned hardware defect risk distribution prediction method, and will not be repeated here.

[0156] Reference Figure 4 , Figure 4 It is a structural block diagram of the first embodiment of the hardware defect risk distribution prediction device of the present invention.

[0157] like Figure 4 As shown, the hardware defect risk distribution prediction device proposed in the embodiment of the present invention includes:

[0158] The data processing module 10 is used to pre-process the historical monitoring data of the target line to obtain an event set and a feature set, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, and the feature data is composed of one or more defect-related elements corresponding to the fault event. The historical monitoring data includes air-based monitoring data and ground-based monitoring data;

[0159] A matrix construction module 20, configured to construct a mapping matrix between fault events and the defect-related elements based on a mapping relationship between the event set and the feature set;

[0160] A knowledge graph construction module 30, configured to construct a knowledge graph of hardware defect risks according to the hardware defect weight of each fault event in the event set and the mapping relationship;

[0161] A prediction model building module 40, used to build a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph;

[0162] The risk prediction module 50 is used to input the real-time monitoring data of the target line into the hardware defect prediction model for risk prediction, so as to obtain the hardware defect risk distribution information of the target line.

[0163] Furthermore, the prediction model construction module 40 is also used to divide the target line into hardware defect intervals according to the hardware defect risk knowledge graph to obtain multiple initial intervals; perform range correction on the initial intervals based on the historical monitoring data to obtain a target interval; obtain defect assessment indicators, and generate a fuzzy rule set corresponding to each target interval based on the defect assessment indicators; generate a defect assessment structure based on the target interval and the fuzzy rule set, the defect assessment structure including the fuzzy risk level of each target interval, defect-related elements and hazard weights corresponding to each defect-related element; divide the weights of the defect assessment structure according to the hardware defect risk knowledge graph to obtain a target rule structure; and construct a hardware defect prediction model based on the target rule structure and the mapping matrix.

[0164] Furthermore, the hardware defect prediction model includes:

[0165]

[0166] Among them, t i represents the i-th fault event in the mapping matrix M, m and n are the total number of records in M ​​and the total number of defect-related elements occurring in X, respectively. i ∈M(i,1)| is the cardinality of all internal standard simultaneously qualified records, x i,j Indicates t i Medium feature fj A defect-related element of the data, t i (x i,j ) indicates t i Medium i,j The specific value of Represents defect related elements c k The importance of represents a threshold coefficient, which is determined based on the fuzzified risk level of the fuzzy rule set and the hazard weight of each defect-related element in the defect assessment structure.

[0167] Furthermore, the knowledge graph construction module 30 is also used to classify each fault event in the event set to obtain the fault type corresponding to each fault event; generate fault cause information of each fault type according to the defect-related elements of each feature data in the feature set; determine the disaster impact level, time inducing factors and space inducing factors of each fault type based on the fault cause information; determine the hardware defect weight of each fault event according to the disaster impact level, the time inducing factors and the space inducing factors; and construct a hardware defect risk knowledge graph based on the hardware defect weight and the mapping relationship.

[0168] Furthermore, the data processing module 10 is also used to identify abnormal data in the historical monitoring data of the target line:

[0169]

[0170] Wherein, X is the data point to be detected in the historical monitoring data, μ is the average value of the historical monitoring data, σ is the standard deviation of the historical monitoring data, Z is the standard deviation of the value of the data point to be detected in the historical monitoring data from the average value, and the data point whose standard deviation exceeds the preset range is determined to be abnormal data;

[0171] Eliminating abnormal data from the historical monitoring data based on the identification result;

[0172] Fill the missing data in the historical monitoring data to obtain candidate data:

[0173]

[0174] Among them, X i represents the load sequence after filling the missing data, i is the time node, β1 and β2 represent the data weighting coefficients of the two time nodes before and after the two days;

[0175] Perform feature analysis on the candidate data to obtain an event set and a feature set.

[0176] Further, the data processing module 10 is also used to obtain the timing information of each modal data in the candidate data; determine the sequence sampling frequency of each modal data based on the timing information; determine the sampling priority of each modal data according to the sequence sampling frequency, and determine the data to be synchronized in the candidate data based on the sampling priority; perform time synchronization on the candidate data based on the data to be synchronized:

[0177]

[0178] Wherein, x(t) is the predicted value at time t, x1 and x2 are the recorded values ​​at time points t1 and t2, and two known data points (t1, x1) and (t2, x2) before and after time point t are connected. The predicted value is used to interpolate the candidate data to synchronize the candidate data in time;

[0179] Perform feature analysis on the candidate data after time synchronization to obtain event sets and feature sets.

[0180] This embodiment obtains an event set and a feature set by preprocessing the historical monitoring data of the target line, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data; a mapping matrix between fault events and the defect-related elements is constructed based on the mapping relationship between the event set and the feature set, a knowledge graph of hardware defect risks is constructed according to the hardware defect weights of each fault event in the event set and the mapping relationship, and a hardware defect prediction model is constructed based on the mapping matrix and the knowledge graph of hardware defect risks. , the real-time monitoring data of the target line is input into the hardware defect prediction model for risk prediction, and the hardware defect risk distribution information of the target line is obtained; since this embodiment constructs a mapping matrix between fault events and defect-related elements based on the mapping relationship between event sets and feature sets, it effectively mines the relevant conditions and factors that lead to fault events, realizes the mining of potential risks of hardware, and constructs a prediction model based on the mapping matrix and the hardware defect risk knowledge graph, and performs risk prediction based on the prediction model, thereby accurately predicting the hardware defect risks in different regions and environments, effectively improving the accuracy and efficiency of hardware defect risk prediction, and thus realizing the reasonable allocation of hardware maintenance resources to different regions.

[0181] The hardware defect risk distribution prediction device provided by the present application adopts the hardware defect risk distribution prediction method in the above-mentioned embodiment, which can solve the technical problem of hardware defect risk distribution prediction. Compared with the prior art, the beneficial effects of the hardware defect risk distribution prediction device provided by the present application are the same as the beneficial effects of the hardware defect risk distribution prediction method provided by the above-mentioned embodiment, and the other technical features in the hardware defect risk distribution prediction device are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0182] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0183] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0184] In addition, for technical details not fully described in this embodiment, reference can be made to the hardware defect risk distribution prediction method provided in any embodiment of the present invention, and will not be repeated here.

[0185] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0186] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0187] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0188] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for predicting risk distribution of hardware defects, characterized in that: The hardware defect risk distribution prediction method comprises: Preprocessing the historical monitoring data of the target line to obtain an event set and a feature set, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data; Constructing a mapping matrix between fault events and defect-related elements based on a mapping relationship between the event set and the feature set; Constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship; Constructing a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph; The real-time monitoring data of the target line is input into the hardware defect prediction model for risk prediction to obtain the hardware defect risk distribution information of the target line.

2. The method for predicting the risk distribution of hardware defects according to claim 1, characterized in that: The constructing of a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph includes: Dividing the target line into hardware defect intervals according to the hardware defect risk knowledge graph to obtain multiple initial intervals; Performing range correction on the initial interval based on the historical monitoring data to obtain a target interval; Obtaining defect assessment indicators, and generating fuzzy rule sets corresponding to each target interval based on the defect assessment indicators; generating a defect assessment structure based on the target interval and the fuzzy rule set, wherein the defect assessment structure includes a fuzzy risk level of each target interval, defect-related elements, and a hazard weight corresponding to each defect-related element; According to the hardware defect risk knowledge graph, the defect assessment structure is weighted to obtain a target rule structure; A hardware defect prediction model is constructed based on the target rule structure and the mapping matrix.

3. The method for predicting the risk distribution of hardware defects according to claim 2, characterized in that: The hardware defect prediction model includes: Among them, t i represents the i-th fault event in the mapping matrix M, m and n are the total number of records in M ​​and the total number of defect-related elements occurring in X, respectively. i ∈M(i,1)| is the cardinality of all internal standard simultaneously qualified records, x i,j Indicates t i Medium feature f j A defect-related element of the data, t i (x i,j ) indicates t i Medium i,j The specific value of Represents defect related elements c k The importance of represents a threshold coefficient, which is determined based on the fuzzified risk level of the fuzzy rule set and the hazard weight of each defect-related element in the defect assessment structure.

4. The method for predicting the risk distribution of hardware defects according to any one of claims 1 to 3, characterized in that: The constructing a hardware defect risk knowledge graph according to the hardware defect weight of each fault event in the event set and the mapping relationship includes: Fault classification is performed on each fault event in the event set to obtain a fault type corresponding to each fault event; Generate fault cause information of each fault type according to the defect-related elements of each feature data in the feature set; Determine the disaster impact level, time inducing factors and space inducing factors of each fault type based on the fault cause information; Determine the hardware defect weight of each fault event according to the disaster impact level, the time inducing factor and the space inducing factor; A hardware defect risk knowledge graph is constructed based on the hardware defect weights and the mapping relationship.

5. The method for predicting the risk distribution of hardware defects according to any one of claims 1 to 3, characterized in that: The preprocessing of the historical monitoring data of the target line to obtain an event set and a feature set includes: Identify abnormal data from the historical monitoring data of the target line: Wherein, X is the data point to be detected in the historical monitoring data, μ is the average value of the historical monitoring data, σ is the standard deviation of the historical monitoring data, Z is the standard deviation of the value of the data point to be detected in the historical monitoring data from the average value, and the data point whose standard deviation exceeds the preset range is determined as abnormal data; Eliminating abnormal data from the historical monitoring data based on the identification result; Fill the missing data in the historical monitoring data to obtain candidate data: Among them, X i represents the load sequence after filling the missing data, i is the time node, β1 and β2 represent the data weighting coefficients of the two time nodes before and after the two days; Perform feature analysis on the candidate data to obtain an event set and a feature set.

6. The method for predicting the risk distribution of hardware defects according to claim 5, characterized in that: The performing feature analysis on the candidate data to obtain an event set and a feature set includes: Acquire time series information of each modal data in the candidate data; Determine a sequence sampling frequency of each modal data based on the timing information; Determine a sampling priority of each modality data according to the sequence sampling frequency, and determine the data to be synchronized among the candidate data based on the sampling priority; Performing time synchronization on the candidate data based on the data to be synchronized: Wherein, x(t) is the predicted value at time t, x1 and x2 are the recorded values ​​at time points t1 and t2, and two known data points (t1, x1) and (t2, x2) before and after time point t are connected. The predicted value is used to interpolate the candidate data to synchronize the candidate data in time; Perform feature analysis on the candidate data after time synchronization to obtain event sets and feature sets.

7. A device for predicting risk distribution of hardware defects, characterized in that: The hardware defect risk distribution prediction device comprises: A data processing module, used to pre-process the historical monitoring data of the target line, and obtain an event set and a feature set, wherein the event set includes multiple fault events, the feature set includes feature data corresponding to each fault event, and the feature data is composed of one or more defect-related elements corresponding to the fault event, and the historical monitoring data includes air-based monitoring data and ground-based monitoring data; A matrix construction module, used to construct a mapping matrix between fault events and the defect-related elements based on the mapping relationship between the event set and the feature set; A knowledge graph construction module, used to construct a knowledge graph of hardware defect risks according to the hardware defect weight of each fault event in the event set and the mapping relationship; A prediction model building module, used to build a hardware defect prediction model based on the mapping matrix and the hardware defect risk knowledge graph; The risk prediction module is used to input the real-time monitoring data of the target line into the hardware defect prediction model for risk prediction, so as to obtain the hardware defect risk distribution information of the target line.

8. A hardware defect risk distribution prediction device, characterized in that: The hardware defect risk distribution prediction device includes: a memory, a processor, and a hardware defect risk distribution prediction program stored in the memory and executable on the processor, wherein the hardware defect risk distribution prediction program is configured to implement the hardware defect risk distribution prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a hardware defect risk distribution prediction program, and when the hardware defect risk distribution prediction program is executed by the processor, the hardware defect risk distribution prediction method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product comprises a hardware defect risk distribution prediction program, and when the hardware defect risk distribution prediction program is executed by a processor, the steps of the hardware defect risk distribution prediction method according to any one of claims 1 to 6 are implemented.

Images