Automation equipment fault monitoring method and system based on industrial Internet
Through the automated equipment failure monitoring method based on the industrial Internet, the target selection thread calculates the possibility of equipment failure, solves the problem that equipment failure is difficult to predict in advance, improves the accuracy and credibility of fault monitoring, and reduces production interruptions and losses.
Patent Information
- Application Number
- CN202311519881.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-15
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2043-11-15
AI Technical Summary
Equipment failures are difficult to predict in advance, resulting in production interruptions and reduced efficiency. The existing technology cannot promptly notify relevant technical personnel for repair, resulting in immeasurable losses.
An automated equipment failure monitoring method based on the industrial Internet, by obtaining device object characteristics and operation data, using target selection threads to calculate the possibility of equipment failure, and processing it in combination with influencing elements, determining outliers to identify the final equipment failure information.
It improves the accuracy and credibility of equipment failure monitoring, ensures that the equipment failure information is highly matched with the requirements, and promptly notifies maintenance personnel, reducing production interruptions and losses.
Smart Images

Figure CN117575565B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data monitoring technology, and more specifically, to an automation equipment fault monitoring method and system based on the Industrial Internet. Background Art
[0002] Equipment failure refers to an event or phenomenon in which the equipment loses or reduces its specified function. It manifests itself as certain parts of the equipment losing their original accuracy or performance, causing the equipment to fail to operate normally, technical performance to be reduced, resulting in production interruption or reduced efficiency, affecting production.
[0003] At present, if the equipment fails and the relevant technicians cannot be notified in time for maintenance, it may cause immeasurable losses to the factory. However, it is difficult to predict the equipment failure in advance, which is a technical problem that is difficult to overcome at present. Therefore, a technical solution is urgently needed to improve the above technical problems. Summary of the invention
[0004] In order to improve the technical problems existing in the related technologies, the present application provides an automation equipment fault monitoring method and system based on the industrial Internet.
[0005] In a first aspect, a method for monitoring faults of automated equipment based on the industrial Internet is provided, wherein the method comprises:
[0006] Obtaining device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each of the target device operation data;
[0007] Input the device object characteristics of the target device object into each of the target selection threads to obtain a first device failure probability corresponding to each of the target selection threads, wherein the first device failure probability is used to represent the device failure probability extracted from the target device operation data of the corresponding information type of each of the target selection threads in the target device object;
[0008] Input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure probability corresponding to each of the target selection threads, wherein the second device failure probability is used to indicate the device failure probability extracted from the target device operation data of the corresponding information type of each target selection thread when interference is applied to the target device operation data of the corresponding information type, and the interference is used to increase the confidence of the target device operation data;
[0009] Obtaining a first device failure possibility and a second device failure possibility corresponding to each of the target device operation data, and obtaining an abnormal value of each of the target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each of the target device operation data;
[0010] The target device operation data with the highest abnormal value among the plurality of target device operation data is obtained and determined as the final device fault information corresponding to the target device object.
[0011] The step of obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data includes:
[0012] Calculate the comparison result of the first device failure probability and the second device failure probability corresponding to each of the target device operation data, and determine the device failure probability result of each of the target device operation data;
[0013] Obtaining a prior evaluation value corresponding to each of the target device operation data;
[0014] The equipment failure possibility result corresponding to each of the target equipment operation data and the pre-evaluation value are calculated and processed by function to determine the abnormal value of each of the target equipment operation data.
[0015] The step of inputting the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibilities corresponding to each of the target selection threads includes:
[0016] Processing the device object features of the target device object through each of the target selection threads to obtain a device object feature vector corresponding to each of the target selection threads;
[0017] Processing the influencing factors through each of the target selection threads to obtain interference feature vectors corresponding to each of the target selection threads;
[0018] Performing attention mechanism operation on the device object feature vector and the interference feature vector corresponding to each of the target selection threads to obtain the attention operation result corresponding to each of the target selection threads; through each of the target selection threads, processing the corresponding device object feature vector and the attention operation result to obtain the regression analysis device failure possibility output by each of the target selection threads, and determining the second device failure possibility corresponding to each of the target selection threads.
[0019] The step of performing attention mechanism operation on the device object feature vector and the interference feature vector corresponding to each of the target selection threads to obtain the attention operation result corresponding to each of the target selection threads includes:
[0020] According to the device object feature vector corresponding to each of the target selection threads, a key vector and a value vector corresponding to each of the target selection threads are obtained;
[0021] Calculate the key vector and interference feature vector corresponding to each of the target selection threads to perform function processing, and determine the function processing results corresponding to each of the target selection threads; perform clustering processing on the function processing results corresponding to each of the target selection threads to obtain clustering results corresponding to each of the target selection threads;
[0022] Analyze and process the clustering results corresponding to each of the target selection threads to obtain analysis and processing results corresponding to each of the target selection threads;
[0023] Performing dimensionless simplification processing on the analysis and processing results corresponding to each of the target selection threads to obtain dimensionless simplified results corresponding to each of the target selection threads;
[0024] Calculate the dimensionless simplified results and value vectors corresponding to each of the target selection threads for function processing, and determine the attention operation results corresponding to each of the target selection threads.
[0025] The process of processing the corresponding device object feature vectors and attention operation results through each of the target selection threads to obtain the regression analysis device failure possibility output by each of the target selection threads includes:
[0026] Performing fusion processing on vectors of several dimensions included in the device object feature vector corresponding to each of the target selection threads to obtain fusion vectors corresponding to each of the target selection threads;
[0027] The sum of the fusion vectors and the attention operation results corresponding to each of the target selection threads is calculated to determine the sum vectors corresponding to each of the target selection threads; through each of the target selection threads, the corresponding sum vectors are processed to obtain the possibility of regression analysis device failure output by each of the target selection threads.
[0028] The step of processing the corresponding sum vectors through each of the target selection threads to obtain the probability of regression analysis equipment failure output by each of the target selection threads includes:
[0029] Performing dimensionless simplification processing on the sum vector corresponding to each of the target selection threads to obtain a dimensionless simplified vector corresponding to each of the target selection threads;
[0030] Through each of the target selection threads, the corresponding dimensionless simplified vectors are processed to obtain the probability of regression analysis equipment failure output by each of the target selection threads.
[0031] The step of processing the corresponding dimensionless simplified vectors through each of the target selection threads to obtain the probability of regression analysis equipment failure output by each of the target selection threads includes:
[0032] By means of each of the target selection threads, the dimensionless simplified vectors corresponding to each of the target selection threads are processed to obtain a first vector corresponding to each of the target selection threads;
[0033] Performing fusion processing on the first vectors and interference vectors corresponding to each of the target selection threads to obtain the second vectors corresponding to each of the target selection threads;
[0034] Through each of the target selection threads, the corresponding second vectors are processed to obtain the probability of regression analysis device failure output by each of the target selection threads.
[0035] Among them, the obtaining of the device object characteristics of the target device object, a plurality of target device operating data and a target selection thread corresponding to the information type of each of the target device operating data includes: in response to a search request for a target display interface from the target device object, obtaining the device object characteristics of the target device object, the plurality of target device operating data and a target selection thread corresponding to the information type of each of the target device operating data.
[0036] Among them, the obtaining of the target selection thread corresponding to the information type of each target device operation data includes: determining the target selection thread corresponding to each target device operation data from a number of preset selection threads according to the information type of each target device operation data, and the several preset selection threads correspond to different information types respectively.
[0037] Wherein, the configuration method of the target selection thread includes: obtaining configuration examples corresponding to each of the information types, the configuration examples corresponding to each of the information types include the device object characteristics of the first device object, the device object characteristics of the second device object, the first extraction result corresponding to the first example information under the information type, and the second extraction result corresponding to the second example information under the information type, the first extraction result refers to the extraction result of the first device object for the first example information when no interference is applied to the first example information, and the second extraction result refers to the extraction result of the second device object for the second example information when interference is applied to the second example information; according to the configuration examples corresponding to each of the information types, the original thread is configured to obtain the target selection thread corresponding to each of the information types.
[0038] In a second aspect, an automation equipment fault monitoring system based on the industrial Internet is provided, comprising a processor and a memory that communicate with each other, wherein the processor is used to read and execute a computer program from the memory to implement the above method.
[0039] The industrial Internet-based automated equipment fault monitoring method and system provided in the embodiments of the present application obtain the first equipment fault possibility and the second equipment fault possibility corresponding to each target equipment operation data through a target selection thread, wherein the second equipment fault possibility reflects the equipment fault possibility of the target equipment operation data being extracted by the target equipment object after interference is imposed on the target equipment operation data, and the abnormal value obtained according to the first equipment failure possibility and the second equipment failure possibility can accurately reflect the target equipment object's demand for the target equipment operation data, thereby making the final equipment failure information determined according to the abnormal value have a higher matching degree with the equipment object's demand, thereby improving the accuracy and reliability of equipment fault monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0041] Figure 1 A flowchart of an automation equipment fault monitoring method based on the Industrial Internet provided in an embodiment of the present application. DETAILED DESCRIPTION
[0042] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.
[0043] See also Figure 1 , shows a method for monitoring faults of automated equipment based on the industrial Internet, which may include the technical solution described in the following steps S101-S150.
[0044] S110, obtaining device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each target device operation data.
[0045] In this embodiment, the target device operating data may refer to data acquired from operating devices on a production line.
[0046] Different target device operation data may correspond to different information types, which may include mechanical operation types, program operation types, and operation operation types. Different information types correspond to their own target selection threads.
[0047] In a possible implementation, the target selection thread corresponding to each target device operation data can be determined from a number of preset selection threads according to the information type of each target device operation data, and the plurality of preset selection threads correspond to different information types.
[0048] The electronic device may store a plurality of preset selection threads, each information type corresponds to a preset selection thread, obtain the information type of the target device operation data for the target device operation data, and determine the preset selection thread corresponding to the information type of the target device operation data as the target selection thread of the target device operation data. Different target device operation data may correspond to different target selection threads.
[0049] S120. Input the device object characteristics of the target device object into each target selection thread to obtain a first device failure possibility corresponding to each target selection thread. The first device failure possibility is used to represent the device failure possibility of the target device operation data of the corresponding information type of each target selection thread being extracted by the target device object.
[0050] The device object features of the target device object can be digitized into vectors, and the vectors after the device object features are digitized are input into each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and determine the first device failure possibility corresponding to each target selection thread.
[0051] For target device operation data under the same information type, the target selection thread of the information type determines that the device failure possibilities extracted by the target device object for each target device operation data under the information type are all first device failure possibilities.
[0052] S130. Input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibility corresponding to each target selection thread. The second device failure possibility is used to indicate the device failure possibility of the target device operation data extracted by the target device object when interference is applied to the target device operation data of the corresponding information type of each target selection thread. The interference is used to increase the confidence in the target device operation data.
[0053] The device object characteristics of the target device object can be digitized into a vector, and the vector of the digitized device object characteristics and the influencing factors can be input into each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and determine the second device failure possibility corresponding to each target selection thread.
[0054] For target device operation data under the same information type, the target selection thread of the information type determines that the device failure possibilities extracted by the target device object for each target device operation data under the information type are all second device failure possibilities.
[0055] In this embodiment, the influencing factor may refer to an interference factor (for example, a factor that interferes with the normal operation of the device). When the device object characteristics of the target device object and the influencing factor are input into the target selection thread, it represents that interference is exerted on the target device operation data corresponding to the target selection thread. At this time, the obtained second device failure possibility represents the possibility of device failure when interference is exerted on the target device operation data of the corresponding information type of each target selection thread, and the target device operation data is extracted by the target device object.
[0056] In some possible implementations, the vector of digitized device object features and non-influencing factors may also be input into each target selection thread to obtain a first device failure possibility output by each target selection thread, wherein the non-influencing factor may be 0. When the device object features and non-influencing factors of the target device object are input into the target selection thread, no interference is applied to the target device operation data corresponding to the target selection thread. At this time, the obtained first device failure possibility represents the device failure possibility of the target device operation data extracted by the target device object when no interference is applied to the target device operation data of the corresponding information type of each target selection thread.
[0057] In a possible implementation embodiment, a configuration method of a target selection thread (the configuration method of a preset selection thread refers to the configuration method of a target selection thread and will not be described in detail) includes: obtaining configuration examples corresponding to each information type, the configuration examples corresponding to each information type include a device object feature of a first device object, a device object feature of a second device object, a first extraction result corresponding to the first example information under the information type, and a second extraction result corresponding to the second example information under the information type, the first extraction result refers to an extraction result of the first device object for the first example information when no interference is applied to the first example information, and the second extraction result refers to an extraction result of the second device object for the second example information when interference is applied to the second example information; according to the configuration examples corresponding to each information type, the original thread is configured to obtain a target selection thread corresponding to each information type.
[0058] For each information type, the configuration example corresponding to the information type may include the device object characteristics of the first device object, the device object characteristics of the second device object, the first extraction results corresponding to each of the several first example information under the information type, and the second extraction results corresponding to each of the several second example information under the information type; the device object characteristics of the first device object, the non-influencing factors, and the first extraction results corresponding to each of the several first example information under the information type may be determined as non-interference configuration examples to configure the original thread; at the same time, the device object characteristics of the second device object, the influencing factors, and the second extraction results corresponding to each of the several second example information under the information type are determined as interference configuration examples to configure the original thread; when the non-interference configuration example and the interference configuration example are both configured, the original thread is determined as the target selection thread corresponding to the information type.
[0059] The device object characteristics and non-influencing factors of the first device object may be input into the original thread to obtain the device failure possibilities extracted by the first device object corresponding to each of the first example information output by the original thread, and the loss value is determined according to the device failure possibilities extracted by the first device object corresponding to each of the first example information and the first extraction results corresponding to each of the first example information, and the original thread is configured by the determined loss value, thereby realizing the configuration of the original thread by the non-interference configuration example. At the same time, the device object characteristics and influencing factors of the second device object may be input into the original thread to obtain the device failure possibilities extracted by the second device object corresponding to each of the second example information output by the original thread, and the loss value is determined according to the device failure possibilities extracted by the second device object corresponding to each of the second example information and the second extraction results corresponding to each of the second example information, and the original thread is configured by the determined loss value, thereby realizing the configuration of the original thread by the interference configuration example.
[0060] It can be understood that through the target selection thread of the present application, accurate regression analysis can be performed to determine the change in the possibility of device failure extracted by the target device object in the target table information before and after interference is applied to the target device operation data (the comparison result distance between the first device failure possibility and the second device failure possibility), thereby determining whether to display the target device operation data based on the change in the device failure possibility.
[0061] S140, obtaining the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data.
[0062] Input the device object characteristics of the target device object into each target selection thread to obtain the first device failure possibility corresponding to each target selection thread, and input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibility corresponding to each target selection thread. For each target device operation data, obtain the first device failure possibility and the second device failure possibility output by the target selection thread corresponding to the information type of the target device operation data, and determine them as the first device failure possibility and the second device failure possibility corresponding to the target device operation data.
[0063] In a possible implementation, a method for determining abnormal values of each target device operation data based on a first device failure possibility and a second device failure possibility corresponding to each target device operation data may include: calculating a comparison result of the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and determining the device failure possibility result of each target device operation data; obtaining a prior evaluation value corresponding to each target device operation data; calculating the device failure possibility result and the prior evaluation value corresponding to each target device operation data, performing function processing, and determining them as abnormal values of each target device operation data.
[0064] In this embodiment, the abnormal value of each target device operation data is determined based on the first device failure possibility and the second device failure possibility of each target device operation data. Since the second device failure possibility accurately reflects the device failure possibility of the target device operation data being extracted by the device object after interference is applied to the target device operation data, the abnormal value of each target device operation data can accurately reflect the degree of demand for the target device operation data by the device object. The higher the abnormal value of a target device operation data, the higher the requirement of the target device object for the target device operation data.
[0065] S150: Obtain target device operation data with the highest abnormal value among a plurality of target device operation data, and determine it as final device failure information of the corresponding target device object.
[0066] After obtaining the abnormal values of several target device operation data, the target device operation data with the highest abnormal value is extracted from each target device operation data to be determined as the final device fault information. The final device fault information has the highest abnormal value, and the final device fault information is the target device operation data that the target device object is most likely to pay attention to.
[0067] In a possible implementation, after obtaining the final device fault information, the final device fault information can be displayed to the target device object. After the final device fault information is displayed, the device fault probability extracted by the target device object from the final device fault information is the highest among all target device operation data.
[0068] In this embodiment, the first device failure possibility and the second device failure possibility corresponding to each target device operation data are obtained through the target selection thread, wherein the second device failure possibility reflects the device failure possibility of the target device operation data being extracted by the target device object after interference is imposed on the target device operation data, and the abnormal value obtained according to the first device failure possibility and the second device failure possibility can accurately reflect the target device object's demand for the target device operation data, thereby making the final device failure information determined according to the abnormal value highly matched with the demand of the device object, thereby improving the accuracy and reliability of equipment fault monitoring.
[0069] Another embodiment of the present application proposes an automation equipment fault monitoring method based on the industrial Internet. The method can be applied to electronic equipment and includes the following content.
[0070] S210: Obtain device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each target device operation data.
[0071] S220. Input the device object characteristics of the target device object into each target selection thread to obtain a first device failure possibility corresponding to each target selection thread. The first device failure possibility is used to represent the device failure possibility of the target device operation data of the corresponding information type of each target selection thread being extracted by the target device object.
[0072] The description of S210 - S220 refers to the description of S110 - S120 above, which will not be repeated here.
[0073] S230, processing the device object characteristics of the target device object through each target selection thread to obtain the device object characteristic vector corresponding to each target selection thread; and processing the influencing factors through each target selection thread to obtain the interference characteristic vector corresponding to each target selection thread.
[0074] Each target selection thread may include a feature representation module, which is used to convert the device object features and influencing factors of the target device object into a vector form to obtain the device object feature vector and interference feature vector corresponding to each target selection thread.
[0075] The target selection thread includes a feature representation module, which includes a device object feature unit and a factor unit. The device object feature unit and the factor unit in the feature representation module can both be composed of an input layer of a neural network, so that after the device object feature of the target device object is input into the device object feature unit, a device object feature vector is obtained, and after the influencing factor is input into the factor unit, an interference feature vector is obtained.
[0076] S240, performing attention mechanism operation on the device object feature vector and interference feature vector corresponding to each target selection thread to obtain the attention operation result corresponding to each target selection thread.
[0077] The attention mechanism operation is performed on the device object feature vector and the interference feature vector corresponding to each target selection thread to obtain the attention operation result. Through the attention mechanism operation, the device object feature vector and the interference feature vector can be well integrated, so that the obtained attention operation result accurately reflects the corresponding state of the target device operation data after the interference is applied, thereby effectively improving the accuracy of the second device failure possibility obtained according to the attention operation result.
[0078] In a possible implementation embodiment, S240 may include: obtaining a key vector and a value vector corresponding to each target selection thread according to a device object feature vector corresponding to each target selection thread; calculating the key vector and the interference feature vector corresponding to each target selection thread for function processing, and determining the function processing result corresponding to each target selection thread; clustering the function processing results corresponding to each target selection thread to obtain the clustering results corresponding to each target selection thread; analyzing the clustering results corresponding to each target selection thread to obtain the analysis and processing results corresponding to each target selection thread; performing dimensionless simplification processing on the analysis and processing results corresponding to each target selection thread to obtain the dimensionless simplification results corresponding to each target selection thread; calculating the dimensionless simplification results and the value vector corresponding to each target selection thread for function processing, and determining them as the attention operation results corresponding to each target selection thread.
[0079] S250. Process the corresponding device object feature vectors and attention operation results through each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and determine the second device failure possibility corresponding to each target selection thread.
[0080] After obtaining the corresponding attention calculation results and device object feature vectors for each target selection thread, the attention calculation results and device object feature vectors corresponding to each target selection thread can be summed, and the summed result can be processed by the target selection thread to obtain the output of the second device failure possibility.
[0081] The target selection thread also includes an output regression analysis module, which uses the attention mechanism module to sum the device object feature vector corresponding to the target selection thread and the attention operation result to obtain the sum result, and then processes the sum result through the output regression analysis module to obtain the second device failure possibility output by the target selection thread.
[0082] S260, obtaining the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data.
[0083] S270: Obtain target device operation data with the highest abnormal value among a plurality of target device operation data, and determine it as final device fault information of the corresponding target device object.
[0084] Among them, the description of S260-S270 refers to the description of S140-S150 mentioned above, and will not be repeated here.
[0085] In this embodiment, an attention mechanism operation is performed on the device object feature vector and the interference feature vector corresponding to each target selection thread to obtain an attention operation result. Through the attention mechanism operation, the device object feature vector and the interference feature vector can be better fused, so that the obtained attention operation result accurately reflects the corresponding state of the target device operating data after interference is applied, thereby effectively improving the accuracy of the second device failure possibility obtained according to the attention operation result.
[0086] Another embodiment of the present application proposes a flowchart of an automation equipment fault monitoring method based on the industrial Internet, which can be applied to electronic equipment and includes the following content.
[0087] S310, obtaining device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each target device operation data.
[0088] S320. Input the device object characteristics of the target device object into each target selection thread to obtain the first device failure possibility corresponding to each target selection thread. The first device failure possibility is used to represent the device failure possibility of the target device operation data of the corresponding information type of each target selection thread being extracted by the target device object.
[0089] S330, processing the device object characteristics of the target device object through each target selection thread to obtain the device object characteristic vector corresponding to each target selection thread; and processing the influencing factors through each target selection thread to obtain the interference characteristic vector corresponding to each target selection thread.
[0090] S340, performing attention mechanism operation on the device object feature vector and interference feature vector corresponding to each target selection thread to obtain the attention operation result corresponding to each target selection thread.
[0091] Among them, the description of S310-S340 refers to the description of S210-S240 above, and will not be repeated here.
[0092] S350, fusing the vectors of several dimensions included in the device object feature vector corresponding to each target selection thread to obtain the fusion vector corresponding to each target selection thread; calculating the sum of the fusion vector corresponding to each target selection thread and the result of the attention operation to determine the sum vector corresponding to each target selection thread.
[0093] Through each target selection thread, the input device object features of several dimensions are processed to obtain vectors of several dimensions corresponding to each target selection thread, and the vectors of several dimensions are determined as the device object feature vector corresponding to the target selection thread. Through each target selection thread, the vectors of several dimensions included in the device object feature vector are fused to obtain a fused vector, which is determined as the fused vector corresponding to each target selection thread.
[0094] After obtaining the fusion vector and attention operation result corresponding to each target selection thread, the fusion vector and attention operation result of each target selection thread are summed to obtain the sum vector corresponding to each target selection thread.
[0095] The attention mechanism module may also include a first fusion unit and a summation unit. The first fusion unit is used to fuse vectors of several dimensions included in the device object feature vector corresponding to each target selection thread to obtain a fusion vector corresponding to each target selection thread; the summation unit is used to calculate the sum of the fusion vector corresponding to each target selection thread and the attention operation result, and determine the sum vector corresponding to each target selection thread.
[0096] S360. Process the corresponding sum vectors through each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and determine the second device failure possibility corresponding to each target selection thread.
[0097] After the sum vector is obtained, the corresponding sum vectors are processed through each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and the second device failure possibility corresponding to each target selection thread is determined.
[0098] In a possible implementation embodiment, S360 may include: performing dimensionless simplification processing on the sum vector corresponding to each target selection thread to obtain a dimensionless simplified vector corresponding to each target selection thread; through each target selection thread, processing the corresponding dimensionless simplified vector to obtain the regression analysis device failure possibility output by each target selection thread, and determining the second device failure possibility corresponding to each target selection thread.
[0099] The attention mechanism module may also include a second normalization unit, which performs dimensionless simplification processing on the sum vector corresponding to the target selection thread through the second normalization unit to obtain the dimensionless simplified vector corresponding to each target selection thread, and then processes the dimensionless simplified vector through the output regression analysis module in the target selection thread to obtain the possibility of regression analysis device failure output by the target selection thread.
[0100] By performing dimensionless simplification on the sum vector, the numerical size of the data is reduced, thereby improving the processing efficiency of the target selection thread on the sum vector, thereby improving the efficiency of obtaining the possibility of regression analysis equipment failure.
[0101] In another possible implementation, S360 may also include: performing dimensionless simplification processing on the sum vector corresponding to each target selection thread to obtain a dimensionless simplified vector corresponding to each target selection thread; processing the corresponding dimensionless simplified vector through each target selection thread to obtain a first vector corresponding to each target selection thread; fusing the first vector and the interference vector corresponding to each target selection thread to obtain a second vector corresponding to each target selection thread; processing the corresponding second vector through each target selection thread to obtain the regression analysis device failure possibility output by each target selection thread, and determining the second device failure possibility corresponding to each target selection thread.
[0102] The dimensionless simplified vector may be processed by the first fully connected network in the target selection thread to obtain the processed first vector, and then the first vector corresponding to the target selection thread and the interference vector are fused to obtain the second vector corresponding to the target selection thread, and then the second vector corresponding to the target selection thread is processed by the second fully connected network and the third fully connected network in the target selection thread in turn to obtain the output result, and the output result is processed by the activation function in the target selection thread to obtain the second device failure possibility.
[0103] The dimensionless simplified vector is processed by the first fully connected network to obtain a first vector, and the first vector and the interference feature vector are fused by the second fusion unit to obtain a second vector; the second vector is then input into the second fully connected network, and the result output by the second fully connected network is input into the third fully connected network to obtain the output result of the third fully connected network, and then the output result is processed by the activation function to obtain the second device failure possibility.
[0104] S370, obtaining the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data.
[0105] S380: Obtain target device operation data with the highest abnormal value among a plurality of target device operation data, and determine it as final device failure information of the corresponding target device object.
[0106] Among them, the description of S370-S380 refers to the description of S140-S150 above, and will not be repeated here.
[0107] In this embodiment, the vectors of several dimensions included in the device object feature vector are fused, and the fused vector after the fusion processing is summed with the attention operation result. The sum vector corresponding to the summation result is processed to obtain the output regression analysis device failure possibility, which is determined as the second device failure possibility. The sum vector fuses the vectors of several dimensions included in the user feature vector and the attention operation result, so that the sum vector more accurately reflects the device failure possibility when the target device operation data is extracted by the target device object after interference, thereby improving the accuracy of the second device failure possibility.
[0108] Another embodiment of the present application proposes a flowchart of an automation equipment fault monitoring method based on the industrial Internet. The method can be applied to electronic equipment, and the method includes the following steps.
[0109] S410 , in response to a search request of a target device object for a target display interface, obtaining device object characteristics of the target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each target device operation data.
[0110] S420. Input the device object characteristics of the target device object into each target selection thread to obtain a first device failure possibility corresponding to each target selection thread. The first device failure possibility is used to represent the device failure possibility of the target device operation data of the corresponding information type of each target selection thread being extracted by the target device object.
[0111] S430. Input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibility corresponding to each target selection thread. The second device failure possibility is used to indicate the device failure possibility of the target device operation data extracted by the target device object when interference is applied to the target device operation data of the corresponding information type of each target selection thread. The interference is used to increase the confidence in the target device operation data.
[0112] S440, obtaining the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data.
[0113] S450: Obtain target device operation data with the highest abnormal value among a plurality of target device operation data, and determine it as final device failure information of the corresponding target device object.
[0114] Among them, the description of S420-S450 refers to the description of S120-S150 above, and will not be repeated here.
[0115] S460. Display final device failure information on the target display interface.
[0116] After the final device failure information is obtained, the final device failure information is displayed on the target display interface of the terminal.
[0117] In one possible implementation, the target device object can send a search request through a terminal, and the terminal sends the search request to a server. The server is determined as the execution subject, and obtains final device failure information based on the search request. The server then returns the final device failure information to the terminal, and the terminal then outputs the final device failure information.
[0118] Another embodiment of the present application proposes a flowchart of an automation equipment fault monitoring method based on the industrial Internet. The method can be applied to electronic equipment, and the method includes the following steps.
[0119] S510: Obtain device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each target device operation data.
[0120] S520. Process the device object characteristics of the target device object through each target selection thread to obtain a device object characteristic vector corresponding to each target selection thread; process the non-influencing factors through each target selection thread to obtain a non-interference characteristic vector corresponding to each target selection thread.
[0121] Each target selection thread may include a feature representation module, which is used to convert the device object features and non-influencing elements of the target device object into vector form to obtain the device object feature vector and non-interference feature vector corresponding to each target selection thread.
[0122] S530. According to the device object feature vector corresponding to each target selection thread, obtain the key vector and value vector corresponding to each target selection thread; calculate the key vector and non-interference feature vector corresponding to each target selection thread, perform function processing, and determine the non-interference product result corresponding to each target selection thread; perform clustering processing on the non-interference product result corresponding to each target selection thread, and obtain the non-interference clustering result corresponding to each target selection thread; perform analysis processing on the non-interference clustering result corresponding to each target selection thread, and obtain the non-interference analysis processing result corresponding to each target selection thread; perform dimensionless simplification processing on the non-interference analysis processing result corresponding to each target selection thread, and obtain the non-interference dimensionless simplification result corresponding to each target selection thread; calculate the non-interference dimensionless simplification result and value vector corresponding to each target selection thread, perform function processing, and determine the non-interference attention operation result corresponding to each target selection thread.
[0123] S540. Fuse the vectors of several dimensions included in the device object feature vector corresponding to each target selection thread to obtain the fused vector corresponding to each target selection thread; calculate the sum of the fused vector corresponding to each target selection thread and the result of the non-interference attention operation to determine the non-interference sum vector corresponding to each target selection thread; perform dimensionless simplification on the non-interference sum vector corresponding to each target selection thread to obtain the non-interference dimensionless simplified vector corresponding to each target selection thread.
[0124] S550. Through each target selection thread, the corresponding non-interference dimensionless simplified vector is processed to obtain the third vector corresponding to each target selection thread; the third vector corresponding to each target selection thread and the non-interference vector are merged to obtain the fourth vector corresponding to each target selection thread; through each target selection thread, the corresponding fourth vector is processed to obtain the regression analysis device failure possibility output by each target selection thread, and it is determined to be the first device failure possibility.
[0125] Among them, the process of obtaining the possibility of failure of the first device in S530-S550 is similar to the calculation process of obtaining the possibility of failure of the second device mentioned above. The difference is that the influencing factors are replaced by non-influencing factors in the calculation process, and the intermediate results in the calculation process are also the intermediate results corresponding to the non-influencing factors, which will not be repeated here.
[0126] It can be understood that the non-influencing factor is 0, and the process of inputting the non-influencing factor and the device object characteristics of the target device object into the target selection thread for processing is actually equivalent to the process of only processing the device object characteristics of the target device object through the target selection thread.
[0127] S560. Input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibility corresponding to each target selection thread. The second device failure possibility is used to indicate the device failure possibility of the target device operation data extracted by the target device object when interference is applied to the target device operation data of the corresponding information type of each target selection thread. The interference is used to increase the confidence in the target device operation data.
[0128] S570, obtaining the first device failure possibility and the second device failure possibility corresponding to each target device operation data, and obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data.
[0129] S580: Obtain target device operation data with the highest abnormal value among a plurality of target device operation data, and determine it as final device failure information of the corresponding target device object.
[0130] Among them, the description of S560-S580 refers to the description of S130-S150 above and will not be repeated here.
[0131] In this embodiment, the device object characteristics including the target device object and the non-influencing factors can be processed through the target selection thread to obtain the first device failure possibility, so there is no need to redeploy a new thread. Only different values (influencing factors or non-influencing factors) need to be input to obtain the first device failure possibility or the second device failure possibility, thereby improving the applicability of the target selection thread.
[0132] On the basis of the above, an automation equipment fault monitoring device based on the industrial Internet is provided, and the device comprises:
[0133] A thread acquisition module, used to obtain device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each of the target device operation data;
[0134] A first possibility determination module is used to input the device object characteristics of the target device object into each of the target selection threads to obtain a first device failure possibility corresponding to each of the target selection threads, wherein the first device failure possibility is used to represent the device failure possibility extracted from the target device operation data of the corresponding information type of each of the target selection threads at the target device object;
[0135] A second possibility determination module is used to input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibility corresponding to each of the target selection threads, wherein the second device failure possibility is used to indicate the device failure possibility extracted from the target device operation data of the corresponding information type of each target selection thread when interference is applied to the target device operation data of the corresponding information type, and the interference is used to increase the confidence of the target device operation data;
[0136] an abnormal value determination module, used to obtain the first device failure possibility and the second device failure possibility corresponding to each of the target device operation data, and obtain the abnormal value of each of the target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each of the target device operation data;
[0137] The fault information determination module is used to obtain the target device operation data with the highest abnormal value among the plurality of target device operation data, and determine it as the final device fault information corresponding to the target device object.
[0138] On the basis of the above, an automation equipment fault monitoring system based on the industrial Internet is shown, including a processor and a memory that communicate with each other, and the processor is used to read and execute a computer program from the memory to implement the above method.
[0139] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.
[0140] In summary, based on the above scheme, the first device failure possibility and the second device failure possibility corresponding to each target device operation data are obtained through the target selection thread, wherein the second device failure possibility reflects the device failure possibility of the target device operation data being extracted by the target device object after interference is imposed on the target device operation data, and the abnormal value obtained according to the first device failure possibility and the second device failure possibility can accurately reflect the target device object's demand for the target device operation data, thereby making the final device failure information determined according to the abnormal value have a higher matching degree with the demand of the device object, thereby improving the accuracy and reliability of equipment fault monitoring.
[0141] It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present application can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software such as executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (e.g., firmware).
[0142] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other beneficial effects that may be obtained.
Claims
1. A method for monitoring faults of automated equipment based on the industrial Internet, characterized in that: The method comprises: Obtaining device object characteristics of a target device object, a plurality of target device operation data, and a target selection thread corresponding to an information type of each of the target device operation data; Input the device object characteristics of the target device object into each of the target selection threads to obtain a first device failure probability corresponding to each of the target selection threads, wherein the first device failure probability is used to represent the device failure probability extracted from the target device operation data of the corresponding information type of each of the target selection threads in the target device object; Input the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure probability corresponding to each of the target selection threads, wherein the second device failure probability is used to indicate the device failure probability extracted from the target device operation data of the corresponding information type of each target selection thread when interference is applied to the target device operation data of the corresponding information type, and the interference is used to increase the confidence of the target device operation data; Obtaining a first device failure possibility and a second device failure possibility corresponding to each of the target device operation data, and obtaining an abnormal value of each of the target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each of the target device operation data; Obtaining target device operation data with the highest abnormal value among the plurality of target device operation data, and determining the data as final device fault information corresponding to the target device object; The step of obtaining the abnormal value of each target device operation data according to the first device failure possibility and the second device failure possibility corresponding to each target device operation data includes: Calculate the comparison result of the first device failure probability and the second device failure probability corresponding to each of the target device operation data, and determine the device failure probability result of each of the target device operation data; Obtaining a prior evaluation value corresponding to each of the target device operation data; The equipment failure possibility result corresponding to each of the target equipment operation data and the pre-evaluation value are calculated and processed by function to determine the abnormal value of each of the target equipment operation data.
2. The method according to claim 1, characterized in that The step of inputting the device object characteristics and influencing factors of the target device object into the target selection thread to obtain the second device failure possibilities corresponding to each of the target selection threads includes: Processing the device object features of the target device object through each of the target selection threads to obtain a device object feature vector corresponding to each of the target selection threads; Processing the influencing factors through each of the target selection threads to obtain interference feature vectors corresponding to each of the target selection threads; Performing attention mechanism operation on the device object feature vector and the interference feature vector corresponding to each of the target selection threads to obtain the attention operation result corresponding to each of the target selection threads; through each of the target selection threads, processing the corresponding device object feature vector and the attention operation result to obtain the regression analysis device failure possibility output by each of the target selection threads, and determining the second device failure possibility corresponding to each of the target selection threads.
3. The method according to claim 2, characterized in that The performing attention mechanism operation on the device object feature vector and the interference feature vector corresponding to each of the target selection threads to obtain the attention operation result corresponding to each of the target selection threads includes: According to the device object feature vector corresponding to each of the target selection threads, a key vector and a value vector corresponding to each of the target selection threads are obtained; Calculate the key vector and interference feature vector corresponding to each of the target selection threads to perform function processing, and determine the function processing results corresponding to each of the target selection threads; perform clustering processing on the function processing results corresponding to each of the target selection threads to obtain clustering results corresponding to each of the target selection threads; Analyze and process the clustering results corresponding to each of the target selection threads to obtain analysis and processing results corresponding to each of the target selection threads; Performing dimensionless simplification processing on the analysis and processing results corresponding to each of the target selection threads to obtain dimensionless simplified results corresponding to each of the target selection threads; Calculate the dimensionless simplified results and value vectors corresponding to each of the target selection threads for function processing, and determine the attention operation results corresponding to each of the target selection threads.
4. The method according to claim 2, characterized in that The method of processing the corresponding device object feature vectors and attention operation results through each of the target selection threads to obtain the probability of device failure output by the regression analysis of each of the target selection threads includes: Performing fusion processing on vectors of several dimensions included in the device object feature vector corresponding to each of the target selection threads to obtain fusion vectors corresponding to each of the target selection threads; The sum of the fusion vectors and the attention operation results corresponding to each of the target selection threads is calculated to determine the sum vectors corresponding to each of the target selection threads; through each of the target selection threads, the corresponding sum vectors are processed to obtain the possibility of regression analysis device failure output by each of the target selection threads.
5. The method according to claim 4, characterized in that The step of processing the corresponding sum vectors through each of the target selection threads to obtain the probability of regression analysis of device failure output by each of the target selection threads includes: Performing dimensionless simplification processing on the sum vector corresponding to each of the target selection threads to obtain a dimensionless simplified vector corresponding to each of the target selection threads; Through each of the target selection threads, the corresponding dimensionless simplified vectors are processed to obtain the probability of regression analysis equipment failure output by each of the target selection threads.
6. The method according to claim 5, characterized in that The step of processing the corresponding dimensionless simplified vectors through each of the target selection threads to obtain the probability of regression analysis equipment failure output by each of the target selection threads includes: By means of each of the target selection threads, the dimensionless simplified vectors corresponding to each of the target selection threads are processed to obtain a first vector corresponding to each of the target selection threads; Performing fusion processing on the first vectors and interference vectors corresponding to each of the target selection threads to obtain the second vectors corresponding to each of the target selection threads; Through each of the target selection threads, the corresponding second vectors are processed to obtain the probability of regression analysis device failure output by each of the target selection threads.
7. The method according to claim 1, characterized in that The method of obtaining the device object characteristics of the target device object, a plurality of target device operation data, and a target selection thread corresponding to the information type of each of the target device operation data includes: in response to a search request for a target display interface from the target device object, obtaining the device object characteristics of the target device object, the plurality of target device operation data, and a target selection thread corresponding to the information type of each of the target device operation data.
8. The method according to claim 1, characterized in that Obtaining a target selection thread corresponding to an information type of each target device operation data comprises: determining a target selection thread corresponding to each target device operation data from a plurality of preset selection threads according to the information type of each target device operation data, wherein the plurality of preset selection threads correspond to different information types respectively.
9. An automation equipment fault monitoring system based on industrial Internet, characterized in that: The invention comprises a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the method according to any one of claims 1 to 8.
