A method for classifying offline fault causes of power IoT intelligent terminals

Abstract

The present invention discloses a method for classifying offline fault causes of an electric power Internet of Things smart terminal. The method comprises: obtaining historical time series data of the electric power Internet of Things smart terminal; inputting the historical time series data into a fault cause analysis model to obtain a fault cause probability set, wherein the fault cause analysis model comprises a GRU-DNN-Attention model combining a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; and determining the fault cause of the electric power Internet of Things smart terminal being offline according to the fault cause probability set. The method can determine the fault cause probability set of the electric power Internet of Things smart terminal being offline through the GRU-DNN-Attention model, and determine the fault cause of the electric power Internet of Things smart terminal being offline according to the fault cause probability set, thereby improving the operation and maintenance efficiency of the electric power Internet of Things smart terminal.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of power Internet of Things, and in particular to a method for classifying offline fault causes of a power Internet of Things intelligent terminal. Background Art

[0002] With the rapid construction of new power systems, the sensing objects extend from the power grid side to the source and load side, new energy and new services are continuously connected, the data scale is growing exponentially, and the number and scale of IoT smart terminals required for accurate collection of each link of source, grid, load and storage are also increasing year by year. Based on the massive data collected by the power IoT smart terminal, combined with the analysis of technical means such as artificial intelligence and big data, the operating status of the power system can be more accurately perceived and controlled. However, during the operation of the power IoT smart terminal, offline faults often occur. It is necessary to determine the cause of the offline fault as soon as possible and perform maintenance to minimize the offline time of the equipment. In the current maintenance work, manual experience is still the main method. The cause of the fault is determined by checking the on-site status of the equipment, and then the fault is handled. Therefore, the existing technology cannot quickly and accurately determine the cause of the smart terminal offline. Summary of the invention

[0003] The present invention provides a method for classifying offline fault causes of an electric power Internet of Things intelligent terminal, so as to solve the problem in the prior art that the fault cause of the intelligent terminal being offline cannot be determined quickly and accurately.

[0004] According to one aspect of the present invention, a method for classifying offline fault causes of a power IoT intelligent terminal is provided, the method comprising:

[0005] Obtain historical time series data of power IoT smart terminals;

[0006] Inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that is a combination of a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention;

[0007] The fault cause of the offline of the power Internet of Things intelligent terminal is determined according to the fault cause probability set.

[0008] According to another aspect of the present invention, a device for classifying offline fault causes of a power IoT intelligent terminal is provided, the device comprising:

[0009] An acquisition module is used to acquire historical time series data of the power Internet of Things smart terminal;

[0010] An analysis module, used for inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that is a combination of a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention;

[0011] A determination module is used to determine the fault cause of the power Internet of Things intelligent terminal being offline according to the fault cause probability set.

[0012] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising: at least one processor; and

[0013] a memory communicatively connected to the at least one processor; wherein,

[0014] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the offline fault cause classification method of the power Internet of Things smart terminal described in any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the offline fault cause classification method of the power Internet of Things smart terminal described in any embodiment of the present invention when executed.

[0016] A method for classifying offline fault causes of an electric IoT smart terminal according to an embodiment of the present invention comprises: obtaining historical time series data of the electric IoT smart terminal; inputting the historical time series data into a fault cause analysis model to obtain a fault cause probability set, wherein the fault cause analysis model comprises a GRU-DNN-Attention model combined with a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; and determining the fault cause of the electric IoT smart terminal being offline according to the fault cause probability set. The method can determine the fault cause probability set of the electric IoT smart terminal being offline through the GRU-DNN-Attention model, and determine the fault cause of the electric IoT smart terminal being offline according to the fault cause probability set, thereby solving the problem that the fault cause of the smart terminal being offline cannot be determined quickly and accurately in the prior art, and improving the operation and maintenance efficiency of the electric IoT smart terminal. Through the fault cause analysis model, the operation and maintenance personnel can be assisted to quickly and accurately locate the fault of the electric IoT smart terminal, thereby improving the terminal operation and maintenance efficiency, changing the operation and maintenance mode relying on manual on-site debugging, effectively saving the debugging and operation and maintenance costs, and improving the digital level of power grid operation.

[0017] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A flowchart of a method for classifying offline fault causes of an electric power IoT intelligent terminal provided in the first embodiment of the present invention;

[0020] Figure 2 A schematic diagram of the network structure of a GRU-DNN-Attention model proposed in an embodiment of the present invention;

[0021] Figure 3 A schematic diagram of the classification of offline fault causes of a power IoT smart terminal provided by an embodiment of the present invention;

[0022] Figure 4 A weight change schematic diagram provided by an embodiment of the present invention;

[0023] Figure 5 A schematic diagram of the structure of an offline fault cause classification device for an electric power IoT intelligent terminal provided in Embodiment 2 of the present invention;

[0024] Figure 6 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should belong to the scope of protection of the present invention. It should be understood that the various steps recorded in the method implementation of the present invention can be performed in different orders and / or in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0026] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0029] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0030] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0031] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0032] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0033] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0034] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0035] Embodiment 1

[0036] Figure 1 A flow chart of a method for classifying offline fault causes of a power Internet of Things smart terminal is provided in accordance with the first embodiment of the present invention. The method can be applied to the situation where the offline causes of the power Internet of Things smart terminal are analyzed. The method can be executed by an offline fault cause classification device of the power Internet of Things smart terminal, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device. In this embodiment, the electronic device includes but is not limited to: computers and other devices.

[0037] like Figure 1 As shown, a method for classifying offline fault causes of a power IoT smart terminal provided by a first embodiment of the present invention includes the following steps:

[0038] S110. Obtain historical time series data of the power Internet of Things smart terminal.

[0039] Among them, the power Internet of Things smart terminal refers to the equipment that is mainly used on the edge side to realize data collection, edge computing and data transmission through the network layer of electrical quantities, environmental quantities, state quantities, video images, etc., and undertakes multiple functions such as data collection, processing, encryption, and transmission. Historical time series data can be the time series data before the power Internet of Things smart terminal. Time series data refers to the state change information of an object in the historical time dimension.

[0040] In this embodiment, during the operation of the power Internet of Things smart terminal, a large amount of equipment status information will be collected. This embodiment can obtain this information of the power Internet of Things smart terminal, that is, historical time series data. Based on this information, the cause of the offline failure of the power Internet of Things smart terminal can be analyzed, and relevant professional operation and maintenance personnel can be promptly assigned to carry out maintenance, thereby improving operation and maintenance efficiency.

[0041] S120: Input the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities.

[0042] The fault cause analysis model includes a GRU-DNN-Attention model that combines a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention.

[0043] Among them, the fault cause analysis model can be a model for analyzing the offline cause of the power Internet of Things intelligent terminal. The fault cause probability set can be a set of probabilities of the fault causes of the power Internet of Things intelligent terminal being offline, and the fault cause probability set can include the probability of each fault cause occurring. The gated recurrent unit (GRU) is a type of recurrent neural network. Deep Neural Networks (DNN) are the basis of deep learning. The attention mechanism (Attention) is a resource allocation scheme that is the main means to solve the problem of information overload, allocating computing resources to more important tasks.

[0044] In this embodiment, a GRU-DNN-Attention model can be constructed to classify the reasons for offline faults of terminal devices. Specifically, historical time series data can be input into the GRU-DNN-Attention model to obtain a set of probability of fault reasons. GRU can retain the features of historical time series data, and the Attention layer can enable the algorithm to notice abnormalities in historical data to improve the accuracy of classification of reasons for offline terminals. For example, Figure 2 A schematic diagram of the network structure of a GRU-DNN-Attention model proposed in an embodiment of the present invention. By introducing the attention mechanism, attention can be paid to key data.

[0045] Specifically, when constructing the GRU-DNN-Attention model, a linear mapping can be used first. Given the hidden state of the input sequence represented as x, the attention mechanism first calculates the attention weight through the following linear mapping:

[0046] s=W att ·x+b att

[0047] Among them, W att is the learning parameter of the attention weight, b att is the bias term.

[0048] Then, the attention score s is normalized by softmax to obtain the attention weight a i :

[0049]

[0050] Where T is the number of time steps in the sequence, a i represents the attention weight at the i-th time step.

[0051] Then, use the attention weight a i For the hidden state x i Perform weighted aggregation to obtain the weighted result c:

[0052]

[0053] Among them, x i is the hidden state at the ith time step.

[0054] Finally, the weighted result c can be passed to the neural network DNN for subsequent processing and classification.

[0055] The attention mechanism allows the model to dynamically assign weights based on the importance of each time step, thereby focusing on the parts of the sequence data that contain key information. This helps improve the performance and representation of the model when processing sequence data.

[0056] The GRU-DNN-Attention network constructed in this embodiment can better mine the information in the historical time series data and classify the reasons why the smart terminal devices go offline.

[0057] S130. Determine the fault cause of the power Internet of Things smart terminal being offline according to the fault cause probability set.

[0058] The fault cause may be a cause of an offline fault of the power IoT smart terminal. The types of fault causes may be classified according to actual conditions.

[0059] In this embodiment, when the fault cause probability set is obtained, the fault cause of the power Internet of Things smart terminal being offline can be determined according to the probability corresponding to each fault cause in the fault cause probability set. For example, the fault cause with the highest probability can be used as the fault cause. When there are two highest fault cause probabilities, the final fault cause of the power Internet of Things smart terminal being offline can also be determined by continuing to compare, and this embodiment does not limit this.

[0060] In one embodiment, the historical timing data includes at least: network delay data, ambient temperature, device temperature, CPU usage and memory usage; the fault causes include at least: network failure, device hardware failure, device software failure and power failure.

[0061] In this embodiment, the historical time series data may include network delay data, ambient temperature, device temperature, CPU usage, and memory usage, and the fault cause may include network failure, device hardware failure, device software failure, and power failure. It is understandable that the historical time series data and the fault cause are not limited to the above data.

[0062] A method for classifying offline fault causes of an electric IoT smart terminal is provided in a first embodiment of the present invention, the method comprising: obtaining historical time series data of the electric IoT smart terminal; inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, the fault cause analysis model comprising a GRU-DNN-Attention model combining a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; and determining the fault cause of the electric IoT smart terminal being offline according to the set of fault cause probabilities. The method can determine the set of fault cause probabilities of the electric IoT smart terminal being offline through the GRU-DNN-Attention model, and determine the fault cause of the electric IoT smart terminal being offline according to the set of fault cause probabilities, thereby solving the problem in the prior art that the fault cause of the smart terminal being offline cannot be determined quickly and accurately, and improving the operation and maintenance efficiency of the electric IoT smart terminal.

[0063] Based on the above embodiment, a variant embodiment of the above embodiment is proposed. It should be noted that in order to make the description concise, only the differences from the above embodiment are described in the variant embodiment.

[0064] In one embodiment, the fault cause analysis model further includes at least one of the following: decision tree, random forest, naive Bayes, support vector machine and logistic regression.

[0065] Among them, the decision tree is a prediction model that represents a mapping relationship between object attributes and object values. Random forest is an algorithm that integrates multiple trees through the bagging idea of ​​ensemble learning. Naive Bayes is a type of Bayesian classification algorithm. Support vector machine is a binary classification model. Logistic regression is a machine learning method used to solve binary classification problems.

[0066] In this embodiment, the fault cause analysis model also includes at least one of the following: decision tree, random forest, naive Bayes, support vector machine and logistic regression. Since the neural network has a large demand for historical data, and in the early stage of the construction of the power smart IoT system, there is still a lack of sufficient data for neural network training, so the performance of the equipment offline fault cause classification method based on the GRU-DNN-Attention network is difficult to meet expectations in the absence of sufficient data support. Therefore, this embodiment can also introduce other models to determine the probability set of fault causes.

[0067] In one embodiment, when the fault cause analysis model includes at least two models, inputting the historical time series data into the fault cause analysis model to obtain a set of fault cause probabilities includes:

[0068] Inputting the historical time series data into a fault cause analysis model to obtain an initial fault cause probability set output by different fault cause analysis models;

[0069] A final failure cause probability set is determined according to the weights corresponding to different failure cause analysis models and the initial failure cause probability set.

[0070] The initial fault cause probability set may be a set of fault cause probabilities output by each different fault cause analysis model.

[0071] In this embodiment, the initial fault cause probability sets may be determined respectively by different fault cause analysis models, and then the final fault cause probability set may be calculated by weighted summation.

[0072] In one embodiment, the weight determination method of the GRU-DNN-Attention model, decision tree, random forest, and naive Bayes includes:

[0073]

[0074] Among them, w G is the GRU-DNN-Attention model, w max,G is the maximum weight of the GRU-DNN-Attention network model, w R is the weight of random forest, w D is the weight of the decision tree, w B is the weight of the naive Bayes method, n is the number of historical time series data, k R is the weight parameter of random forest, k D is the weight parameter of the decision tree, k B is the weight parameter of Naive Bayes, w min,R is the minimum weight of random forest, w min,D is the minimum weight of the decision tree, w min,B is the minimum weight of the naive Bayes method, w 0,R is the initial weight of the random forest, w 0,D is the initial weight of the decision tree, w 0,B is the initial weight of Naive Bayes, and σ is the adjustment parameter.

[0075] In this embodiment, the decision tree, random forest, and naive Bayes algorithms are combined with the GRU-DNN-Attention network model to build an integrated fault cause analysis model. It should be noted that the integrated learning architecture of the present invention can integrate other types of algorithms, such as support vector machines, logistic regression, etc., according to actual conditions. For example, Figure 3 A schematic diagram of offline fault cause classification of a power IoT smart terminal provided by an embodiment of the present invention. After determining the weight of each model by the above weight determination method, the historical time series data can be input into different fault cause analysis models to obtain the fault cause probability set output by each model. The probability of each fault cause is then calculated by the following formula:

[0076]

[0077] Where M is the number of fault cause analysis models participating in ensemble learning, P i is the probability of occurrence of the fault cause output by the i-th fault cause analysis model, w i is the weight of the i-th fault cause analysis model.

[0078] The weights in this embodiment can be set dynamically. When the amount of data is sufficient, the GRU-DNN-Attention network model has a higher classification ability for the reasons why smart devices are offline than other methods. Therefore, the weights can be set dynamically according to the amount of data.

[0079] The weight determination formula in this embodiment introduces the sigmoid function to limit the weight value between 0 and 1, and introduces the concept of maximum and minimum weights so that the range of weight variation can be controlled. Figure 4 A weight change schematic diagram provided by an embodiment of the present invention, such as Figure 4 As shown in the figure, under the above weight formula, the weight value of each method changes with the training data. It can be seen from the figure that the weight value of each method can be well limited between the maximum and minimum weights. At the same time, the change speed of different weights can also be controlled by changing the amount of data. In general, with the increase of historical time series data for training, the weight value of the GRU-DNN-Attention network model is gradually increasing, while the weight values ​​of other methods are gradually decreasing.

[0080] In order to solve the problem that neural networks perform poorly in small data sets, this embodiment can propose a generalized ensemble learning algorithm based on dynamic weight adaptation based on the idea of ​​ensemble learning. According to the changes in the number of data sets, the classification weights of different algorithm results are gradually changed, which not only avoids the impact of the degradation of neural network results in small data sets, but also gives full play to the accuracy advantage of neural networks in large data sets.

[0081] In one embodiment, the method further comprises:

[0082] Obtain historical offline status data of different power Internet of Things smart terminals; determine the offline association relationship of different power Internet of Things smart terminals according to the historical offline status data; and update the power supply failure probability in the fault cause probability set according to the association relationship.

[0083] The historical offline status data may be the offline status data of the power IoT smart terminal in the historical time, and whether the power IoT smart terminal is online can be determined through the historical offline status data. The association relationship may be the offline relationship of different power IoT smart terminals. For example, when device A is offline, if device B is also offline, it can be considered that device A and device B have an association relationship.

[0084] In this embodiment, the historical offline status data of different power IoT smart terminals can be obtained, and the association relationship of the offline of different power IoT smart terminals can be determined according to the historical offline status data, and the power failure probability in the fault cause probability set can be updated according to the association relationship. The offline reasons of smart devices such as power failure are difficult to classify through historical time series data, but it can be relatively convenient to make artificial knowledge judgments through the online / offline status of other adjacent devices. Therefore, this embodiment can also integrate knowledge into the data-driven method, based on the method of generalized integrated learning data-knowledge coupling classification.

[0085] This embodiment can introduce the Pearson correlation coefficient to assist in the extraction of artificial knowledge, and then use the artificial knowledge to correct the power failure probability in the fault cause probability set to further improve the algorithm performance.

[0086] In one embodiment, the method for determining the association relationship includes:

[0087]

[0088] Among them, ij is the offline association relationship between power IoT smart terminals i and j, x k is the offline state (0 or 1) of the power IoT smart terminal i at the kth time point, y k is the offline state of the power IoT smart terminal j at the kth time point, and is the mean offline status of the power IoT smart terminals i and j. After calculating the correlation coefficient, if the absolute value of the correlation coefficient is large (close to 1), it means that the offline conditions of the two devices are highly correlated and may be powered by the same power supply. Otherwise, it means that their offline conditions are weakly correlated and may be powered by different power supplies.

[0089] Furthermore, the probability that the offline reason of each offline device is power failure can be calculated through the association relationship:

[0090]

[0091] Among them, P i (E|O) is the probability that the offline reason of the i-th offline device is power failure, Z in is the online / offline status of the nth device when the i-th device is offline, υ in is the association relationship between the ith device and the nth device. If the power supply line connection status of each terminal is known, it is easy to obtain υ in If the power supply line connection status of each terminal device is unknown, it can be determined by fitting historical data. For example, if two devices are always offline at the same time, the weight of their association relationship is large.

[0092] Furthermore, the knowledge-data fusion power failure probability update formula can be constructed:

[0093]

[0094] Among them, P E is the probability of the device being offline due to power failure, and γ is the weight coefficient, which is used to balance the weight values ​​of data-driven results and knowledge-driven results.

[0095] Embodiment 2

[0096] Figure 5 This is a structural schematic diagram of an offline fault cause classification device for an electric power Internet of Things smart terminal provided in Embodiment 2 of the present invention. The device can be used to analyze the offline causes of electric power Internet of Things smart terminals, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device.

[0097] like Figure 5 As shown, the device comprises:

[0098] An acquisition module 210 is used to acquire historical time series data of a power IoT smart terminal;

[0099] An analysis module 220 is used to input the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that combines a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention;

[0100] The determination module 230 is used to determine the fault cause of the offline of the power Internet of Things intelligent terminal according to the fault cause probability set.

[0101] The second embodiment of the present invention provides an offline fault cause classification device for an electric IoT smart terminal, including an acquisition module for acquiring historical time series data of the electric IoT smart terminal; an analysis module for inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that combines a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; and a determination module for determining the fault cause of the electric IoT smart terminal being offline according to the set of fault cause probabilities. The set of fault cause probabilities of the electric IoT smart terminal being offline is determined by the GRU-DNN-Attention model, and the fault cause of the electric IoT smart terminal being offline is determined according to the set of fault cause probabilities, thereby solving the problem in the prior art that the fault cause of the smart terminal being offline cannot be determined quickly and accurately, and improving the operation and maintenance efficiency of the electric IoT smart terminal.

[0102] Furthermore, the fault cause analysis model also includes at least one of the following: decision tree, random forest, naive Bayes, support vector machine and logistic regression.

[0103] Further, when the fault cause analysis model includes at least two models, the analysis module 220 includes:

[0104] Inputting the historical time series data into a fault cause analysis model to obtain an initial fault cause probability set output by different fault cause analysis models;

[0105] A final failure cause probability set is determined according to the weights corresponding to different failure cause analysis models and the initial failure cause probability set.

[0106] Furthermore, the weight determination methods of the GRU-DNN-Attention model, decision tree, random forest, and naive Bayes include:

[0107]

[0108] Among them, w G is the GRU-DNN-Attention model, w max,G is the maximum weight of the GRU-DNN-Attention network model, w R is the weight of random forest, w D is the weight of the decision tree, w B is the weight of the naive Bayes method, n is the number of historical time series data, k R is the weight parameter of random forest, k D is the weight parameter of the decision tree, k B is the weight parameter of Naive Bayes, w min,Ris the minimum weight of random forest, w min,D is the minimum weight of the decision tree, w min,B is the minimum weight of the naive Bayes method, w 0,R is the initial weight of the random forest, w 0,D is the initial weight of the decision tree, w 0,B is the initial weight of Naive Bayes, and σ is the adjustment parameter.

[0109] Furthermore, the historical time series data includes at least: network delay data, ambient temperature, device temperature, CPU usage and memory usage; the fault causes include at least: network failure, device hardware failure, device software failure and power failure.

[0110] Furthermore, the device also includes:

[0111] Obtain historical offline status data of different power IoT smart terminals;

[0112] Determine the offline association relationship of different power IoT smart terminals according to the historical offline status data;

[0113] The power supply failure probability in the failure cause probability set is updated according to the association relationship.

[0114] Furthermore, the method for determining the association relationship includes:

[0115]

[0116] Among them, ij is the offline association relationship between power IoT smart terminals i and j, x k is the offline state of the power IoT smart terminal i at the kth time point, y k is the offline state of the power IoT smart terminal j at the kth time point, and It is the mean value of the offline state corresponding to the power IoT smart terminals i and j.

[0117] The above-mentioned offline fault cause classification device of the power Internet of Things intelligent terminal can execute the offline fault cause classification method of the power Internet of Things intelligent terminal provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0118] Embodiment 3

[0119] Figure 6A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0120] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0121] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0122] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, such as the offline fault cause classification method of the power Internet of Things smart terminal.

[0123] In some embodiments, the offline fault cause classification method of the power Internet of Things intelligent terminal can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the offline fault cause classification method of the power Internet of Things intelligent terminal described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the offline fault cause classification method of the power Internet of Things intelligent terminal in any other appropriate manner (for example, by means of firmware).

[0124] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0125] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0126] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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 foregoing.

[0127] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0128] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0129] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0130] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0131] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for classifying offline fault causes of an electric power IoT intelligent terminal, characterized in that: The method comprises: Obtain historical time series data of the power Internet of Things intelligent terminal; the historical time series data at least includes: network delay data, ambient temperature, device temperature, CPU usage rate and memory usage rate; Inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that is a combination of a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; Determine the fault cause of the offline of the power IoT intelligent terminal according to the fault cause probability set; The fault cause analysis model also includes at least one of the following: decision tree, random forest, naive Bayes, support vector machine and logistic regression; The weight determination methods for the GRU-DNN-Attention model, decision tree, random forest, and naive Bayes include: in, is the weight of the GRU-DNN-Attention model, is the maximum weight of the GRU-DNN-Attention network model, is the weight of random forest, is the weight of the decision tree, is the weight of the naive Bayes method, is the number of historical time series data, is the weight parameter of random forest, is the weight parameter of the decision tree, is the weight parameter of Naive Bayes, is the minimum weight of random forest, is the minimum weight of the decision tree, is the minimum weight of the naive Bayes method, is the initial weight of random forest, is the initial weight of the decision tree, is the initial weight of Naive Bayes, To adjust the parameters; When the fault cause analysis model includes at least two models, inputting the historical time series data into the fault cause analysis model to obtain a set of fault cause probabilities includes: Inputting the historical time series data into a fault cause analysis model to obtain an initial fault cause probability set output by different fault cause analysis models; A final failure cause probability set is determined according to the weights corresponding to different failure cause analysis models and the initial failure cause probability set.

2. The method according to claim 1, characterized in that The failure causes include at least: network failure, device hardware failure, device software failure and power failure.

3. The method according to claim 1, characterized in that The method further comprises: Obtain historical offline status data of different power IoT smart terminals; Determine the offline association relationship of different power IoT smart terminals according to the historical offline status data; The power supply failure probability in the failure cause probability set is updated according to the association relationship.

4. The method according to claim 3, characterized in that: The method for determining the association relationship includes: in, is the offline association relationship between power IoT smart terminals i and j, x k is the offline state of the power IoT smart terminal i at the kth time point, y k is the offline state of the power IoT smart terminal j at the kth time point, and It is the mean value of the offline state corresponding to the power IoT smart terminals i and j.

5. An offline fault cause classification device for an electric power IoT intelligent terminal, characterized in that: The device comprises: An acquisition module is used to acquire historical time series data of the power Internet of Things intelligent terminal; the historical time series data at least includes: network delay data, ambient temperature, device temperature, CPU usage rate and memory usage rate; An analysis module, used for inputting the historical time series data into a fault cause analysis model to obtain a set of fault cause probabilities, wherein the fault cause analysis model includes a GRU-DNN-Attention model that is a combination of a gated recurrent unit GRU, a deep neural network DNN, and an attention mechanism Attention; A determination module, used to determine the fault cause of the offline of the power Internet of Things intelligent terminal according to the fault cause probability set; The fault cause analysis model also includes at least one of the following: decision tree, random forest, naive Bayes, support vector machine and logistic regression; The weight determination methods for the GRU-DNN-Attention model, decision tree, random forest, and naive Bayes include: in, is the weight of the GRU-DNN-Attention model, is the maximum weight of the GRU-DNN-Attention network model, is the weight of random forest, is the weight of the decision tree, is the weight of the naive Bayes method, is the number of historical time series data, is the weight parameter of random forest, is the weight parameter of the decision tree, is the weight parameter of Naive Bayes, is the minimum weight of random forest, is the minimum weight of the decision tree, is the minimum weight of the naive Bayes method, is the initial weight of random forest, is the initial weight of the decision tree, is the initial weight of Naive Bayes, To adjust the parameters; When the fault cause analysis model includes at least two models, the analysis module includes: Inputting the historical time series data into a fault cause analysis model to obtain an initial fault cause probability set output by different fault cause analysis models; A final failure cause probability set is determined according to the weights corresponding to different failure cause analysis models and the initial failure cause probability set.

6. An electronic device, characterized in that: The device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the offline fault cause classification method of the power Internet of Things intelligent terminal described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the offline fault cause classification method for the power Internet of Things smart terminal according to any one of claims 1 to 4 when executed.

Images