CN118228200A - Multi-mode model-based power equipment abnormality identification method, device and equipment - Google Patents

Abstract

The embodiment of the invention discloses a method, a device and equipment for identifying power equipment abnormality based on a multi-mode model. One embodiment of the method comprises the following steps: for each historical device information set, the following steps are performed: model training is carried out on each initial power equipment anomaly identification network according to the historical equipment information group so as to generate each network weight information and each network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality identification network according to the target network weight information and the target network structure information; and inputting the target power equipment information into a multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result. The embodiment can accurately, efficiently and quickly detect the abnormality of the operation information of the power equipment so as to determine whether the power equipment is abnormal.

Description

Multi-mode model-based power equipment abnormality identification method, device and equipment

Technical Field

The embodiment of the disclosure relates to the technical field of computers, in particular to a method, a device and equipment for identifying power equipment abnormality based on a multi-mode model.

Background

In power networks, a variety of power devices are involved, and the distribution is very broad. Therefore, in order to ensure the normal operation of the power network, it is necessary to detect the power equipment to find abnormal power equipment in advance. At present, the abnormality detection of the power equipment is generally performed by the following modes: the electrical equipment is periodically checked by a technician.

However, when the abnormality detection is performed on the power equipment in the above manner, there are often the following technical problems: the time interval exists in manual detection, and timeliness is lower, and abnormal power equipment is difficult to detect in time.

The above information disclosed in this background section is only for enhancement of understanding of the background of the inventive concept and, therefore, may contain information that does not form the prior art that is already known to those of ordinary skill in the art in this country.

Disclosure of Invention

The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Some embodiments of the present disclosure propose a method, apparatus, electronic device, and computer-readable medium for identifying anomalies in power devices based on multimodal models to solve one or more of the technical problems mentioned in the background section above.

In a first aspect, some embodiments of the present disclosure provide a method for identifying an abnormality of a power device based on a multi-modal model, the method comprising: acquiring historical equipment information groups of each piece of electric equipment to obtain a historical equipment information group set, wherein each historical equipment information group corresponds to an information tag, and the historical equipment information comprises electric equipment operation data and electric equipment appearance images; for each of the above-described historical device information groups, the following steps are performed: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group; fusing each multi-source power equipment abnormality recognition network into a multi-mode power equipment abnormality recognition model; and inputting the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

In a second aspect, some embodiments of the present disclosure provide a multi-modal model-based power equipment anomaly identification device, the device comprising: the system comprises an acquisition unit, a display unit and a display unit, wherein the acquisition unit is configured to acquire historical equipment information groups of each power equipment to obtain a historical equipment information group set, each historical equipment information group corresponds to an information tag, and the historical equipment information comprises power equipment operation data and power equipment appearance images; a construction unit configured to perform, for each of the history device information groups in the history device information group set described above, the steps of: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group; a fusion unit configured to fuse each of the multi-source power device anomaly identification networks into a multi-mode power device anomaly identification model; and the input unit is configured to input the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors causes the one or more processors to implement the method described in any of the implementations of the first aspect above.

In a fourth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

The above embodiments of the present disclosure have the following advantageous effects: by the method for identifying the abnormality of the power equipment based on the multi-mode model, which is disclosed by the embodiment of the invention, the operation information of the power equipment can be accurately, efficiently and rapidly detected to determine whether the power equipment is abnormal or not. In particular, the inaccuracy of the manually delimited transportation device area is caused by the fact that the manually delimited transportation device area is a fixed transportation device area, and during operation of the transportation device, situations may occur in which different electrical devices need to be stored in different areas. Based on this, in the method for identifying abnormal power equipment based on the multi-mode model according to some embodiments of the present disclosure, first, a historical equipment information set of each power equipment is obtained, and a historical equipment information set is obtained, where each historical equipment information set corresponds to an information tag, and the historical equipment information includes power equipment operation data and a power equipment appearance image. Thus, the power equipment abnormality recognition model can be trained by using the historical operation information of the power equipment. Next, for each of the above-described history device information sets, the following steps are performed: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with target information labels, wherein the target information labels are information labels corresponding to the historical equipment information groups. Therefore, the generalization capability and the prediction accuracy of each model structure for the current power equipment type can be determined by training the models of different network structures by using the historical equipment information sets collected from a plurality of power equipment. Then, the multi-source power equipment abnormality recognition networks are fused into a multi-mode power equipment abnormality recognition model. Thus, the trained multimodal power apparatus anomaly identification model is enabled to identify a variety of different power apparatuses. And finally, inputting the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result. Therefore, the abnormality detection can be accurately, efficiently and quickly carried out on the operation information of the power equipment so as to determine whether the power equipment is abnormal.

Drawings

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.

FIG. 1 is a flow chart of some embodiments of a multi-modal model-based power device anomaly identification method according to the present disclosure;

FIG. 2 is a flow chart of some embodiments of a multi-modal model-based power device anomaly identification apparatus according to the present disclosure;

Fig. 3 is a schematic structural diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1 is a flow chart of some embodiments of a multi-modal model-based power device anomaly identification method according to the present disclosure. A flow 100 of some embodiments of a multi-modal model-based power device anomaly identification method in accordance with the present disclosure is shown. The power equipment abnormality identification method based on the multi-mode model comprises the following steps:

step 101, acquiring a historical equipment information set of each piece of power equipment, and obtaining a historical equipment information set.

In some embodiments, an execution body (for example, a computing device) of the power device abnormality recognition method based on the multi-mode model may obtain a historical device information set of each power device through a wired connection or a wireless connection, so as to obtain a historical device information set. Each historical equipment information group corresponds to an information tag, and the historical equipment information comprises power equipment operation data and power equipment appearance images. Wherein, the historical virtual power plant data of the historical equipment information group comprises: the method comprises the steps of power equipment operation data, power equipment appearance images, operation data labels and equipment appearance image labels. The power device may be various power devices, for example, a power generation device, a power transmission and transformation device, a power consumer, a power electronic device, and the like. The information tag may be an information text tag that characterizes a source of each historical device information in the corresponding set of historical device information. For example, a data source tag corresponding to a historical equipment information set is "power plant," and may characterize that the historical equipment information set originates from the power plant. For another example, a data source tag corresponding to a historical equipment information set is "power transmission and transformation equipment", and may represent that the historical equipment information set originates from the power transmission and transformation equipment. For example, the power device operational data may include: power plant load values, power plant maximum operating voltage, maximum operating current, etc. The operational data tag may indicate whether the power device operational data is abnormal.

In practice, the execution subject may acquire the historical device information group of each power device by:

First, acquiring an initial equipment information group of the power equipment in a preset time period. The preset time period may refer to a history time period. One initial device information corresponds to one time granularity.

And secondly, detecting abnormal information of each piece of initial equipment information included in the initial equipment information group so as to generate an abnormal information tag group. The abnormality information detection includes: null value detection, abnormal discrete data detection, and the like. Then, an information tag corresponding to the initial device information in which the null value/abnormal discrete data exists may be determined as an abnormal information tag.

And thirdly, performing data restoration processing on each piece of abnormal initial equipment information in the initial equipment information group to generate a restoration initial equipment information group. Wherein the abnormality initial device information in the abnormality initial device information corresponds to an abnormality information tag in the abnormality information tag group. For example, the data repair process may refer to operations such as filling in null/repairing outliers.

And step four, resampling the pieces of repair initial equipment information included in the repair initial equipment information set to obtain a sampling repair initial equipment information set. Resampling is mainly carried out by nearest neighbor method, bilinear interpolation method and three-time convolution interpolation method.

And fifthly, performing noise reduction processing on each piece of sampling repair initial equipment information included in the sampling repair initial equipment information group to obtain a noise reduction initial equipment information group serving as a historical equipment information group. The method of noise reduction processing may include: data cleaning, filtering and integrating methods.

Step 102, for each historical device information group in the above historical device information group set, executing the following steps:

And 1021, performing model training on each initial power equipment abnormality recognition network according to the historical equipment information set so as to generate each network weight information and each network structure information.

In some embodiments, the executing entity may perform model training on each initial power device anomaly identification network according to the historical device information set to generate each network weight information and each network structure information. Wherein, the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to one training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information. The initial power device anomaly identification network may refer to an untrained neural network model. For example, the initial power equipment anomaly identification Network may be a DCN (Deep & Cross Network) model or a transformer model, or may be a YoLoV model.

In practice, the execution subject may perform model training on each initial power equipment anomaly identification network by:

And combining the power equipment operation data and the power equipment appearance image included in each piece of history equipment information in the history equipment information group into equipment information sample data to obtain an equipment information sample data group.

Second, for each of the above-described individual initial power equipment anomaly identification networks, the following training steps are performed:

And a first training step of inputting at least one piece of equipment information sample data in the equipment information sample data set into an initial power equipment abnormality recognition network to obtain an initial power equipment abnormality recognition result corresponding to each piece of equipment information sample data in the at least one piece of equipment information sample data. The initial power equipment abnormality identification result comprises: and the power equipment operation data identification result and the power equipment appearance image identification result.

And a second training step, namely determining a model loss value according to the operation data label and the equipment appearance image label corresponding to the at least one initial power equipment abnormality recognition result and the at least one equipment information sample data. That is, a data loss value between each power equipment operation data identification result and a corresponding operation data tag may be determined through a preset loss function, and an image loss value between each power equipment appearance image identification result and a corresponding equipment appearance image tag may be determined. And determining the sum of the data loss value and the image loss value corresponding to each piece of equipment information sample data as a model loss value.

And a third training step, determining whether the initial power equipment abnormality identification network reaches a preset optimization target according to the model loss value. That is, it is determined whether the model loss value is equal to or less than a preset threshold. If the model loss value is smaller than or equal to a preset threshold value, the initial power equipment abnormality identification network reaches a preset optimization target.

A fourth training step, in response to determining that the initial power equipment abnormality recognition network reaches the preset optimization target, of executing the following adjustment steps:

first, an initial power device anomaly identification network is determined as a power device anomaly identification network.

And secondly, performing adjustment processing on the power equipment abnormality recognition network to generate network weight information and network structure information.

In some optional implementations of some embodiments, the executing entity may perform adjustment processing on the electrical device anomaly identification network by:

And step one, carrying out layer-by-layer weight export processing on the power equipment abnormality identification network so as to generate an initial network layer weight information sequence. Wherein each initial network layer weight information in the initial network layer weight information sequence comprises a respective neuron weight, and the respective neuron weight corresponds to the neuron position information.

Second, for each initial network layer weight information in the initial network layer weight information sequence, the following processing steps are executed:

first, an initial weight adjustment value is generated according to each neuron weight included in the initial network layer weight information. For example, the average value of the individual neuron weights may be determined as the initial weight adjustment value

Second, a weight adjustment value is generated based on the initial weight adjustment value. The initial weight adjustment value may be determined as the weight adjustment value.

Third, based on the initial weight adjustment value, the following structure adjustment steps are performed:

1. And selecting the neuron weight meeting the preset weight condition from the initial network layer weight information as a target neuron weight to obtain a target neuron weight group. The preset weight condition may mean that the neuron weight is equal to or greater than the initial weight adjustment value.

2. And deleting each model neuron corresponding to the target neuron weight group in the power equipment abnormality identification network to obtain an updated power equipment abnormality identification network.

3. And verifying the updated power equipment abnormality identification network to generate a verification loss value. For example, the updated power device anomaly identification network may be validated through a validation dataset to generate a validation loss value.

And thirdly, in response to determining that the verification loss value is larger than the model loss value corresponding to the power equipment abnormality identification network and the execution times of the structure adjustment step are larger than the preset execution times, performing weight derivation processing on the updated power equipment abnormality identification network to generate network weight information.

Fourth, according to the updated power equipment abnormality identification network, generating network structure information. That is, the configuration information of each layer of the network layer of the updated power device abnormality recognition network may be determined as the network configuration information.

Therefore, by iteratively carrying out structural adjustment and verification on the model, the adjusted model is ensured to still maintain or improve the prediction precision, the complexity and the calculation requirement of the model are reduced, the calculation force requirement in the model task process is further effectively reduced, and the original prediction performance of the model is maintained.

Step 1022, generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information.

In some embodiments, the executing entity may generate the target network weight information and the target network structure information according to each training result, each network weight information, and each network structure information.

In practice, the execution subject may generate the target network weight information and the target network structure information by:

First, selecting a training result satisfying the information condition corresponding to the target information label from the training results as a target training result. The information condition may be: the model loss value corresponding to the training result is minimum.

And secondly, determining the network weight information corresponding to the target training result as target network weight information.

And thirdly, determining the network structure information corresponding to the target training result as target network structure information.

Step 1023, constructing a multisource power equipment abnormality recognition network according to the target network weight information and the target network structure information, and binding the multisource power equipment abnormality recognition network with the target information tag.

In some embodiments, the executing body may construct a multisource power equipment abnormality recognition network according to the target network weight information and the target network structure information, and bind the multisource power equipment abnormality recognition network and the target information tag. The target information tag is an information tag corresponding to the historical equipment information group.

In practice, the execution subject may construct a multi-source power device anomaly identification network by the following steps, and bind the multi-source power device anomaly identification network and a target information tag:

first, determining an initial power equipment abnormality recognition network corresponding to the target training result.

And secondly, carrying out structural adjustment on the initial power equipment abnormality identification network according to the target network structural information so as to update the initial power equipment abnormality identification network. That is, the network configuration information of the initial power equipment abnormality recognition network may be adjusted to the target network configuration information.

Thirdly, weighting the updated initial power equipment abnormality identification network according to the target network weight information to construct a multi-source power equipment abnormality identification network. That is, each weight in the updated initial power device abnormality identification network may be replaced with a weight included in the target network weight information.

And fourthly, binding the multisource power equipment abnormality recognition network with the target information tag.

And step 103, fusing the multi-source power equipment abnormality recognition networks into a multi-mode power equipment abnormality recognition model.

In some embodiments, the executing entity may fuse each multi-source power device anomaly identification network into a multi-modal power device anomaly identification model.

And 104, inputting the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

In some embodiments, the executing body may input the target power device information into the multimodal power device abnormality identification model to obtain a target power device abnormality identification result. The multi-modal power device anomaly identification model may include a power device anomaly identification model for identifying various types of power devices. For example, the power equipment abnormality recognition model may be a neural network model that is trained in advance with power equipment information as input and power equipment abnormality recognition results as output. For example, the power device anomaly identification model may be a YoLoV model. The target power device information may be device information of a certain power device currently acquired, and may include: the power equipment operation data and the power equipment appearance image. That is, the target power equipment information may be input into the power equipment abnormality recognition model corresponding to the multi-mode power equipment abnormality recognition model according to the information tag corresponding to the target power equipment information, so as to obtain the target power equipment abnormality recognition result. The target power device abnormality recognition result may indicate whether there is an abnormality in the target power device information.

Optionally, in response to determining that the target power device abnormality identification result characterizes the power device abnormality, the target power device abnormality identification result is sent to an associated power device abnormality processing terminal.

In some embodiments, the execution body may send the target power device anomaly identification result to an associated power device anomaly processing terminal in response to determining that the target power device anomaly identification result characterizes a power device anomaly. The power device abnormality processing terminal may refer to a terminal that performs abnormality processing on an abnormal power device.

With further reference to fig. 2, as an implementation of the method shown in the foregoing figures, the present disclosure provides some embodiments of a multi-modal model-based power device anomaly identification apparatus, which correspond to those method embodiments shown in fig. 1, and which are particularly applicable to various electronic devices.

As shown in fig. 2, the multi-modal model-based power device anomaly identification apparatus 200 of some embodiments includes: an acquisition unit 201, a construction unit 202, a fusion unit 203, and an input unit 204. The acquiring unit 201 is configured to acquire a historical equipment information group of each power equipment to obtain a historical equipment information group set, wherein each historical equipment information group corresponds to an information tag, and the historical equipment information comprises power equipment operation data and power equipment appearance images; a construction unit 202 configured to perform, for each of the above-described history device information sets, the steps of: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group; a fusion unit 203 configured to fuse each of the multi-source power device anomaly identification networks into a multi-modal power device anomaly identification model; and an input unit 204 configured to input the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

It will be appreciated that the elements described in the multi-modal model-based power equipment anomaly identification device 200 correspond to the various steps in the method described with reference to FIG. 1. Thus, the operations, features and advantages described above for the method are equally applicable to the multi-modal model-based power equipment anomaly identification device 200 and the units contained therein, and are not described herein.

Referring now to FIG. 3, a schematic diagram of an electronic device (e.g., computing device) 300 suitable for use in implementing some embodiments of the present disclosure is shown. The electronic devices in some embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), car terminals (e.g., car navigation terminals), and the like, as well as stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is merely an example and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 3, the electronic device 300 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 301 that may perform various suitable actions and processes in accordance with a program stored in a Read Only Memory (ROM) 302 or a program loaded from a storage means 308 into a Random Access Memory (RAM) 303. In the RAM303, various programs and data required for the operation of the electronic apparatus 300 are also stored. The processing device 301, the ROM302, and the RAM303 are connected to each other via a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.

In general, the following devices may be connected to the I/O interface 305: input devices 306 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 307 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 308 including, for example, magnetic tape, hard disk, etc.; and communication means 309. The communication means 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. While fig. 3 shows an electronic device 300 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead. Each block shown in fig. 3 may represent one device or a plurality of devices as needed.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via communications device 309, or from storage device 308, or from ROM 302. The above-described functions defined in the methods of some embodiments of the present disclosure are performed when the computer program is executed by the processing means 301.

It should be noted that, the computer readable medium described in some embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, the computer-readable signal medium may comprise a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some embodiments, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol ), and may be interconnected with 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"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring historical equipment information groups of each piece of electric equipment to obtain a historical equipment information group set, wherein each historical equipment information group corresponds to an information tag, and the historical equipment information comprises electric equipment operation data and electric equipment appearance images; for each of the above-described historical device information groups, the following steps are performed: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group; fusing each multi-source power equipment abnormality recognition network into a multi-mode power equipment abnormality recognition model; and inputting the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

Computer program code for carrying out operations for some embodiments of the present disclosure may be written in one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. The described units may also be provided in a processor, for example, described as: a processor comprising: the device comprises an acquisition unit, a construction unit, a fusion unit and an input unit. The names of these units do not limit the unit itself in some cases, and for example, the input unit may also be described as "a unit that inputs the target power device information into the above-described multi-mode power device abnormality recognition model to obtain the target power device abnormality recognition result".

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above technical features, but encompasses other technical features formed by any combination of the above technical features or their equivalents without departing from the spirit of the invention. Such as the above-described features, are mutually substituted with (but not limited to) the features having similar functions disclosed in the embodiments of the present disclosure.

Claims (7)

1. A power equipment abnormality identification method based on a multi-mode model comprises the following steps:

Acquiring historical equipment information groups of each piece of electric equipment to obtain a historical equipment information group set, wherein each historical equipment information group corresponds to an information tag, and the historical equipment information comprises electric equipment operation data and electric equipment appearance images;

For each of the set of historical device information, performing the steps of:

Performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information;

generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information;

Constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group;

Fusing each multi-source power equipment abnormality recognition network into a multi-mode power equipment abnormality recognition model;

and inputting the target power equipment information into the multi-mode power equipment abnormality identification model to obtain a target power equipment abnormality identification result.

2. The method of claim 1, wherein the obtaining a historical device information set for each power device comprises:

acquiring an initial equipment information group of the power equipment in a preset time period;

Detecting abnormal information of each piece of initial equipment information included in the initial equipment information group to generate an abnormal information tag group;

Performing data restoration processing on each piece of abnormal initial equipment information in the initial equipment information group to generate a restoration initial equipment information group, wherein the abnormal initial equipment information in each piece of abnormal initial equipment information corresponds to an abnormal information label in the abnormal information label group;

resampling the information of each piece of repair initial equipment included in the repair initial equipment information group to obtain a sampled repair initial equipment information group;

and carrying out noise reduction processing on each piece of sampling repair initial equipment information included in the sampling repair initial equipment information group to obtain a noise reduction initial equipment information group serving as a historical equipment information group.

3. The method of claim 1, wherein the constructing a multi-source power device anomaly identification network according to the target network weight information and the target network structure information, and binding the multi-source power device anomaly identification network and the target information tag, comprises:

determining an initial power equipment abnormality identification network corresponding to the target training result;

According to the target network structure information, carrying out structure adjustment on the initial power equipment abnormality identification network so as to update the initial power equipment abnormality identification network;

Weighting the updated initial power equipment abnormality identification network according to the target network weight information to construct a multi-source power equipment abnormality identification network;

binding the multisource power equipment abnormality identification network with the target information tag.

4. The method of claim 1, wherein the method further comprises:

And in response to determining that the target power equipment abnormality identification result characterizes the power equipment abnormality, sending the target power equipment abnormality identification result to an associated power equipment abnormality processing terminal.

5. An electrical equipment anomaly identification device based on a multi-modal model, comprising:

The system comprises an acquisition unit, a display unit and a display unit, wherein the acquisition unit is configured to acquire historical equipment information groups of each power equipment to obtain a historical equipment information group set, each historical equipment information group corresponds to an information tag, and the historical equipment information comprises power equipment operation data and power equipment appearance images;

A construction unit configured to perform, for each of the set of historical device information, the steps of: performing model training on each initial power equipment abnormality recognition network according to the historical equipment information group to generate each network weight information and each network structure information, wherein the network structures of the initial power equipment abnormality recognition networks are different, one network weight information corresponds to a training result, one network weight information corresponds to one initial power equipment abnormality recognition network, and one initial power equipment abnormality recognition network corresponds to one network structure information; generating target network weight information and target network structure information according to each training result, each network weight information and each network structure information; constructing a multisource power equipment abnormality recognition network according to target network weight information and target network structure information, and binding the multisource power equipment abnormality recognition network with a target information tag, wherein the target information tag is an information tag corresponding to the historical equipment information group;

A fusion unit configured to fuse each of the multi-source power device anomaly identification networks into a multi-mode power device anomaly identification model;

and the input unit is configured to input the target power equipment information into the multi-mode power equipment abnormality recognition model to obtain a target power equipment abnormality recognition result.

6. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon;

When executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-4.

7. A computer readable medium having stored thereon a computer program, wherein the program when executed by a processor implements the method of any of claims 1-4.