CN116683648B - Intelligent power distribution cabinet and control system thereof - Google Patents
Intelligent power distribution cabinet and control system thereof Download PDFInfo
- Publication number
- CN116683648B CN116683648B CN202310691904.2A CN202310691904A CN116683648B CN 116683648 B CN116683648 B CN 116683648B CN 202310691904 A CN202310691904 A CN 202310691904A CN 116683648 B CN116683648 B CN 116683648B
- Authority
- CN
- China
- Prior art keywords
- training
- encoder
- classification
- feature
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000009826 distribution Methods 0.000 title claims abstract description 84
- 239000013598 vector Substances 0.000 claims abstract description 91
- 239000011159 matrix material Substances 0.000 claims abstract description 83
- 238000010586 diagram Methods 0.000 claims abstract description 37
- 238000012549 training Methods 0.000 claims description 111
- 238000012545 processing Methods 0.000 claims description 27
- 238000001514 detection method Methods 0.000 claims description 17
- 238000011176 pooling Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 2
- 230000003213 activating effect Effects 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 claims description 2
- 230000006870 function Effects 0.000 description 30
- 238000000034 method Methods 0.000 description 22
- 238000013527 convolutional neural network Methods 0.000 description 17
- 238000012544 monitoring process Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 10
- 230000008859 change Effects 0.000 description 8
- 230000002159 abnormal effect Effects 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000005065 mining Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 5
- 238000013145 classification model Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000003709 image segmentation Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000004148 unit process Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/18—Status alarms
- G08B21/185—Electrical failure alarms
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02B—BOARDS, SUBSTATIONS OR SWITCHING ARRANGEMENTS FOR THE SUPPLY OR DISTRIBUTION OF ELECTRIC POWER
- H02B1/00—Frameworks, boards, panels, desks, casings; Details of substations or switching arrangements
- H02B1/24—Circuit arrangements for boards or switchyards
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00001—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by the display of information or by user interaction, e.g. supervisory control and data acquisition systems [SCADA] or graphical user interfaces [GUI]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Image Analysis (AREA)
Abstract
An intelligent power distribution cabinet and a control system thereof are disclosed. Firstly, arranging electric energy load values at a plurality of preset time points into electric energy load time sequence input vectors according to a time dimension, enabling a current waveform diagram of a preset time period to pass through a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder to obtain a multi-mode feature matrix of an electric power system, then enabling the multi-mode feature matrix of the electric power system to pass through a space attention module to obtain a classification feature matrix, enabling the classification feature matrix to pass through a classifier to obtain a classification result used for indicating whether a monitored electric power system is normal or not, and finally, generating a control signal used for indicating whether an alarm signal is generated or not based on the classification result. In this way, the stability and reliability of the power system can be improved.
Description
Technical Field
The application relates to the field of power distribution cabinets, and more particularly, to an intelligent power distribution cabinet and a control system thereof.
Background
A power distribution cabinet is a device for an electrical power system, also known as a low voltage electrical switchboard or a power control box. The system is mainly used for distributing, controlling, protecting and monitoring electric energy loads and current signals in the electric power system so as to ensure the normal operation, safety and stability of the electric power system. In operation of the power system, faults such as overload, short circuit and the like can cause unstable current and equipment damage, and even endanger life safety. Therefore, monitoring and prediction of the power system becomes particularly important.
However, conventional power distribution cabinets generally can only perform simple switch control and short-circuit protection, and cannot monitor whether overload, short-circuit or other faults occur in the power system in real time. The equipment lacks intelligence and cannot timely early warn and control abnormal conditions in the power system, so that the safety and stability of the power system are threatened.
Accordingly, an optimized intelligent power distribution cabinet is desired.
Disclosure of Invention
In view of this, the disclosure provides an intelligent power distribution cabinet and a control system thereof, which can evaluate the running state of electric equipment in real time by monitoring the electric energy load value, discover potential fault hidden trouble and process in time, and improve the stability and reliability of an electric power system.
According to an aspect of the present disclosure, there is provided an intelligent power distribution cabinet, including:
the data acquisition module is used for acquiring electric energy load values of a monitored electric power system at a plurality of preset time points in preset time and a current waveform diagram of the preset time period;
the data parameter time sequence arrangement module is used for arranging the electric energy load values of the plurality of preset time points into electric energy load time sequence input vectors according to the time dimension;
the joint coding module is used for enabling the electric energy load time sequence input vector and the current waveform diagram of the preset time period to pass through a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder so as to obtain a multi-mode characteristic matrix of the electric power system;
The space feature enhancement module is used for enabling the multi-mode feature matrix of the power system to pass through the space attention module to obtain a classification feature matrix;
the power system detection module is used for passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a monitored power system is normal or not; and
and the early warning module is used for generating a control signal based on the classification result, wherein the control signal is used for indicating whether an alarm signal is generated or not.
According to another aspect of the present disclosure, there is provided a control system of an intelligent power distribution cabinet, including:
the control system is used for controlling any one of the intelligent power distribution cabinets.
According to the embodiment of the disclosure, firstly, electric energy load values at a plurality of preset time points are arranged into electric energy load time sequence input vectors according to a time dimension, and a current waveform chart of a preset time period passes through a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder to obtain a multi-mode characteristic matrix of an electric power system, then, the multi-mode characteristic matrix of the electric power system passes through a spatial attention module to obtain a classification characteristic matrix, then, the classification characteristic matrix passes through a classifier to obtain a classification result used for indicating whether a monitored electric power system is normal or not, and finally, a control signal used for indicating whether an alarm signal is generated or not is generated based on the classification result. In this way, the stability and reliability of the power system can be improved.
Other features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments, which proceeds with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate exemplary embodiments, features and aspects of the present disclosure and together with the description, serve to explain the principles of the disclosure.
Fig. 1 shows a block diagram of an intelligent power distribution cabinet according to an embodiment of the present disclosure.
Fig. 2 shows a block diagram of the joint coding module in the intelligent power distribution cabinet, according to an embodiment of the disclosure.
Fig. 3 shows a block diagram of a training module further included in the intelligent power distribution cabinet, according to an embodiment of the present disclosure.
Fig. 4 shows a flow chart of an intelligent power distribution method according to an embodiment of the present disclosure.
Fig. 5 shows an architectural schematic diagram of an intelligent power distribution method according to an embodiment of the present disclosure.
Fig. 6 illustrates an application scenario diagram of an intelligent power distribution cabinet according to an embodiment of the present disclosure.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present application without making any inventive effort, are also within the scope of the present application.
As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Various exemplary embodiments, features and aspects of the disclosure will be described in detail below with reference to the drawings. In the drawings, like reference numbers indicate identical or functionally similar elements. Although various aspects of the embodiments are illustrated in the accompanying drawings, the drawings are not necessarily drawn to scale unless specifically indicated.
The word "exemplary" is used herein to mean "serving as an example, embodiment, or illustration. Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.
In addition, numerous specific details are set forth in the following detailed description in order to provide a better understanding of the present disclosure. It will be understood by those skilled in the art that the present disclosure may be practiced without some of these specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art have not been described in detail in order not to obscure the present disclosure.
As described above, a conventional power distribution cabinet is a device for distributing, protecting, and controlling electric energy in an electric power system. It is typically composed of a housing, power switch, fuse, circuit breaker, contactor, relay, transformer, etc. These elements are connected by cables and wires to form a circuit system to implement distribution, protection and control functions of the power system. Moreover, the conventional power distribution cabinet generally can only perform simple switch control and short-circuit protection, can only perform basic power distribution, protection and control, lacks an intelligent function, cannot monitor whether overload, short-circuit or other faults occur in the power system in real time, generally requires manual maintenance and management, and lacks the capability of automation and remote monitoring. The equipment lacks intelligence and cannot timely early warn and control abnormal conditions in the power system, so that the safety and stability of the power system are threatened. Accordingly, an optimized intelligent power distribution cabinet is desired.
Accordingly, considering that in the intelligent power distribution cabinet, in order to monitor whether overload, short circuit or other faults occur in the power system in real time, in the technical scheme of the application, it is expected that the change trend of the power load and the current signal in the power system is detected based on analysis of the power load and the current signal, so that the overload, the short circuit or the other faults in the power system are judged, and an alarm is sent out through the intelligent power distribution cabinet. However, since the electric energy load value and the current signal have time-sequence variation characteristics in the time dimension, and have time-sequence cooperative association relationship, the electric energy load value and the current signal have influence on fault detection of the electric power system. Therefore, it is difficult to efficiently analyze the trend of the change in the power system parameters, thereby making it difficult to detect the failure of the power system.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. Deep learning and development of a neural network provide new solutions and schemes for mining correlation characteristic information between time sequence change characteristics of the electric energy load values and time sequence implicit characteristics of the current signals.
Fig. 1 shows a block diagram schematic of an intelligent power distribution cabinet according to an embodiment of the present disclosure. As shown in fig. 1, an intelligent power distribution cabinet 100 according to an embodiment of the present application includes: the data acquisition module 110 is configured to acquire electrical energy load values of a monitored electrical power system at a plurality of predetermined time points within a predetermined time period and a current waveform diagram of the predetermined time period; a data parameter time sequence arrangement module 120, configured to arrange the electrical energy load values at the plurality of predetermined time points into an electrical energy load time sequence input vector according to a time dimension; the joint coding module 130 is configured to pass the electrical energy load time sequence input vector and the current waveform diagram of the predetermined period through a cross-mode joint encoder including a sequence encoder and a waveform image encoder to obtain a multi-mode feature matrix of the electrical power system; the spatial feature enhancement module 140 is configured to pass the multi-modal feature matrix of the power system through the spatial attention module to obtain a classification feature matrix; the power system detection module 150 is configured to pass the classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether the monitored power system is normal; and an early warning module 160, configured to generate a control signal based on the classification result, where the control signal is used to indicate whether to generate an alarm signal.
More specifically, in the embodiment of the present application, the data acquisition module 110 is configured to acquire electrical energy load values of the monitored electrical power system at a plurality of predetermined time points within a predetermined time period and a current waveform diagram of the predetermined time period. When short circuit or other abnormal conditions occur in the power system, obvious changes in the current waveform diagram can occur, at the moment, the power distribution cabinet can timely identify the abnormal conditions and take measures to protect the abnormal conditions, and accidents are avoided. In addition, the power distribution cabinet can also evaluate the running state of the electric equipment in real time by monitoring the electric energy load value, discover potential fault hidden danger and process in time, and improve the stability and reliability of the electric power system.
For example, in one example, a data collection instrument may be used, with the current and voltage sensors of the monitored power system connected to the data collection instrument, which is then connected to a computer. And setting a preset time point and a preset time period by using data acquisition software on a computer, so as to obtain the required electric energy load value and current waveform diagram. The data acquisition instrument can acquire current and voltage data in real time, calculate an electric energy load value at the same time, and store the data on a computer, so that the subsequent data analysis and processing are convenient.
In another example, the intelligent power monitoring system can also monitor the current, voltage, power and other data of the power system in real time, and store the data on the cloud server. By using the intelligent power monitoring system, the electric energy load values and the current waveform diagrams of the monitored power system at a plurality of preset time points in preset time can be conveniently obtained. The user can acquire the required data only by setting a preset time point and a preset time period on the interface of the intelligent power monitoring system.
More specifically, in the embodiment of the present application, the data parameter timing arrangement module 120 is configured to arrange the electrical energy load values at the plurality of predetermined time points into an electrical energy load timing input vector according to a time dimension. In order to extract time sequence change characteristics of the electric energy load values, so as to comprehensively detect faults of the electric power system by combining current time sequence implicit characteristic information of the electric power system, the electric energy load values at a plurality of preset time points need to be further arranged into electric energy load time sequence input vectors according to the time dimension, and distribution information of the electric energy load values in time sequence is integrated.
In one example, the electrical energy load values at each point in time may be arranged in a time series order to form a vector. For example, if we want to obtain the power load values per hour of the day, we can time-sequence these values to form a vector. The input vector may be obtained by the following electrical energy load arrangement formula:
x＝[x 1 ,x 2 ,x 3 ,…,x n ]
wherein,x n represents the electrical energy load value at the nth time point, n representing the total time point.
More specifically, in the embodiment of the present application, the joint encoding module 130 is configured to pass the electrical energy load time sequence input vector and the current waveform chart of the predetermined period through a cross-mode joint encoder including a sequence encoder and a waveform image encoder to obtain a multi-mode feature matrix of the electrical power system. And processing the current waveform diagram of the preset time period by using a waveform image encoder which has excellent performance in the aspect of extracting the implicit characteristics of the image and comprises a convolutional neural network model, so as to mine the current signal implicit characteristic distribution information about the power system in the image, thereby being beneficial to fault detection of the power system. Particularly, the power distribution cabinet is considered to be capable of real-time evaluating the running state of the electric equipment by monitoring the time sequence change condition of the electric energy load value, so that potential fault hidden danger is found and timely processed. Therefore, in order to be able to further improve the fault detection accuracy for the power system, in the technical solution of the present application, it is desirable to optimize the expression of the current waveform timing implication feature based on the timing variation feature of the electrical energy load value by using cross-mode joint coding including a sequence encoder and a waveform image encoder.
In particular, here, the waveform image encoder performs feature mining of the current waveform map using a convolutional neural network model as a filter to extract current waveform implicit feature distribution information about a power system in the current waveform map; the sequence encoder uses a convolutional neural network based on a one-dimensional convolutional kernel to perform feature mining of the power load time sequence input vector so as to extract time sequence related feature information of the power load in a time dimension. And then, based on a joint coding module of the cross-mode joint coder, joint coding optimization of the current waveform characteristics and the power load time sequence change characteristics is completed, so that the characteristic optimization expression is carried out on the current waveform characteristic distribution based on the power load time sequence change characteristic distribution, and the accuracy of subsequent classification is improved.
It is worth mentioning that convolutional neural network (Convolutional Neural Network, CNN) is a deep learning model, and is commonly used in the fields of image recognition, natural language processing, and the like. The convolution layer in CNN can be regarded as a filter, and the convolution operation is performed on the input data by means of a sliding window, so as to extract the characteristics of the data. In power monitoring, a CNN model may be used as a filter to perform feature mining on the current waveform map. Specifically, the CNN model may gradually extract local features and global features in the current waveform map through a combination of multiple convolution layers and pooling layers. For example, the convolution layer may extract local features in the current waveform map, such as peaks, valleys, etc.; the pooling layer can downsample the feature map output by the convolution layer, so that the size of the feature map is reduced, and the calculation efficiency of the model is improved. Through multiple rolling and pooling operations, the CNN model can gradually extract more abstract features, and finally the features are input into a full-connection layer for classification or regression and other tasks.
A convolutional neural network (1D CNN) based on one-dimensional convolution kernels is a neural network model that is dedicated to processing time-series data. The 1D CNN performs feature extraction and dimension reduction on time series data by a convolution operation similar to that in image processing. The convolutional neural network based on the one-dimensional convolutional kernel is used for carrying out characteristic mining on the time sequence input vector of the electric energy load, and tasks such as prediction, analysis and optimization of the electric energy load can be realized. Specifically, the 1D CNN may gradually extract local and global features in the power load time sequence input vector through a combination of multiple convolution layers and pooling layers, for example, may extract features such as periodicity, trending, and peak of the power load. Through multiple rolling and pooling operations, the 1D CNN can gradually extract more abstract features, and finally the features are input into a full connection layer for classification or regression and other tasks. In summary, convolutional neural networks based on one-dimensional convolutional kernels are a very efficient time-series data processing tool, and can be widely applied to various tasks in power systems.
Accordingly, in one possible implementation, as shown in fig. 2, the joint coding module 130 includes: a sequence encoding unit 131, configured to pass the electrical energy load time sequence input vector through the sequence encoder of the cross-mode joint encoder to obtain an electrical energy load time sequence feature vector; a waveform image coding unit 132 for passing the current waveform diagram of the predetermined period of time through the waveform image encoder of the cross-mode joint encoder to obtain a waveform image coding feature vector; and a cross-mode fusion unit 133, configured to fuse the electrical energy load time sequence feature vector and the waveform image coding feature vector by using a cross-mode fusion device of the cross-mode joint encoder to obtain the multi-mode feature matrix of the electrical power system.
Accordingly, in one possible implementation, the sequence encoding unit 131 is configured to: and respectively carrying out one-dimensional convolution processing, pooling processing and nonlinear activation processing on input data in forward transfer of layers by using each layer of the sequence encoder of the cross-mode joint encoder to output the power load time sequence characteristic vector by the last layer of the sequence encoder of the cross-mode joint encoder, wherein the input of the first layer of the sequence encoder of the cross-mode joint encoder is the power load time sequence input vector.
Accordingly, in one possible implementation, the waveform image coding unit 132 is configured to: performing convolution processing, feature matrix-based averaging pooling processing and nonlinear activation processing on input data in forward transfer of layers by using each layer of the waveform image encoder of the cross-mode joint encoder to output the waveform image encoding feature vector by the last layer of the waveform image encoder of the cross-mode joint encoder, wherein an input of a first layer of the waveform image encoder of the cross-mode joint encoder is a current waveform diagram of the predetermined period of time.
More specifically, in the embodiment of the present application, the spatial feature enhancement module 140 is configured to pass the multi-modal feature matrix of the power system through a spatial attention module to obtain a classification feature matrix. Considering that the hidden characteristic of the current waveform and the electric energy load characteristic have special key associated characteristic information in certain time periods when fault detection of the electric power system is actually carried out, such as the condition that the current waveform suddenly increases in certain time periods, the key associated characteristic information has important significance for fault detection of the electric power system. Therefore, in order to further improve the accuracy of classification, in the technical scheme of the application, the multi-mode feature matrix of the power system is passed through a spatial attention module to obtain a classification feature matrix. It should be understood that the associated features extracted by the spatial attention module reflect weights of the spatial dimension feature differences, so as to suppress or strengthen features of different spatial positions, that is, the spatial attention module can adjust weights of different areas of the multi-modal feature matrix of the power system, so that the classifier focuses more on the areas with the greatest contribution to the classification result, thereby effectively improving the recognition and classification capability of the intelligent power distribution cabinet and improving the classification accuracy.
The spatial attention module is an attention mechanism used in deep learning and is mainly used for processing spatial features. The method can adaptively learn the importance of the features at different positions in the image, so that excellent performance is obtained in tasks such as image classification, target detection, image segmentation and the like. When using the spatial attention module, the input feature map needs to be first subjected to convolution operation to obtain the channel attention and the spatial attention respectively. The channel attention and the spatial attention are then multiplied to obtain a final attention map, which is then multiplied to the input feature map to obtain a weighted feature map. The spatial attention module has the main beneficial effect of improving the performance of the model, and particularly has outstanding performance when processing large-scale images. In addition, the space attention module can also reduce the parameter number of the model and improve the calculation efficiency of the model. Typically, the spatial attention module is applied in tasks such as image classification, object detection, image segmentation, etc. In image classification, the spatial attention module can adaptively learn the importance of features at different positions in an image, so that the classification accuracy is improved. In target detection, the spatial attention module can improve the accuracy of the detection frame, so that the detection accuracy is improved. In image segmentation, the spatial attention module can adaptively learn the importance of features at different positions, thereby improving the accuracy of segmentation.
Accordingly, in one possible implementation, the spatial feature enhancement module 140 is configured to: input data are respectively carried out in the forward transmission process of each layer of the spatial attention module: convolving the input data to generate a convolved feature map; pooling the convolution feature map to generate a pooled feature map; non-linearly activating the pooled feature map to generate an activated feature map; calculating the mean value of each position of the activation feature map along the channel dimension to generate a spatial feature matrix; calculating a Softmax-like function value of each position in the space feature matrix to obtain a space score matrix; calculating the position-wise dot multiplication of the spatial feature matrix and the spatial score matrix to obtain a feature matrix; wherein the feature matrix output by the last layer of the spatial attention module is the classification feature matrix.
More specifically, in the embodiment of the present application, the power system detection module 150 is configured to pass the classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether the monitored power system is normal. That is, in the technical solution of the present application, the tag of the classifier includes a monitored power system being normal (first tag) and a monitored power system being abnormal (second tag), wherein the classifier determines to which classification tag the classification feature matrix belongs through a soft maximum function. It should be noted that the first tag p1 and the second tag p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether the monitored power system is normal", which is simply that there are two kinds of classification tags and the probability that the output characteristic is under the two kinds of classification tags, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the monitored power system is normal is actually converted into the classified probability distribution conforming to the classification of the natural law through classifying the labels, and the physical meaning of the natural probability distribution of the labels is essentially used instead of the language text meaning of whether the monitored power system is normal. It should be understood that, in the technical solution of the present application, the classification label of the classifier is a detection evaluation label for whether the monitored power system is normal, so after the classification result is obtained, a control signal may be generated based on the classification result, where the control signal is used to indicate whether an alarm signal is generated. Therefore, fault detection of the power system can be accurately performed in real time, and an alarm is sent out through the intelligent power distribution cabinet when the fault is detected, so that safety and stability of the power system are improved.
It should be appreciated that the role of the classifier is to learn the classification rules and classifier using a given class, known training data, and then classify (or predict) the unknown data. Logistic regression (logistics), SVM, etc. are commonly used to solve the classification problem, and for multi-classification problems (multi-class classification), logistic regression or SVM can be used as well, but multiple bi-classifications are required to compose multiple classifications, but this is error-prone and inefficient, and the commonly used multi-classification method is the Softmax classification function.
More specifically, in the embodiment of the present application, the early warning module 160 is configured to generate a control signal based on the classification result, where the control signal is used to indicate whether to generate the warning signal. The early warning module is typically built on the basis of a classification model. In the classification model, input data can be classified to obtain different classification results. In the early warning module, a control signal can be generated according to the classification result and used for indicating whether an alarm signal is generated or not.
Specifically, the early warning module may set thresholds according to the classification result, and generate a control signal when the classification result exceeds the thresholds. For example, in a classification model, the classification result may be set to 0 or 1, and when the classification result is 1, a control signal may be generated to indicate that an alarm signal needs to be generated. The control signal generated by the alert module is typically a binary number, such as 0 or 1, indicating whether an alert signal needs to be generated. When the control signal is 1, it indicates that the warning signal needs to be generated, and when the control signal is 0, it indicates that the warning signal does not need to be generated. The control signal can be implemented by a circuit, software and the like, for example, warning is performed by controlling light, sound and the like. In the aspect of controlling the light, the brightness and the flicker frequency of the light can be controlled by controlling a switch or adjusting voltage and the like; in terms of controlling sound, the size and tone of sound can be controlled by controlling the vibration frequency and amplitude of the horn.
Of course, the control can also be performed mechanically, the mechanical control is a control signal mode realized based on mechanical equipment, and the light, sound and the like can be controlled through a mechanical device. For example, in controlling the light, the brightness and flicker frequency of the light may be controlled by a mechanical switch or an adjustable mechanical structure; in controlling sound, the size and pitch of sound can be controlled by mechanical vibration means.
Accordingly, in one possible implementation, the intelligent power distribution cabinet further includes a training module for training the cross-modal joint encoder including the sequence encoder and the waveform image encoder, the spatial attention module, and the classifier. As shown in fig. 3, the training module 200 includes: a training data acquisition unit 210, configured to acquire training data, where the training data includes training power load values of a monitored power system at a plurality of predetermined time points in a predetermined time and training current waveform diagrams of the predetermined time period, and a true value of whether to generate a warning signal; a training data parameter time sequence arrangement unit 220, configured to arrange training power load values at the plurality of predetermined time points into training power load time sequence input vectors according to a time dimension; a training joint encoding unit 230, configured to pass the training power load time sequence input vector and the training current waveform diagram of the predetermined period through the cross-mode joint encoder including the sequence encoder and the waveform image encoder to obtain a training power load time sequence feature vector and a training waveform image encoding feature vector; the training feature fusion unit 240 is configured to fuse the training power load time sequence feature vector and the training waveform image coding feature vector to obtain a multi-mode feature matrix of the training power system; the training spatial feature enhancement unit 250 is configured to pass the training power system multi-modal feature matrix through the spatial attention module to obtain a training classification feature matrix; a classification loss unit 260, configured to pass the training classification feature matrix through the classifier to obtain a classification loss function value; a pseudo-cyclic difference penalty loss unit 270, configured to calculate a pseudo-cyclic difference penalty factor of the training power load timing feature vector and the training waveform image coding feature vector as a pseudo-cyclic difference penalty loss function value; and a model training unit 280 for training the cross-modal joint encoder including the sequence encoder and the waveform image encoder, the spatial attention module, and the classifier with a weighted sum of the classification loss function value and the pseudo-cyclic difference penalty loss function value as a loss function value, and by back propagation of gradient descent.
The training module is mainly responsible for training the cross-modal joint encoder, the spatial attention module and the classifier. In particular, the training module trains the cross-modal joint encoder, the spatial attention module, and the classifier with existing datasets to enable accurate classification and prediction of data. During training, the cross-modal joint encoder encodes the sequence data and the waveform image data in the dataset to obtain a richer and more comprehensive feature representation. The spatial attention module weights the important parts of the power system data by using the characteristic representation extracted by the encoder to improve the accuracy of classification and prediction. Finally, the classifier classifies and predicts the power system data using the feature representations extracted by the encoder and the spatial attention module.
The classification loss unit is mainly used for calculating the classification loss function value. In the training process, the classification loss unit classifies the data in the training data set through a classifier, and then compares the classification result with the real label to calculate a classification loss function value.
Specifically, the classification loss function value is a numerical value representing the degree of difference between the classifier prediction result and the real label. The smaller the classification loss function value is, the closer the classifier prediction result is to the real label, and the higher the classification accuracy of the classifier is. The classification loss function value is very important for the training module, and can be used as an optimization target of the training module, and parameters in the cross-mode joint encoder, the spatial attention module and the classifier are updated through a back propagation algorithm, so that the classification and prediction capability of the intelligent power distribution cabinet is improved. By continuously optimizing the classification loss function value, the intelligent power distribution cabinet can gradually improve the classification and prediction accuracy of the intelligent power distribution cabinet to the power system data, and intelligent monitoring and control of the power system are realized.
Accordingly, in one possible implementation manner, the classification loss unit 260 is configured to: processing the training classification feature matrix by using the classifier according to the following classification formula to obtain a classification result, wherein the classification formula is as follows: softmax { (W) n ,B n ):…:(W 1 ,B 1 ) Project (F), where Project (F) represents projecting the training classification feature matrix as a vector, W 1 To W n Weight matrix for all the connection layers of each layer, B 1 To B n Representing the bias matrix of each fully connected layer; and calculating a cross entropy value between the classification result and the true value as the classification loss function value.
Particularly, when the training electric energy load time sequence input vector is obtained by a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder, the training electric energy load time sequence input vector obtained by the sequence encoder is associated with the training waveform image coding feature vector obtained by the waveform image encoder in the training current waveform image in a position-by-position mode so as to obtain the multi-mode feature matrix of the training electric power system. Here, considering the direction difference of feature codes caused by the data source difference and the data mode difference of the electric energy load value and the current waveform diagram, a significant imbalance exists between the training electric energy load time sequence feature vector and the integral feature distribution of the training waveform image coding feature vector, so that the expression effect of the training power system multi-mode feature matrix obtained by the association code is affected, and the accuracy of the classification result obtained by the classifier of the training classification feature matrix obtained from the training power system multi-mode feature matrix is also affected. Therefore, the applicant of the present application further introduces a pseudo-cyclic difference penalty factor for the training power load timing feature vector and the training waveform image coding feature vector as a loss function in addition to the classification loss function for the classification feature matrix.
Accordingly, in one possible implementation, the pseudo-loop difference penalty unit 270 is configured to: calculating a pseudo-cycle difference penalty factor of the training power load time sequence feature vector and the training waveform image coding feature vector as the pseudo-cycle difference penalty loss function value according to the following loss formula; wherein, the loss formula is:
wherein V is 1 Is the time sequence characteristic vector of the training electric energy load, V 2 Is the training waveform image coding feature vector, D (V 1 ,V 2 ) Encoding a distance matrix between the training power load time sequence feature vector and the training waveform image feature vector, I.I F The Frobenius norm of the matrix, L is the length of the eigenvector, d (V 1 ,V 2 ) Is the distance between the training power load time sequence feature vector and the training waveform image coding feature vector, I.I 2 Is the two norms of the vector, log represents a logarithmic function based on 2, and alpha and beta are weighted hyper-parameters,is the pseudo-cyclic difference penalty loss function value, < >>Is vector subtraction, ++>Is vector addition.
Here, considering that the unbalanced distribution between the training power load time sequence feature vector and the training waveform image coding feature vector can cause gradient propagation abnormality in a model training process based on gradient descent back propagation, thereby forming a pseudo-loop of model parameter updating, the pseudo-loop difference penalty factor regards the pseudo-loop of model parameter updating as a real loop in a model training process of minimizing a loss function by introducing a penalty factor for expressing both spatial relations and numerical relations of closely related numerical pairs of feature values, so as to realize progressive coupling of the respective feature distributions of the training power load time sequence feature vector and the training waveform image coding feature vector in a mode of simulation activation of gradient propagation, thereby improving the expression effect of the training power system multi-mode feature matrix obtained by association coding, and improving the accuracy of the classification result obtained by the classifier of the training classification feature matrix. Therefore, fault detection of the power system can be accurately performed in real time based on the parameter change trend of the power system, and an alarm is sent out through the intelligent power distribution cabinet when the fault is detected, so that the safety and stability of the power system are improved, and the accident rate and loss are reduced.
It should be appreciated that the pseudo-loop difference penalty loss unit is primarily used to calculate the pseudo-loop difference penalty loss function value. In the training process, the pseudo-cycle difference penalty loss unit processes the electric energy load time sequence characteristic vector and the waveform image coding characteristic vector in the training data set, calculates a pseudo-cycle difference penalty factor, and uses the pseudo-cycle difference penalty factor as a pseudo-cycle difference penalty loss function value. Specifically, the pseudo-cyclic difference penalty factor is an indicator of how much the power load is different between the timing feature vector and the waveform image encoding feature vector. The pseudo-cyclic difference penalty loss function value is a value obtained by weighting the pseudo-cyclic difference penalty factors and is used for measuring the degree of difference between the sequence encoder and the waveform image encoder in the cross-mode joint encoder. The larger the pseudo-loop difference penalty loss function value, the larger the difference between the sequence encoder and the waveform image encoder, and the need for adjustment to improve the accuracy of the cross-mode joint encoder.
The pseudo-cycle difference penalty loss function value is very important for the training module, and can be used as one of optimization targets of the training module, and parameters in the cross-mode joint encoder, the spatial attention module and the classifier can be updated through a back propagation algorithm, so that the classification and prediction capability of the intelligent power distribution cabinet is improved. The intelligent power distribution cabinet can gradually improve the feature extraction capacity and classification and prediction accuracy of the power system data by continuously optimizing the pseudo-cycle difference penalty loss function value, and intelligent monitoring and control of the power system are realized.
In summary, the intelligent power distribution cabinet 100 according to the embodiment of the present application is illustrated, firstly, a current waveform diagram of a plurality of preset time points is arranged into an electrical energy load time sequence input vector according to a time dimension and then a preset time period passes through a cross-mode joint encoder including a sequence encoder and a waveform image encoder to obtain a multi-mode feature matrix of an electrical power system, then the multi-mode feature matrix of the electrical power system passes through a spatial attention module to obtain a classification feature matrix, then the classification feature matrix passes through a classifier to obtain a classification result for indicating whether a monitored electrical power system is normal, and finally, a control signal for indicating whether to generate a warning signal is generated based on the classification result. In this way, the stability and reliability of the power system can be improved.
As described above, the intelligent power distribution cabinet 100 according to the embodiment of the present application may be implemented in various terminal apparatuses, for example, a server having an intelligent power distribution control algorithm, and the like. In one example, intelligent power distribution cabinet 100 may be integrated into a terminal device as a software module and/or hardware module. For example, the intelligent power distribution cabinet 100 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the intelligent power distribution cabinet 100 may also be one of a plurality of hardware modules of the terminal apparatus.
Alternatively, in another example, the intelligent power distribution cabinet 100 and the terminal apparatus may be separate apparatuses, and the intelligent power distribution cabinet 100 may be connected to the terminal apparatus through a wired and/or wireless network and transmit the interactive information in a agreed data format.
Further, the present disclosure also provides a control system of an intelligent power distribution cabinet, which includes a control system, where the control system is configured to control any one of the foregoing intelligent power distribution cabinets.
In one example, the control system of the intelligent power distribution cabinet is a system integrating multiple technologies such as computer technology, communication technology and control technology, and is mainly used for realizing the functions of monitoring, controlling, managing and the like of the intelligent power distribution cabinet. It mainly includes two aspects of hardware and software. The hardware part mainly comprises a sensor, an actuator, a controller and the like. The sensor is used for collecting electric energy parameters of each circuit in the power distribution cabinet, such as current, voltage, power and the like, the actuator is used for controlling the switching state of each circuit in the power distribution cabinet, and the controller is used for processing and analyzing the data collected by the sensor and controlling the actuator according to the processing result. The software part mainly comprises a control algorithm, a communication protocol, a human-computer interface and other components. The control algorithm is used for processing and analyzing the data acquired by the sensor so as to realize intelligent control of the power distribution cabinet; the communication protocol is used for realizing the communication between the control system and the external equipment, such as the communication with an upper computer, a SCADA system and the like; the man-machine interface is used for providing a graphical operation interface so as to facilitate the operation and monitoring of a user.
Correspondingly, the control system of the intelligent power distribution cabinet has the following advantages: 1. the intelligent control of the power distribution cabinet is realized, the load of each circuit can be automatically adjusted, and the efficiency and the stability of the power system are improved; 2. by monitoring and analyzing the electric energy parameters in real time, faults and abnormal conditions in the electric power system can be found in time, and corresponding processing is performed; 3. by communicating with an upper computer, an SCADA system and the like, remote monitoring and control of the power system can be realized, and the management efficiency and reliability of the power system are improved; 4. through providing graphical operation interface, can make things convenient for the user to operate and monitor, improve user's use experience.
Fig. 4 shows a flow chart of an intelligent power distribution method according to an embodiment of the present disclosure. Fig. 5 shows a schematic diagram of a system architecture of an intelligent power distribution method according to an embodiment of the present disclosure. As shown in fig. 4 and 5, an intelligent power distribution method according to an embodiment of the present application includes: s110, acquiring electric energy load values of a monitored electric power system at a plurality of preset time points in preset time and a current waveform diagram of the preset time period; s120, arranging the electric energy load values of the plurality of preset time points into an electric energy load time sequence input vector according to a time dimension; s130, enabling the electric energy load time sequence input vector and the current waveform diagram of the preset time period to pass through a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder so as to obtain a multi-mode feature matrix of the electric power system; s140, passing the multi-mode feature matrix of the power system through a spatial attention module to obtain a classification feature matrix; s150, the classification feature matrix passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a monitored power system is normal or not; and S160, generating a control signal based on the classification result, wherein the control signal is used for indicating whether an alarm signal is generated or not.
The implementation process and system architecture of the intelligent power distribution method can be better understood through the above-mentioned flowcharts and architecture diagrams. Through the combination of the architecture diagram and the flow chart, the implementation process and the system architecture of the intelligent power distribution method can be more clearly understood, and the technical scheme of the disclosure can be better understood and applied.
Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described intelligent power distribution method have been described in detail in the above description of the intelligent power distribution cabinet with reference to fig. 1 to 3, and thus, repetitive descriptions thereof will be omitted.
Fig. 6 illustrates an application scenario diagram of an intelligent power distribution cabinet according to an embodiment of the present disclosure. As shown in fig. 6, in this application scenario, first, electrical energy load values (for example, D1 illustrated in fig. 6) of a monitored electrical power system at a plurality of predetermined time points and a current waveform diagram of the predetermined time period (for example, D2 illustrated in fig. 6) are acquired, then, the electrical energy load values of the plurality of predetermined time points and the current waveform diagram of the predetermined time period are input to a server (for example, S illustrated in fig. 6) in which an intelligent power distribution algorithm is deployed, wherein the server is capable of processing the electrical energy load values of the plurality of predetermined time points and the current waveform diagram of the predetermined time period using the intelligent power distribution algorithm to obtain a classification result for indicating whether the monitored electrical power system is normal, and finally, a control signal for indicating whether an alarm signal is generated based on the classification result.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (7)
1. An intelligent power distribution cabinet, characterized by comprising:
the data acquisition module is used for acquiring electric energy load values of a plurality of preset time points of the monitored electric power system in preset time and a current waveform diagram of a preset time period;
the data parameter time sequence arrangement module is used for arranging the electric energy load values of the plurality of preset time points into electric energy load time sequence input vectors according to the time dimension;
the joint coding module is used for enabling the electric energy load time sequence input vector and the current waveform diagram of the preset time period to pass through a cross-mode joint encoder comprising a sequence encoder and a waveform image encoder so as to obtain a multi-mode characteristic matrix of the electric power system;
The space feature enhancement module is used for enabling the multi-mode feature matrix of the power system to pass through the space attention module to obtain a classification feature matrix;
the power system detection module is used for passing the classification feature matrix through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a monitored power system is normal or not; and
the early warning module is used for generating a control signal based on the classification result, wherein the control signal is used for indicating whether an alarm signal is generated or not;
the intelligent power distribution cabinet further comprises a training module for training the cross-mode joint encoder comprising the sequence encoder and the waveform image encoder, the spatial attention module and the classifier;
wherein, training module includes:
the training data acquisition unit is used for acquiring training data, wherein the training data comprises training electric energy load values of a monitored electric power system at a plurality of preset time points in preset time and training current waveform diagrams of the preset time period, and whether the real value of the warning signal is generated or not;
the training data parameter time sequence arrangement unit is used for arranging the training electric energy load values of the plurality of preset time points into training electric energy load time sequence input vectors according to the time dimension;
The training joint coding unit is used for enabling the training electric energy load time sequence input vector and the training current waveform diagram of the preset time period to pass through the cross-mode joint coder comprising the sequence coder and the waveform image coder so as to obtain a training electric energy load time sequence characteristic vector and a training waveform image coding characteristic vector;
the training feature fusion unit is used for fusing the training electric energy load time sequence feature vector and the training waveform image coding feature vector to obtain a multi-mode feature matrix of the training electric power system;
the training space feature enhancement unit is used for enabling the multi-mode feature matrix of the training power system to pass through the space attention module to obtain a training classification feature matrix;
the classification loss unit is used for passing the training classification characteristic matrix through the classifier to obtain a classification loss function value;
the pseudo-cycle difference penalty loss unit is used for calculating the pseudo-cycle difference penalty factors of the training power load time sequence feature vector and the training waveform image coding feature vector as pseudo-cycle difference penalty loss function values; and
a model training unit for training the cross-modal joint encoder including the sequence encoder and the waveform image encoder, the spatial attention module, and the classifier by back propagation of gradient descent with a weighted sum of the classification loss function value and the pseudo-cyclic difference penalty loss function value as a loss function value;
Wherein the pseudo-loop difference penalty loss unit is configured to:
calculating a pseudo-cycle difference penalty factor of the training power load time sequence feature vector and the training waveform image coding feature vector as the pseudo-cycle difference penalty loss function value according to the following loss formula;
wherein, the loss formula is:
wherein V is 1 Is the time sequence characteristic vector of the training electric energy load, V 2 Is the training waveform image coding feature vector, D (V 1 ,V 2 ) Encoding a distance matrix between the training power load time sequence feature vector and the training waveform image feature vector F The Frobenius norm of the matrix, L is the length of the eigenvector, d (V 1 ,V 2 ) Is the distance between the training power load time sequence characteristic vector and the training waveform image coding characteristic vector，‖·‖ 2 Is the two norms of the vector, log represents a logarithmic function based on 2, and alpha and beta are weighted hyper-parameters,is the pseudo-loop difference penalty loss function value,is vector subtraction, ++>Is vector addition.
2. The intelligent power distribution cabinet of claim 1, wherein the joint encoding module comprises:
the sequence coding unit is used for enabling the electric energy load time sequence input vector to pass through the sequence coder of the cross-mode joint coder so as to obtain an electric energy load time sequence characteristic vector;
A waveform image coding unit, configured to pass the current waveform diagram of the predetermined period of time through the waveform image encoder of the cross-mode joint encoder to obtain a waveform image coding feature vector; and
and the cross-modal fusion unit is used for fusing the electric energy load time sequence feature vector and the waveform image coding feature vector by using a cross-modal fusion device of the cross-modal joint encoder so as to obtain the multi-modal feature matrix of the electric power system.
3. The intelligent power distribution cabinet of claim 2, wherein the sequence encoding unit is configured to:
and respectively carrying out one-dimensional convolution processing, pooling processing and nonlinear activation processing on input data in forward transfer of layers by using each layer of the sequence encoder of the cross-mode joint encoder to output the power load time sequence characteristic vector by the last layer of the sequence encoder of the cross-mode joint encoder, wherein the input of the first layer of the sequence encoder of the cross-mode joint encoder is the power load time sequence input vector.
4. An intelligent power distribution cabinet according to claim 3, wherein the waveform image coding unit is configured to:
Performing convolution processing, feature matrix-based averaging pooling processing and nonlinear activation processing on input data in forward transfer of layers by using each layer of the waveform image encoder of the cross-mode joint encoder to output the waveform image encoding feature vector by the last layer of the waveform image encoder of the cross-mode joint encoder, wherein an input of a first layer of the waveform image encoder of the cross-mode joint encoder is a current waveform diagram of the predetermined period of time.
5. The intelligent power distribution cabinet of claim 4, wherein the spatial signature enhancement module is configured to:
input data are respectively carried out in the forward transmission process of each layer of the spatial attention module:
convolving the input data to generate a convolved feature map;
pooling the convolution feature map to generate a pooled feature map;
non-linearly activating the pooled feature map to generate an activated feature map;
calculating the mean value of each position of the activation feature map along the channel dimension to generate a spatial feature matrix;
calculating a Softmax-like function value of each position in the space feature matrix to obtain a space score matrix; and
Calculating the position-wise dot multiplication of the space feature matrix and the space score matrix to obtain a feature matrix;
wherein the feature matrix output by the last layer of the spatial attention module is the classification feature matrix.
6. The intelligent power distribution cabinet of claim 5, wherein the classification loss unit is configured to:
processing the training classification feature matrix by using the classifier according to the following classification formula to obtain a classification result, wherein the classification formula is as follows: softmax { (W) n ,B n ):…:(W 1 ,B 1 ) Project (F), where Project (F) represents projecting the training classification feature matrix as a vector, W 1 To W n Weight matrix for all the connection layers of each layer, B 1 To B n Representing the bias matrix of each fully connected layer; and
and calculating a cross entropy value between the classification result and the true value as the classification loss function value.
7. The utility model provides a control system of intelligent switch board which characterized in that includes:
a control system for controlling the intelligent power distribution cabinet of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310691904.2A CN116683648B (en) | 2023-06-13 | 2023-06-13 | Intelligent power distribution cabinet and control system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310691904.2A CN116683648B (en) | 2023-06-13 | 2023-06-13 | Intelligent power distribution cabinet and control system thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116683648A CN116683648A (en) | 2023-09-01 |
| CN116683648B true CN116683648B (en) | 2024-02-20 |
Family
ID=87788740
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310691904.2A Active CN116683648B (en) | 2023-06-13 | 2023-06-13 | Intelligent power distribution cabinet and control system thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116683648B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117241443B (en) * | 2023-09-15 | 2024-06-04 | 中山市松伟照明电器有限公司 | Intelligent lighting lamp based on Internet of things and intelligent control method thereof |
| CN117202430B (en) * | 2023-09-20 | 2024-03-19 | 浙江炯达能源科技有限公司 | Energy-saving control method and system for intelligent lamp post |
| CN117292576A (en) * | 2023-10-10 | 2023-12-26 | 漳州建源电力工程有限公司 | Charging parking space management system and method based on Internet of Things |
| CN117498262A (en) * | 2023-10-31 | 2024-02-02 | 神州技测(深圳)科技有限公司 | High-voltage direct-current electronic load switch protection circuit |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101790092A (en) * | 2010-03-15 | 2010-07-28 | 河海大学常州校区 | Intelligent filter designing method based on image block encoding information |
| CN114580666A (en) * | 2020-11-30 | 2022-06-03 | 西安市双合软件技术有限公司 | Multi-mode fusion motor intelligent maintenance system |
| CN114707754A (en) * | 2022-04-25 | 2022-07-05 | 广西电网有限责任公司 | Intelligent ammeter fault prediction method and system based on BiLSTM-CNN model |
| CN114779819A (en) * | 2022-05-23 | 2022-07-22 | 泰昌科技(杭州)有限公司 | High-precision lifting control system and control method based on unmanned aerial vehicle mobile airport |
| CN115456012A (en) * | 2022-08-24 | 2022-12-09 | 华能新能源股份有限公司 | Wind power plant fan major component state monitoring system and method |
| CN115456047A (en) * | 2022-08-23 | 2022-12-09 | 华能新能源股份有限公司 | Offshore wind turbine structure state monitoring system and method based on artificial intelligence |
| CN115510962A (en) * | 2022-09-19 | 2022-12-23 | 浙江佳雪微特电机有限公司 | Water pump permanent magnet synchronous motor with fault self-monitoring function and method thereof |
| WO2023078070A1 (en) * | 2021-11-04 | 2023-05-11 | 北京有竹居网络技术有限公司 | Character recognition method and apparatus, device, medium, and product |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113393025A (en) * | 2021-06-07 | 2021-09-14 | 浙江大学 | Non-invasive load decomposition method based on Informer model coding structure |
-
2023
- 2023-06-13 CN CN202310691904.2A patent/CN116683648B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101790092A (en) * | 2010-03-15 | 2010-07-28 | 河海大学常州校区 | Intelligent filter designing method based on image block encoding information |
| CN114580666A (en) * | 2020-11-30 | 2022-06-03 | 西安市双合软件技术有限公司 | Multi-mode fusion motor intelligent maintenance system |
| WO2023078070A1 (en) * | 2021-11-04 | 2023-05-11 | 北京有竹居网络技术有限公司 | Character recognition method and apparatus, device, medium, and product |
| CN114707754A (en) * | 2022-04-25 | 2022-07-05 | 广西电网有限责任公司 | Intelligent ammeter fault prediction method and system based on BiLSTM-CNN model |
| CN114779819A (en) * | 2022-05-23 | 2022-07-22 | 泰昌科技(杭州)有限公司 | High-precision lifting control system and control method based on unmanned aerial vehicle mobile airport |
| CN115456047A (en) * | 2022-08-23 | 2022-12-09 | 华能新能源股份有限公司 | Offshore wind turbine structure state monitoring system and method based on artificial intelligence |
| CN115456012A (en) * | 2022-08-24 | 2022-12-09 | 华能新能源股份有限公司 | Wind power plant fan major component state monitoring system and method |
| CN115510962A (en) * | 2022-09-19 | 2022-12-23 | 浙江佳雪微特电机有限公司 | Water pump permanent magnet synchronous motor with fault self-monitoring function and method thereof |
Non-Patent Citations (2)
| Title |
|---|
| Penalty Function-Based Joint Diagonalization Approach for Convolutive Blind Separation of Nonstationary Sources;Wenwu Wang et al.;《IEEE TRANSACTIONS ON SIGNAL PROCESSING》;第53卷(第5期);第1654-1669页 * |
| 大型电力系统的信息分类过滤方法研究;肖招娣;《电气应用》;第34卷(第09期);第110-115页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116683648A (en) | 2023-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116683648B (en) | Intelligent power distribution cabinet and control system thereof | |
| CN116639010B (en) | Intelligent control system and method for charging pile | |
| CN113156917B (en) | Power grid equipment fault diagnosis method and system based on artificial intelligence | |
| CN111858526B (en) | Failure time space prediction method and system based on information system log | |
| CN109714324B (en) | User network abnormal behavior discovery method and system based on machine learning algorithm | |
| CN102749572B (en) | IECMAC parameter identification-based power electronic circuit failure predicting method | |
| CN116934304B (en) | Intelligent power distribution room equipment operation maintenance management system and method thereof | |
| CN113408458B (en) | Breaker fault diagnosis method based on vibration signal characteristic entropy fusion | |
| CN113762329A (en) | Method and system for constructing state prediction model of large rolling mill | |
| CN116893336B (en) | High-low voltage switching device and method thereof | |
| CN113111591B (en) | Automatic diagnosis method, device and equipment based on internal fault of modular power distribution terminal | |
| CN116824517B (en) | Substation operation and maintenance safety control system based on visualization | |
| CN117375237B (en) | Substation operation and maintenance method and system based on digital twin technology | |
| CN117113218A (en) | Visual data analysis system and method thereof | |
| CN116684878B (en) | 5G information transmission data safety monitoring system | |
| CN110046663A (en) | A kind of complex electromechanical systems fault critical state discrimination method | |
| CN114722950B (en) | Multi-mode multi-variable time sequence automatic classification method and device | |
| CN116720324A (en) | Traction substation key equipment fault early warning method and system based on prediction model | |
| CN117034123B (en) | Fault monitoring system and method for fitness equipment | |
| CN117407770A (en) | High-voltage switch cabinet fault mode classification and prediction method based on neural network | |
| CN116106672B (en) | Vehicle network resonance detection method and device based on data driving and engineering knowledge | |
| CN117691756B (en) | Safety early warning management method and system for power distribution cabinet | |
| CN116930666A (en) | Intelligent diagnosis system and diagnosis method for low-voltage complete switch cabinet | |
| CN117254593A (en) | Cloud-edge-collaboration-based intelligent management and control platform and method for power grid inspection operation | |
| Lian | Research on health analysis and prediction algorithm of coal mine underground system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |