CN112307287A - Cloud edge cooperative architecture based power internet of things data classification processing method and device - Google Patents

Cloud edge cooperative architecture based power internet of things data classification processing method and device Download PDF

Info

Publication number
CN112307287A
CN112307287A CN202011252601.3A CN202011252601A CN112307287A CN 112307287 A CN112307287 A CN 112307287A CN 202011252601 A CN202011252601 A CN 202011252601A CN 112307287 A CN112307287 A CN 112307287A
Authority
CN
China
Prior art keywords
data
layer
classification
cloud
edge
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.)
Granted
Application number
CN202011252601.3A
Other languages
Chinese (zh)
Other versions
CN112307287B (en
Inventor
卢媛
范春磊
冷小洁
栾卫平
徐康
杨尉
穆芮
顾建伟
王伟
荣俊兴
李柔霏
赵慧群
张睿
杨冉昕
王丽锋
王艳红
周子程
张志浩
黄征
贺艳丽
冯逊
周学军
张赟
杨禹太
孔亮
杜廷文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Corp of China SGCC
Weihai Power Supply Co of State Grid Shandong Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
Weihai Power Supply Co of State Grid Shandong Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by State Grid Corp of China SGCC, Weihai Power Supply Co of State Grid Shandong Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN202011252601.3A priority Critical patent/CN112307287B/en
Publication of CN112307287A publication Critical patent/CN112307287A/en
Application granted granted Critical
Publication of CN112307287B publication Critical patent/CN112307287B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/906—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5072—Grid computing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00—Indexing scheme relating to G06F9/00
    • G06F2209/50—Indexing scheme relating to G06F9/50
    • G06F2209/502—Proximity

Abstract

The invention provides a cloud edge collaborative architecture-based power Internet of things data classification processing method and device, wherein the method comprises the steps of collecting and collecting original data in an edge layer, and dividing the data into upload data and data to be processed according to the source of the data; classifying and processing the data to be processed by utilizing a random forest algorithm to obtain classified result data; and the edge layer uploads the uploaded data and the classification result data to the cloud layer, and the uploaded data and the classification result data are classified and stored by the cloud layer by using an LSTM-FCN data classification model. According to the method and the device for classifying and processing the data of the power internet of things based on the cloud edge cooperative architecture, data uploading and delay of waiting for data return can be avoided, quick response to the data is achieved, and accuracy of fault diagnosis and classification is improved.

Description

Cloud edge cooperative architecture based power internet of things data classification processing method and device
Technical Field
The invention relates to the technical field of data classification, in particular to a cloud-edge collaborative framework-based data classification processing method and device for an electric power internet of things.
Background
With the continuous development of the power internet of things technology, data in a power grid are also continuously increased. The data sources of the power industry are very complex, including equipment state information from power plants and substations, power consumption information from many regions, information from sensors monitoring various facilities in remote regions, and the like. The existing electric power internet of things based on a cloud computing architecture can transmit all the data into the cloud end for data analysis and processing, so that the cloud end bears great computing pressure, and huge broadband burden of a data transmission line can be caused. This model has not been able to meet the ever-increasing data processing needs of the industry.
Disclosure of Invention
In order to solve the problems in the prior art, the application provides a method and a device for classifying and processing data of the power internet of things based on a cloud-edge cooperative architecture, so that delay of data uploading and data return waiting is avoided, quick response to the data is achieved, and accuracy of fault diagnosis and classification is improved.
In order to achieve the above object, an aspect of the present application provides a method for classifying and processing data of an electric power internet of things based on a cloud-edge collaborative architecture, including the following steps:
step 1, collecting and collecting original data in an edge layer, and dividing the data into upload data and data to be processed according to the source of the data;
step 2, classifying and processing the data to be processed by utilizing a random forest algorithm to obtain classified result data;
and 3, uploading the uploaded data and the classification result data to a cloud layer by the edge layer, and classifying and storing the uploaded data and the classification result data by the cloud layer by using an LSTM-FCN data classification model.
In some embodiments, in the step 2, the following steps are further included: step 21, performing data preprocessing on data to be processed to obtain a characteristic data set; step 22, randomly extracting a training data subset from the feature data set by using a bagging algorithm to train a random forest model; and 23, testing the trained random forest model by using the test data set, wherein the output result of the random forest model is the classification result data of the comprehensive T CART trees.
In some embodiments, the step 2 further comprises the following steps: and 24, downloading the classification result data to an equipment layer, and packaging and waiting to send to a cloud end layer.
In some embodiments, in step 21, the data preprocessing comprises extracting mean, standard deviation, peak, kurtosis, pulse index and skewness features from the data to obtain a feature data set.
In some embodiments, in step 23, theThe CART tree was constructed as follows: each node of each tree has its corresponding data set
Figure BDA0002772070370000021
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure BDA0002772070370000022
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure BDA0002772070370000023
and selecting the jth attribute with the minimum corresponding Gini coefficient as the basis for splitting the data set to segment the data set, wherein alpha is an adaptive threshold.
In some embodiments, said step 3 comprises the steps of: step 31, receiving upload data and classification result data from an edge layer at a cloud layer; step 32, identifying and classifying the received upload data and classification result data from the edge layer, and dividing the received data into upload data and classification result data; step 33, making a decision for the classification result data according to the processing result; and step 34, classifying and storing the uploaded data by adopting an LSTM-FCN algorithm.
In some embodiments, the classification process using the LSTM-FCN algorithm in step 34 is as follows: step 341, the upload data is transferred into a time convolution block, the data passes through a time convolution layer in the block, then batch normalization is used, then the output data of the block is obtained through a ReLU activation function, the output data is transmitted to the next convolution block as input data, and the above processes are repeated twice; meanwhile, the upload data is sent to a dimension shuffling layer, then the changed data is input into an LSTM block consisting of Basic LSTM and Attention LSTM, and then is subjected to Dropout; step 342, entering data of the three stacked time volume blocks into a global average pooling layer; and 343, serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
The invention provides a cloud edge collaborative architecture-based power internet of things data classification processing device, which comprises the following modules:
the acquisition module is used for collecting and collecting original data in the edge layer and dividing the data into upload data and data to be processed according to the source of the data;
the processing module is used for classifying and processing the data to be processed by utilizing a random forest algorithm to obtain classified result data;
and the execution module is used for uploading the uploaded data and the classification result data to a cloud layer through an edge layer, and classifying and storing the uploaded data and the classification result data by using an LSTM-FCN data classification model in the cloud layer.
In some embodiments, the processing module is specifically configured to: carrying out data preprocessing on data to be processed to obtain a characteristic data set; randomly extracting a training data subset from the feature data set by using a bagging algorithm to train a random forest model; and testing the trained random forest model by using a test data set, wherein the output result of the random forest model is the classification result data of the comprehensive T CART trees.
In some embodiments, the processing module is specifically configured to: and downloading the classification result data to an equipment layer, and packaging and waiting for sending to a cloud end layer.
In some embodiments, the processing module is specifically configured to: the data preprocessing comprises extracting average, standard deviation, peak, kurtosis, pulse index and skewness characteristics from the data to obtain a characteristic data set.
In some embodiments, the processing module is specifically configured to: the CART tree is constructed as follows: each node of each tree has its corresponding data set
Figure BDA0002772070370000031
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure BDA0002772070370000041
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure BDA0002772070370000042
and selecting the jth attribute with the minimum corresponding Gini coefficient as the basis for splitting the data set to segment the data set, wherein alpha is an adaptive threshold.
In some embodiments, the execution module is specifically configured to: receiving upload data and classification result data from an edge layer at a cloud layer; identifying and classifying the received upload data and classification result data from the edge layer, and dividing the received data into upload data and classification result data; making a decision for the classification result data according to the processing result; and for uploading data, performing classified storage by adopting an LSTM-FCN algorithm.
In some embodiments, the execution module is specifically configured to: transmitting the upload data into a time convolution block, enabling the data to pass through a time convolution layer in the block, then using batch normalization, obtaining output data of the block through a ReLU activation function, enabling the output data to be transmitted to the next convolution block as input data, and repeating the process twice; meanwhile, the upload data is sent to a dimension shuffling layer, then the changed data is input into an LSTM block consisting of Basic LSTM and Attention LSTM, and then is subjected to Dropout; entering data of three stacked time volume blocks into a global average pooling layer; and serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
The beneficial effects of the scheme of the application are that the method and the device for classifying and processing the data of the power internet of things based on the cloud edge cooperative architecture avoid delay of data uploading and data return waiting by processing the data at the equipment terminal, and achieve quick response to the data; the random forest model is used for data classification, so that higher accuracy can be achieved; accuracy can also be improved by using the LSTM-FCN model.
Drawings
Fig. 1 shows an architecture diagram of a data classification processing method of an electric power internet of things based on a cloud edge collaborative architecture in an embodiment.
Detailed Description
The following further describes embodiments of the present application with reference to the drawings.
As shown in fig. 1, the method for classifying and processing the data of the power internet of things based on the cloud-edge collaborative architecture includes the following steps:
step 1, collecting and collecting original data in an edge layer, and dividing the data into upload data UpD and data to be processed PeD according to the source of the data. Specifically, the edge layer does not process the upload data UpD, and directly transmits the upload data UpD upwards to the cloud end layer; and the data to be processed PeD needs to be processed at the edge layer, the result is downloaded to the equipment layer by the insulating layer after the processing is finished, meanwhile, the processed data is uploaded to the cloud end layer, and further analysis is waited at the cloud end layer.
And 2, classifying and processing the data to be processed PeD by using a random forest algorithm to obtain classified result data PeD'.
And 3, uploading the upload data UpD and the classification result data PeD 'to a cloud layer by the edge layer, and classifying and storing the upload data UpD and the classification result data PeD' by the cloud layer by using an LSTM-FCN data classification model.
Wherein in the step 2, the following steps are further included:
and 21, preprocessing the data to be processed PeD to obtain a characteristic data set. The data preprocessing comprises extracting average, standard deviation, peak, kurtosis, pulse index and skewness characteristics from the data to obtain a characteristic data set.
And step 22, randomly extracting a training data subset from the feature data set by using a Bagging algorithm (Bagging algorithm) to train a random forest model.
And 23, testing the trained random forest model by using the test data set, wherein the output result of the random forest model is classification result data PeD' of the comprehensive T CART trees.
And 24, downloading the classification result data PeD' to an equipment layer, and packaging and waiting to send to a cloud end layer.
Specifically, in step 23, the CART tree is constructed as follows: each node of each tree has its corresponding data set
Figure BDA0002772070370000061
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure BDA0002772070370000062
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure BDA0002772070370000063
and selecting the jth attribute with the minimum corresponding Gini coefficient as the basis for splitting the data set to segment the data set, wherein alpha is an adaptive threshold.
The step 3 comprises the following steps:
step 31, receiving the upload data UpD and the classification result data PeD' from the edge layer at the cloud layer.
And 32, identifying and classifying the received upload data UpD and classification result data PeD 'from the edge layer, and dividing the received data into upload data UpD and classification result data PeD'.
And step 33, making a decision for the classification result data PeD' according to the processing result.
Step 34, for the uploaded data UpD, a classification storage is performed by using an LSTM-FCN algorithm to facilitate the analysis of the big data.
Specifically, in step 34, the process of classifying by using the LSTM-FCN algorithm is as follows:
step 341, the upload data UpD is transferred into a time volume block, the data passes through the time volume layer in the block, then batch normalization is used, the output data of the block is obtained through a ReLU activation function, the output data is transmitted to the next volume block as input data, and the above processes are repeated twice; at the same time, the upload data UpD is sent to a dimension shuffle layer (dimension shuffle layer), and then the changed data is input to an LSTM block composed of Basic LSTM and Attention LSTM, and then goes through Dropout.
Step 342, data for three stacked time volume blocks enters the global average pooling layer.
And 343, serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
The application relates to an electric power thing networking data classification processing apparatus based on cloud limit is framework in coordination, includes following module: the acquisition module is used for collecting and collecting original data in the edge layer and dividing the data into upload data UpD and data to be processed PeD according to the source of the data; the processing module is used for classifying and processing the data to be processed PeD by utilizing a random forest algorithm to obtain classified result data PeD'; and the execution module is used for uploading the upload data UpD and the classification result data PeD 'to a cloud layer through an edge layer, and classifying and storing the upload data UpD and the classification result data PeD' by using an LSTM-FCN data classification model in the cloud layer.
Wherein the processing module is specifically configured to: performing data preprocessing on data to be processed PeD to obtain a characteristic data set, wherein the data preprocessing comprises extracting features of an average value, a standard deviation, a peak value, a kurtosis, a pulse index and a skewness of the data to obtain the characteristic data set; randomly extracting a training data subset from the feature data set by using a bagging algorithm to train a random forest model; testing the trained random forest model by using a test data set, wherein the output result of the random forest model is classification result data PeD' of the comprehensive T CART trees; and downloading the classification result data PeD' to an equipment layer, and packaging and waiting for sending to a cloud end layer.
The CART tree is constructed as follows: each node of each tree has its corresponding data set
Figure BDA0002772070370000071
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure BDA0002772070370000072
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure BDA0002772070370000073
and selecting the jth attribute with the minimum corresponding Gini coefficient as the basis for splitting the data set to segment the data set, wherein alpha is an adaptive threshold.
The execution module is specifically configured to: at the cloud layer, receiving upload data UpD and classification result data PeD' from the edge layer; identifying and classifying the received upload data UpD and classification result data PeD 'from the edge layer, and dividing the received data into upload data UpD and classification result data PeD'; making a decision for the classification result data PeD' according to the processing result; for the upload data UpD, the LSTM-FCN algorithm is used for sorted storage.
The execution module is further specifically configured to: transmitting the upload data into a time convolution block, enabling the data to pass through a time convolution layer in the block, then using batch normalization, obtaining output data of the block through a ReLU activation function, enabling the output data to be transmitted to the next convolution block as input data, and repeating the process twice; meanwhile, the upload data is sent to a dimension shuffling layer, then the changed data is input into an LSTM block consisting of Basic LSTM and Attention LSTM, and then is subjected to Dropout; entering data of three stacked time volume blocks into a global average pooling layer; and serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
According to the method and the device for classifying and processing the data of the power internet of things based on the cloud-edge cooperative architecture, the data is processed at the equipment terminal, so that the delay of uploading the data and waiting for returning the data is avoided, and the quick response to the data is realized; the random forest model is used for data classification, so that higher accuracy can be achieved; accuracy can also be improved by using the LSTM-FCN model.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.

Claims (14)

1. A cloud edge collaborative architecture-based power Internet of things data classification processing method is characterized by comprising the following steps: the method comprises the following steps:
step 1, collecting and collecting original data in an edge layer, and dividing the data into upload data and data to be processed according to the source of the data;
step 2, classifying and processing the data to be processed by utilizing a random forest algorithm to obtain classified result data;
and 3, uploading the uploaded data and the classification result data to a cloud layer by the edge layer, and classifying and storing the uploaded data and the classification result data by the cloud layer by using an LSTM-FCN data classification model.
2. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 1, characterized in that: in the step 2, the method further comprises the following steps:
step 21, performing data preprocessing on data to be processed to obtain a characteristic data set;
step 22, randomly extracting a training data subset from the feature data set by using a bagging algorithm to train a random forest model;
and 23, testing the trained random forest model by using the test data set, wherein the output result of the random forest model is the classification result data of the comprehensive T CART trees.
3. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 2, characterized in that: the step 2 further comprises the following steps:
and 24, downloading the classification result data to an equipment layer, and packaging and waiting to send to a cloud end layer.
4. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 2, characterized in that: in the step 21, the data preprocessing includes extracting features of mean, standard deviation, peak, kurtosis, pulse index and skewness of the data to obtain a feature data set.
5. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 2, characterized in that: in step 23, the CART tree is constructed as follows: each node of each tree has its corresponding data set
Figure FDA0002772070360000011
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure FDA0002772070360000021
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure FDA0002772070360000022
and selecting the jth attribute with the minimum corresponding Gini coefficient as the basis for splitting the data set to segment the data set, wherein alpha is an adaptive threshold.
6. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 1, characterized in that: the step 3 comprises the following steps:
step 31, receiving upload data and classification result data from an edge layer at a cloud layer;
step 32, identifying and classifying the received upload data and classification result data from the edge layer, and dividing the received data into upload data and classification result data;
step 33, making a decision for the classification result data according to the processing result;
and step 34, classifying and storing the uploaded data by adopting an LSTM-FCN algorithm.
7. The cloud-edge collaborative architecture-based power internet of things data classification processing method according to claim 6, characterized in that: in step 34, the classification process using the LSTM-FCN algorithm is as follows:
step 341, the upload data is transferred into a time convolution block, the data passes through a time convolution layer in the block, then batch normalization is used, then the output data of the block is obtained through a ReLU activation function, the output data is transmitted to the next convolution block as input data, and the above processes are repeated twice; meanwhile, the upload data is sent to a dimension shuffling layer, then the changed data is input into an LSTM block consisting of Basic LSTM and Attention LSTM, and then is subjected to Dropout;
step 342, entering data of the three stacked time volume blocks into a global average pooling layer;
and 343, serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
8. The utility model provides an electric power thing networking data classification processing device based on cloud limit is framework in coordination which characterized in that: the system comprises the following modules:
the acquisition module is used for collecting and collecting original data in the edge layer and dividing the data into upload data and data to be processed according to the source of the data;
the processing module is used for classifying and processing the data to be processed by utilizing a random forest algorithm to obtain classified result data;
and the execution module is used for uploading the uploaded data and the classification result data to a cloud layer through an edge layer, and classifying and storing the uploaded data and the classification result data by using an LSTM-FCN data classification model in the cloud layer.
9. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the processing module is specifically configured to: carrying out data preprocessing on data to be processed to obtain a characteristic data set; randomly extracting a training data subset from the feature data set by using a bagging algorithm to train a random forest model; and testing the trained random forest model by using a test data set, wherein the output result of the random forest model is the classification result data of the comprehensive T CART trees.
10. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the processing module is specifically configured to: and downloading the classification result data to an equipment layer, and packaging and waiting for sending to a cloud end layer.
11. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the processing module is specifically configured to: the data preprocessing comprises extracting average, standard deviation, peak, kurtosis, pulse index and skewness characteristics from the data to obtain a characteristic data set.
12. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the processing module is specifically configured to: the CART tree is constructed as follows: each node of each tree has its corresponding data set
Figure FDA0002772070360000031
According to its properties, it is divided into two data subsets X by adaptive damping coefficients1，X2Wherein x isiRepresents the ith data with the expression of xi＝{d1,d2,...,dmN is the number of data, m is the number of features, and the adaptive damping coefficient formula corresponding to the data set X is:
Figure FDA0002772070360000032
wherein the function f (d)j) To count djThe number of occurrences; the adaptive kurtosis coefficient formula of the data set X divided according to the jth attribute is as follows:
Figure FDA0002772070360000033
wherein alpha is an adaptive threshold, and the jth attribute with the minimum corresponding Gini coefficient is selected as the basis for the data set splittingThe data set is line partitioned.
13. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the execution module is specifically configured to: receiving upload data and classification result data from an edge layer at a cloud layer; identifying and classifying the received upload data and classification result data from the edge layer, and dividing the received data into upload data and classification result data; making a decision for the classification result data according to the processing result; and for uploading data, performing classified storage by adopting an LSTM-FCN algorithm.
14. The cloud-edge-collaborative-architecture-based data classification processing device for the power internet of things is characterized in that: the execution module is specifically configured to: transmitting the upload data into a time convolution block, enabling the data to pass through a time convolution layer in the block, then using batch normalization, obtaining output data of the block through a ReLU activation function, enabling the output data to be transmitted to the next convolution block as input data, and repeating the process twice; meanwhile, the upload data is sent to a dimension shuffling layer, then the changed data is input into an LSTM block consisting of Basic LSTM and Attention LSTM, and then is subjected to Dropout; entering data of three stacked time volume blocks into a global average pooling layer; and serially connecting the output data of the global average pooling layer and the LSTM block, and sending the output data to the softmax classification layer for classification to obtain a classification result.
CN202011252601.3A 2020-11-11 2020-11-11 Cloud edge cooperative architecture based power internet of things data classification processing method and device Active CN112307287B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011252601.3A CN112307287B (en) 2020-11-11 2020-11-11 Cloud edge cooperative architecture based power internet of things data classification processing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011252601.3A CN112307287B (en) 2020-11-11 2020-11-11 Cloud edge cooperative architecture based power internet of things data classification processing method and device

Publications (2)

Publication Number Publication Date
CN112307287A true CN112307287A (en) 2021-02-02
CN112307287B CN112307287B (en) 2022-08-02

Family

ID=74326001

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011252601.3A Active CN112307287B (en) 2020-11-11 2020-11-11 Cloud edge cooperative architecture based power internet of things data classification processing method and device

Country Status (1)

Country Link
CN (1) CN112307287B (en)

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105930794A (en) * 2016-04-20 2016-09-07 东北大学 Indoor scene identification method based on cloud computing
CN107770263A (en) * 2017-10-16 2018-03-06 电子科技大学 A kind of internet-of-things terminal safety access method and system based on edge calculations
CN108460089A (en) * 2018-01-23 2018-08-28 哈尔滨理工大学 Diverse characteristics based on Attention neural networks merge Chinese Text Categorization
CN109347924A (en) * 2018-09-20 2019-02-15 西北大学 A kind of recommended method based on intelligent perception
CN109907733A (en) * 2019-04-10 2019-06-21 西北工业大学 A kind of ECG signal analysis method towards abnormal heart rhythms classification
CN109961097A (en) * 2019-03-20 2019-07-02 西北大学 Image classification dispatching method based on edge calculations under a kind of embedded scene
CN111046931A (en) * 2019-12-02 2020-04-21 北京交通大学 Turnout fault diagnosis method based on random forest
CN111142049A (en) * 2020-01-16 2020-05-12 合肥工业大学 Intelligent transformer fault diagnosis method based on edge cloud cooperation mechanism
CN111436944A (en) * 2020-04-20 2020-07-24 电子科技大学 Falling detection method based on intelligent mobile terminal
CN111832417A (en) * 2020-06-16 2020-10-27 杭州电子科技大学 Signal modulation pattern recognition method based on CNN-LSTM model and transfer learning
CN111830408A (en) * 2020-06-23 2020-10-27 朗斯顿科技(北京)有限公司 Motor fault diagnosis system and method based on edge calculation and deep learning

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105930794A (en) * 2016-04-20 2016-09-07 东北大学 Indoor scene identification method based on cloud computing
CN107770263A (en) * 2017-10-16 2018-03-06 电子科技大学 A kind of internet-of-things terminal safety access method and system based on edge calculations
CN108460089A (en) * 2018-01-23 2018-08-28 哈尔滨理工大学 Diverse characteristics based on Attention neural networks merge Chinese Text Categorization
CN109347924A (en) * 2018-09-20 2019-02-15 西北大学 A kind of recommended method based on intelligent perception
CN109961097A (en) * 2019-03-20 2019-07-02 西北大学 Image classification dispatching method based on edge calculations under a kind of embedded scene
CN109907733A (en) * 2019-04-10 2019-06-21 西北工业大学 A kind of ECG signal analysis method towards abnormal heart rhythms classification
CN111046931A (en) * 2019-12-02 2020-04-21 北京交通大学 Turnout fault diagnosis method based on random forest
CN111142049A (en) * 2020-01-16 2020-05-12 合肥工业大学 Intelligent transformer fault diagnosis method based on edge cloud cooperation mechanism
CN111436944A (en) * 2020-04-20 2020-07-24 电子科技大学 Falling detection method based on intelligent mobile terminal
CN111832417A (en) * 2020-06-16 2020-10-27 杭州电子科技大学 Signal modulation pattern recognition method based on CNN-LSTM model and transfer learning
CN111830408A (en) * 2020-06-23 2020-10-27 朗斯顿科技(北京)有限公司 Motor fault diagnosis system and method based on edge calculation and deep learning

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
李度洋: "智能视觉传感网络中的多目标检测识别系统的研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》, 15 August 2020 (2020-08-15) *
杨崧: "基于LSTM-Attention的中文新闻标题分类研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》, 15 September 2020 (2020-09-15) *
詹国旗等: "基于特征空间优化的随机森林算法在GF-2影像湿地分类中的研究", 《地球信息科学学报》, vol. 20, no. 10, 25 October 2018 (2018-10-25) *

Also Published As

Publication number Publication date
CN112307287B (en) 2022-08-02

Similar Documents

Publication Publication Date Title
CN109271374A (en) A kind of database health scoring method and scoring system based on machine learning
CN108231086A (en) A kind of deep learning voice enhancer and method based on FPGA
CN113159345A (en) Power grid fault identification method and system based on fusion neural network model
CN112668630B (en) Lightweight image classification method, system and equipment based on model pruning
CN112381763A (en) Surface defect detection method
CN112418175A (en) Rolling bearing fault diagnosis method and system based on domain migration and storage medium
CN112347910B (en) Signal fingerprint identification method based on multi-mode deep learning
CN109800795A (en) A kind of fruit and vegetable recognition method and system
CN112288700A (en) Rail defect detection method
CN115576293B (en) Pressure-sensitive adhesive on-line production analysis method and system based on data monitoring
CN111290922A (en) Service operation health degree monitoring method and device
CN110780878A (en) Method for carrying out JavaScript type inference based on deep learning
CN111310918A (en) Data processing method and device, computer equipment and storage medium
CN109446931B (en) Animal movement behavior discrimination method and device based on time sequence correlation analysis
CN112307287B (en) Cloud edge cooperative architecture based power internet of things data classification processing method and device
CN110726813B (en) Electronic nose prediction method based on double-layer integrated neural network
CN114626426A (en) Industrial equipment behavior detection method based on K-means optimization algorithm
CN112395952A (en) A unmanned aerial vehicle for rail defect detection
CN112101487A (en) Compression method and device for fine-grained recognition model
CN113535667A (en) Method, device and system for automatically analyzing system logs
CN110543675A (en) Power transmission line fault identification method
CN111797991A (en) Deep network model compression system, method and device
CN115238806A (en) Sample class imbalance federal learning method and related equipment
CN113723442B (en) Electronic nose gas identification method, system, electronic equipment and storage medium
CN115205712A (en) Unmanned aerial vehicle inspection image quick duplicate checking method and system based on twin network

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