CN115828165B - New energy intelligent micro-grid data processing method and system - Google Patents
New energy intelligent micro-grid data processing method and system Download PDFInfo
- Publication number
- CN115828165B CN115828165B CN202310116871.9A CN202310116871A CN115828165B CN 115828165 B CN115828165 B CN 115828165B CN 202310116871 A CN202310116871 A CN 202310116871A CN 115828165 B CN115828165 B CN 115828165B
- Authority
- CN
- China
- Prior art keywords
- data
- fault
- intelligent energy
- saving cabinet
- time
- 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
- 238000003672 processing method Methods 0.000 title claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 49
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 45
- 238000013145 classification model Methods 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims description 44
- 238000007781 pre-processing Methods 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 19
- 238000005065 mining Methods 0.000 claims description 18
- 238000012795 verification Methods 0.000 claims description 11
- 230000004927 fusion Effects 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 6
- 238000004519 manufacturing process Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 230000002123 temporal effect Effects 0.000 claims description 2
- 238000010200 validation analysis Methods 0.000 claims 4
- 238000004458 analytical method Methods 0.000 description 8
- 238000010606 normalization Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 101100083446 Danio rerio plekhh1 gene Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000004141 dimensional analysis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to a new energy intelligent micro-grid data processing method and a system, wherein the method collects time sequence data of an intelligent energy-saving cabinet; generating derivative class features A, B and C based on the preprocessed time series data, inputting the preprocessed data in the step S2 into an LSTM to mine out time features related to the fault of the intelligent energy-saving cabinet, inputting the derivative class features A, B and C generated in the step S3 into a convolutional neural network CNN to mine out space features related to the fault of the intelligent energy-saving cabinet, and fusing the time features and the space features; and constructing a fault classification model SVM of the intelligent energy-saving cabinet to realize fault classification. By combining the time sequence data high-efficiency processing level of the LSTM and the space data high-efficiency processing level of the CNN, the time characteristics and the space characteristics related to the faults of the intelligent energy-saving cabinet are mined out and combined and fused, and the accuracy and the efficiency of model data processing classification are improved.
Description
Technical Field
The invention relates to the field of data processing, in particular to a new energy intelligent micro-grid data processing method and system.
Background
The current prediction technology related to micro-grid monitoring is to directly send data collected by a sensor into a pre-established model for processing and analysis. For example, patent CN112070379a (publication day: 20201211) discloses a micro-grid risk monitoring and early warning system, which comprises a micro-grid, a sensor, a blockchain, a data processing device, a neural network prediction unit and a threshold abnormality detection unit which are sequentially connected, wherein the data processing device is connected with the micro-grid through a manual control center, a CNN-GRU neural network model for predicting power generation is arranged in the neural network prediction unit, a power prediction model is carried out based on the CNN-GRU neural network, and monitoring and early warning of the micro-grid are carried out by combining a threshold abnormality analysis method. The method does not consider the accuracy of the collected data and the relevance relation between the data, the related data is single, the accuracy of the data processing analysis result cannot be guaranteed, and if a large amount of data is collected for analysis, a long time is necessary to take, so that the efficiency of the data processing analysis is low; and none of the prior art process analysis of sequence data has considered a representative mode parameter. Therefore, a high-efficiency, high-accuracy and high-reliability smart micro-grid data processing method is needed.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a new energy intelligent micro-grid data processing method and system, which can improve the accuracy of the fault analysis and classification results of the intelligent energy-saving cabinets by combining the time sequence data high-efficiency processing level of the LSTM and the space data high-efficiency processing level of the CNN, mining the time characteristics and the space characteristics related to the faults of the intelligent energy-saving cabinets and combining and fusing the time characteristics and the space characteristics.
In order to achieve the above purpose, the technical solution adopted in the embodiment of the present application is as follows:
in a first aspect, the present invention provides a new energy smart micro-grid data processing method, including:
s1, collecting time sequence data of the intelligent energy-saving cabinet, wherein the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
s2, preprocessing the time sequence data, wherein the preprocessing specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Criteria to determine the contentError data are removed;
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value;
s3, generating derivative type characteristics A, B and C based on the preprocessed time series data, wherein C is the mode in one type of data in the time series data preprocessed in the step S2;
Wherein m is the total number of one type of data in the time series data preprocessed in the step S2, and x j The j-th data in one type of data in the time series data preprocessed in the step S2;
Wherein x is max For one of the types of data in the time-series data pre-processed in step S2Maximum value, x min The minimum value in one type of data in the time series data preprocessed in the step S2;
s4, inputting the data preprocessed in the step S2 into an LSTM to mine out the time characteristic related to the fault of the intelligent energy-saving cabinet, inputting the derivative type characteristic A, B generated in the step S3 and C into a convolutional neural network CNN to mine out the space characteristic related to the fault of the intelligent energy-saving cabinet, and fusing the time characteristic and the space characteristic to generate a fused characteristic as a data sample;
s5, dividing the data sample in the step S4 into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet to perform fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
s51, corresponding fault class values are arranged in the data samples, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
S52, calculating the output value in S51 and the actual fault class value q in the training set by using the design loss function k If the error is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinet is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
In an alternative embodiment, the method further comprises: and S2, the operation of preprocessing the time series data further comprises normalization of the data.
In an alternative embodiment, the method further comprises: in step S4, the LSTM network and the convolutional neural network CNN for feature mining are in parallel, that is, the LSTM mining of the temporal feature related to the fault of the intelligent energy-saving cabinet and the convolutional neural network CNN mining of the spatial feature related to the fault of the intelligent energy-saving cabinet are performed in parallel.
In an alternative embodiment, the method further comprises: and after model training is finished, inputting data in the verification set and the test set into the model to verify and test the completed model.
In a second aspect, the present invention provides a new energy smart micro-grid data processing system, the system comprising:
the acquisition module is used for acquiring time sequence data of the intelligent energy-saving cabinet, and the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
the preprocessing module is used for preprocessing the time sequence data, and specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Judging the data containing errors according to the criteria and eliminating the data; />
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value;
a feature generation module, configured to generate features A, B of the derivative class and C based on the preprocessed time-series data, where C is a mode in one type of data in the preprocessed time-series data;
Wherein m is the total number of one type of data in the preprocessed time series data, and x j J-th data in one type of data in the preprocessed time series data;
Wherein x is max Is the maximum value, x, in one type of data in the preprocessed time series data min Is the minimum value in one type of data in the preprocessed time series data;
the feature fusion module is used for inputting the preprocessed data into the LSTM to dig out the time feature related to the fault of the intelligent energy-saving cabinet, inputting the generated derivative type feature A, B and C into the convolutional neural network CNN to dig out the space feature related to the fault of the intelligent energy-saving cabinet, and fusing the time feature and the space feature to generate a fusion feature as a data sample;
the model construction module is used for dividing the data sample into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet for fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
the data samples have corresponding fault class values, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
Design loss function calculation output value and actual fault class value q in training set k If the error is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinet is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
In an alternative embodiment, the system further comprises: the operation of preprocessing the time series data in the preprocessing module further comprises normalization of the data.
In an alternative embodiment, the system further comprises: the LSTM network and the convolutional neural network CNN for mining features are in a parallel mode, namely, the LSTM mining of time features related to the fault of the intelligent energy-saving cabinet and the convolutional neural network CNN mining of space features related to the fault of the intelligent energy-saving cabinet are performed in parallel.
In an alternative embodiment, the system further comprises: and after model training is finished, inputting data in the verification set and the test set into the model to verify and test the completed model.
The beneficial effects are that:
1. the method of the invention comprises the steps of S1, collecting time sequence data of the intelligent energy-saving cabinet; s2, preprocessing the time sequence data, wherein the preprocessing specifically comprises error data processing and missing data processing; s3, generating derivative class features A, B and C and S4 based on the preprocessed time series data, inputting the preprocessed data in the step S2 into an LSTM to mine out time features related to the fault of the intelligent energy-saving cabinet, inputting the derivative class features A, B and C generated in the step S3 into a convolutional neural network CNN to mine out space features related to the fault of the intelligent energy-saving cabinet, and fusing the time features and the space features to generate fusion features as data samples; s5, dividing the data sample in the step S4 into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet to realize fault classification. According to the invention, through the steps of error data processing and missing data processing on the acquired time sequence data, errors brought by noise data to model training are reduced, the data containing the errors are judged by adopting the 3 sigma criterion, interference information is restrained, and the true reliability of the data is ensured. By combining the time sequence data high-efficiency processing level of the LSTM and the space data high-efficiency processing level of the CNN, the time characteristics and the space characteristics related to the fault of the intelligent energy-saving cabinet are mined, and the time characteristics and the space characteristics are combined and fused, so that the relevance relation between the data is mined, multi-dimensional analysis data is obtained, and the accuracy and the efficiency of model data processing and classification are improved.
2. According to the method, the characteristics of the derivative class are generated based on the preprocessed time sequence data, and the richer characteristic information is extracted; one of the derived characteristics is the mode of one type of data in the preprocessed time series data, so that the data with the largest occurrence number in each type of data is extracted as one type of data for subsequent processing analysis, the data is more representative, the redundant time consumed by data acquisition can be reduced, and the data acquisition efficiency is improved.
3. Aiming at the fault classification model SVM, a special loss function is designed to train the SVM, and the accuracy and reliability of data analysis of the fault classification model SVM are improved.
Drawings
Fig. 1 is a flow chart of steps of a new energy intelligent micro-grid data processing method.
Fig. 2 is a flowchart of a step of generating fusion features in a new energy smart micro-grid data processing method.
Detailed Description
The present application is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only the portions related to the invention are shown in the drawings.
As shown in fig. 1-2, the present embodiment provides a new energy smart micro-grid data processing method, including:
s1, collecting time sequence data of the intelligent energy-saving cabinet, wherein the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
s2, preprocessing the time sequence data, wherein the preprocessing specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Judging the data containing errors according to the criteria and eliminating the data;
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value; in particular, the method comprises the steps of,
s3, generating derivative type characteristics A, B and C based on the preprocessed time series data, wherein C is the mode in one type of data in the time series data preprocessed in the step S2;
Wherein m is the total number of one type of data in the time series data preprocessed in the step S2, and x j The j-th data in one type of data in the time series data preprocessed in the step S2;
Wherein x is max For the maximum value, x, in one type of data in the time series data preprocessed in the step S2 min The minimum value in one type of data in the time series data preprocessed in the step S2;
s4, inputting the data preprocessed in the step S2 into an LSTM to mine out the time characteristic related to the fault of the intelligent energy-saving cabinet, inputting the derivative type characteristic A, B generated in the step S3 and C into a convolutional neural network CNN to mine out the space characteristic related to the fault of the intelligent energy-saving cabinet, and fusing the time characteristic and the space characteristic to generate a fused characteristic as a data sample;
the LSTM has a good effect on mining time sequence data, the CNN has a good effect on mining space data, and the step is combined with the time sequence data high-efficiency processing level of the LSTM and the space data high-efficiency processing level of the CNN to mine time characteristics and space characteristics related to faults of the intelligent energy-saving cabinet.
S5, dividing the data sample in the step S4 into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet to perform fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
s51, corresponding fault class values are arranged in the data samples, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
S52, calculating the output value in S51 and the actual fault class value q in the training set by using the design loss function k If the error is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinet is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
Specifically, the fault types are classified into undervoltage, overload, overcurrent, overtemperature, bypass faults and the like, corresponding values can be set for each fault type according to requirements, and the corresponding values set for each fault type are different and are distinguished from each other.
In an alternative embodiment, the method further comprises: the step S2 of preprocessing the time series data further comprises normalization of the data, and specifically comprises the following steps:
in an alternative embodiment, as shown in fig. 2, the method further comprises: in the step S4, the LSTM network and the convolutional neural network CNN for excavating features are in a parallel mode, namely, the LSTM excavating time features related to the faults of the intelligent energy-saving cabinet and the convolutional neural network CNN excavating space features related to the faults of the intelligent energy-saving cabinet are simultaneously conducted in parallel, and the sequence relation is not separated, so that the data processing efficiency is improved.
In an alternative embodiment, the method further comprises: and after model training is finished, inputting data in the verification set and the test set into the model to verify and test the completed model.
Based on the same inventive concept, the present embodiment provides a new energy smart micro-grid data processing system, which includes:
the acquisition module is used for acquiring time sequence data of the intelligent energy-saving cabinet, and the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
the preprocessing module is used for preprocessing the time sequence data, and specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Judging the data containing errors according to the criteria and eliminating the data;
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value;
a feature generation module, configured to generate features A, B of the derivative class and C based on the preprocessed time-series data, where C is a mode in one type of data in the preprocessed time-series data;
Wherein m is the total number of one type of data in the preprocessed time series data, and x j J-th data in one type of data in the preprocessed time series data;
Wherein x is max Is the maximum value, x, in one type of data in the preprocessed time series data min Is the minimum value in one type of data in the preprocessed time series data;
the feature fusion module is used for inputting the preprocessed data into the LSTM to dig out the time feature related to the fault of the intelligent energy-saving cabinet, inputting the generated derivative type feature A, B and C into the convolutional neural network CNN to dig out the space feature related to the fault of the intelligent energy-saving cabinet, and fusing the time feature and the space feature to generate a fusion feature as a data sample;
the model construction module is used for dividing the data sample into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet for fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
the data samples have corresponding fault class values, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
Design loss function calculation output value and actual fault class value q in training set k When the error isWhen the number of the intelligent energy-saving cabinets is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinets is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
In an alternative embodiment, the system further comprises: the operation of preprocessing the time series data in the preprocessing module further comprises normalization of the data.
In an alternative embodiment, the system further comprises: the LSTM network and the convolutional neural network CNN for mining features are in a parallel mode, namely, the LSTM mining of time features related to the fault of the intelligent energy-saving cabinet and the convolutional neural network CNN mining of space features related to the fault of the intelligent energy-saving cabinet are performed in parallel.
In an alternative embodiment, the system further comprises: and after model training is finished, inputting data in the verification set and the test set into the model to verify and test the completed model.
Claims (8)
1. The new energy intelligent micro-grid data processing method is characterized by comprising the following steps of:
s1, collecting time sequence data of the intelligent energy-saving cabinet, wherein the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
s2, preprocessing the time sequence data, wherein the preprocessing specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Judging the data containing errors according to the criteria and eliminating the data;
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value;
s3, generating derivative type characteristics A, B and C based on the preprocessed time series data, wherein C is the mode in one type of data in the time series data preprocessed in the step S2;
Wherein m is the total number of one type of data in the time series data preprocessed in the step S2, and x j The j-th data in one type of data in the time series data preprocessed in the step S2;
Wherein x is max For the maximum value, x, in one type of data in the time series data preprocessed in the step S2 min After the pretreatment of the step S2A minimum value in one type of data in the time-series data;
s4, inputting the data preprocessed in the step S2 into an LSTM to mine out the time characteristic related to the fault of the intelligent energy-saving cabinet, inputting the derivative type characteristic A, B generated in the step S3 and C into a convolutional neural network CNN to mine out the space characteristic related to the fault of the intelligent energy-saving cabinet, and fusing the time characteristic and the space characteristic to generate a fused characteristic as a data sample;
s5, dividing the data sample in the step S4 into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet to perform fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
s51, corresponding fault class values are arranged in the data samples, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
S52, calculating the output value in S51 and the actual fault class value q in the training set by using the design loss function k If the error is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinet is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
2. The method of claim 1, wherein the step of S2 pre-processing the time series data further comprises normalizing the data.
3. The method as recited in claim 1, further comprising: in step S4, the LSTM network and the convolutional neural network CNN for feature mining are in parallel, that is, the LSTM mining of the temporal feature related to the fault of the intelligent energy-saving cabinet and the convolutional neural network CNN mining of the spatial feature related to the fault of the intelligent energy-saving cabinet are performed in parallel.
4. The method of claim 1, wherein after model training is completed, data in the validation set and the test set is entered into the model to complete validation and testing of the model.
5. The utility model provides a new forms of energy intelligence micro-grid data processing system which characterized in that, the system includes:
the acquisition module is used for acquiring time sequence data of the intelligent energy-saving cabinet, and the types of the time sequence data specifically comprise: input voltage, output voltage, load size, operating current, operating temperature, operating frequency and output efficiency;
the preprocessing module is used for preprocessing the time sequence data, and specifically comprises error data processing and missing data processing;
wherein the error data processing step adopts 3
Judging the data containing errors according to the criteria and eliminating the data;
Wherein n is the total number of one type of data in the time series data, and x i For the ith data in one type of data in the time series data, u i For the mean value of one of the types of data in the corresponding time-series data,
standard deviation for one type of data in the corresponding time series data;
The formula 2 is a judging condition of error data, and the error value is deleted after being identified;
the step of missing data processing is to use the average value of two nearest neighbors of missing values as a filling value;
a feature generation module, configured to generate features A, B of the derivative class and C based on the preprocessed time-series data, where C is a mode in one type of data in the preprocessed time-series data;
Wherein m is the total number of one type of data in the preprocessed time series data, and x j J-th data in one type of data in the preprocessed time series data;
Wherein x is max Is the maximum value, x, in one type of data in the preprocessed time series data min Is the minimum value in one type of data in the preprocessed time series data;
the feature fusion module is used for inputting the preprocessed data into the LSTM to dig out the time feature related to the fault of the intelligent energy-saving cabinet, inputting the generated derivative type feature A, B and C into the convolutional neural network CNN to dig out the space feature related to the fault of the intelligent energy-saving cabinet, and fusing the time feature and the space feature to generate a fusion feature as a data sample;
the model construction module is used for dividing the data sample into a training set, a verification set and a test set, and constructing a fault classification model SVM of the intelligent energy-saving cabinet for fault classification;
the method for constructing the intelligent energy-saving cabinet fault classification model SVM for fault classification specifically comprises the following steps:
the data samples have corresponding fault class values, a training set is used as training data of the SVM, and forward propagation is adopted to obtain a fault class output value p of the intelligent energy-saving cabinet k ；
Design loss function calculation output value and actual fault class value q in training set k If the error is smaller than the threshold value, training is finished, and a fault classification model SVM of the intelligent energy-saving cabinet is obtained;
wherein, the loss function is:
Where H is the total number of training sets, k=1, 2,.. k Is a predefined weight value.
6. The system of claim 5, wherein the operation of preprocessing the time series data in the preprocessing module further comprises normalizing the data.
7. The system of claim 5, further comprising: the LSTM network and the convolutional neural network CNN for mining features are in a parallel mode, namely, the LSTM mining of time features related to the fault of the intelligent energy-saving cabinet and the convolutional neural network CNN mining of space features related to the fault of the intelligent energy-saving cabinet are performed in parallel.
8. The system of claim 5, wherein after model training is completed, data in the validation set and the test set is entered into the model to complete validation and testing of the model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310116871.9A CN115828165B (en) | 2023-02-15 | 2023-02-15 | New energy intelligent micro-grid data processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310116871.9A CN115828165B (en) | 2023-02-15 | 2023-02-15 | New energy intelligent micro-grid data processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115828165A CN115828165A (en) | 2023-03-21 |
| CN115828165B true CN115828165B (en) | 2023-05-02 |
Family
ID=85521473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310116871.9A Active CN115828165B (en) | 2023-02-15 | 2023-02-15 | New energy intelligent micro-grid data processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115828165B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111717753A (en) * | 2020-06-29 | 2020-09-29 | 浙江新再灵科技股份有限公司 | Self-adaptive elevator fault early warning system and method based on multi-dimensional fault characteristics |
| WO2020207214A1 (en) * | 2019-04-08 | 2020-10-15 | 腾讯科技（深圳）有限公司 | Data processing method and apparatus, electronic device and storage medium |
| CN112528891A (en) * | 2020-12-16 | 2021-03-19 | 重庆邮电大学 | Bidirectional LSTM-CNN video behavior identification method based on skeleton information |
| CN112580263A (en) * | 2020-12-24 | 2021-03-30 | 湖南工业大学 | Turbofan engine residual service life prediction method based on space-time feature fusion |
| CN113222265A (en) * | 2021-05-21 | 2021-08-06 | 内蒙古大学 | Mobile multi-sensor space-time data prediction method and system in Internet of things |
| CN113988477A (en) * | 2021-11-26 | 2022-01-28 | 西安化奇数据科技有限公司 | Photovoltaic power short-term prediction method and device based on machine learning and storage medium |
| CN114544155A (en) * | 2022-01-28 | 2022-05-27 | 江苏科技大学 | AUV propeller multi-information-source fusion fault diagnosis method and system based on deep learning |
| WO2022116570A1 (en) * | 2020-12-04 | 2022-06-09 | 东北大学 | Microphone array-based method for locating and identifying fault signal in industrial equipment |
| CN114722909A (en) * | 2022-03-14 | 2022-07-08 | 山西三友和智慧信息技术股份有限公司 | Solar flare time sequence classification method based on low-dimensional convolutional neural network |
| CN115659249A (en) * | 2022-12-28 | 2023-01-31 | 成都大汇物联科技有限公司 | Intelligent station-finding control system anomaly detection method |
-
2023
- 2023-02-15 CN CN202310116871.9A patent/CN115828165B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020207214A1 (en) * | 2019-04-08 | 2020-10-15 | 腾讯科技（深圳）有限公司 | Data processing method and apparatus, electronic device and storage medium |
| CN111717753A (en) * | 2020-06-29 | 2020-09-29 | 浙江新再灵科技股份有限公司 | Self-adaptive elevator fault early warning system and method based on multi-dimensional fault characteristics |
| WO2022116570A1 (en) * | 2020-12-04 | 2022-06-09 | 东北大学 | Microphone array-based method for locating and identifying fault signal in industrial equipment |
| CN112528891A (en) * | 2020-12-16 | 2021-03-19 | 重庆邮电大学 | Bidirectional LSTM-CNN video behavior identification method based on skeleton information |
| CN112580263A (en) * | 2020-12-24 | 2021-03-30 | 湖南工业大学 | Turbofan engine residual service life prediction method based on space-time feature fusion |
| CN113222265A (en) * | 2021-05-21 | 2021-08-06 | 内蒙古大学 | Mobile multi-sensor space-time data prediction method and system in Internet of things |
| CN113988477A (en) * | 2021-11-26 | 2022-01-28 | 西安化奇数据科技有限公司 | Photovoltaic power short-term prediction method and device based on machine learning and storage medium |
| CN114544155A (en) * | 2022-01-28 | 2022-05-27 | 江苏科技大学 | AUV propeller multi-information-source fusion fault diagnosis method and system based on deep learning |
| CN114722909A (en) * | 2022-03-14 | 2022-07-08 | 山西三友和智慧信息技术股份有限公司 | Solar flare time sequence classification method based on low-dimensional convolutional neural network |
| CN115659249A (en) * | 2022-12-28 | 2023-01-31 | 成都大汇物联科技有限公司 | Intelligent station-finding control system anomaly detection method |
Non-Patent Citations (1)
| Title |
|---|
| 汪友明 等.改进的CNN-LSTM承故障诊断方法.《西安邮电大学学报》.2021,第26卷(第1期),第97-103页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115828165A (en) | 2023-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106054104B (en) | A kind of intelligent electric meter failure real-time predicting method based on decision tree | |
| CN106779505B (en) | Power transmission line fault early warning method and system based on big data driving | |
| CN110458230A (en) | A kind of distribution transforming based on the fusion of more criterions is with adopting data exception discriminating method | |
| CN112749509B (en) | Intelligent substation fault diagnosis method based on LSTM neural network | |
| CN105846780A (en) | Decision tree model-based photovoltaic assembly fault diagnosis method | |
| CN110535159B (en) | Method and system for early warning of faults of operation units of large-scale energy storage power station | |
| CN107807860B (en) | Power failure analysis method and system based on matrix decomposition | |
| Shi et al. | Expected output calculation based on inverse distance weighting and its application in anomaly detection of distributed photovoltaic power stations | |
| CN116739829B (en) | Big data-based power data analysis method, system and medium | |
| CN116308304A (en) | New energy intelligent operation and maintenance method and system based on meta learning concept drift detection | |
| CN117411436B (en) | Photovoltaic module state detection method, system and storage medium | |
| CN115758151A (en) | Combined diagnosis model establishing method and photovoltaic module fault diagnosis method | |
| CN116127354A (en) | High-voltage cable partial discharge fault positioning method and system based on deep learning | |
| CN116720095A (en) | Electrical characteristic signal clustering method for optimizing fuzzy C-means based on genetic algorithm | |
| CN111894814A (en) | Fault processing system and method for power generation system | |
| CN114581699A (en) | Transformer state evaluation method based on deep learning model in consideration of multi-source information | |
| Dong et al. | Fault diagnosis and classification in photovoltaic systems using scada data | |
| CN115828165B (en) | New energy intelligent micro-grid data processing method and system | |
| CN106646106B (en) | Electric network fault detection method based on outlier's detection technology | |
| CN117452062A (en) | Method for monitoring line loss rate of transformer area in consideration of operation load | |
| Zhang et al. | Distributed Architecture of Power Grid Asset Management and Future Research Directions | |
| Du et al. | Fault diagnosis method of automation equipment in independent and controllable substation based on deep reinforcement learning | |
| CN118133947B (en) | AI processing method and system for air compressor energy system | |
| CN118294750B (en) | Automatic point-to-point verification system based on relay protection | |
| CN117472898B (en) | Fusion-based power distribution network abnormal data error correction method and 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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240226 Address after: No.198 GANGLONG Road, Yangzhong Development Zone, Zhenjiang City, Jiangsu Province, 212000 Patentee after: Zhenjiang Anhua Electric Group Co.,Ltd. Country or region after: China Address before: Room 1001-3, Tower C, Zhimengyuan, 4 Gupinggang, Gulou District, Nanjing, Jiangsu, 210000 Patentee before: Nanjing University of Technology Jinhong Energy Technology Co.,Ltd. Country or region before: China |
|
| TR01 | Transfer of patent right |