CN117114712A - Power utilization side carbon monitoring method, model training method, device and computer equipment - Google Patents
Power utilization side carbon monitoring method, model training method, device and computer equipment Download PDFInfo
- Publication number
- CN117114712A CN117114712A CN202311055286.9A CN202311055286A CN117114712A CN 117114712 A CN117114712 A CN 117114712A CN 202311055286 A CN202311055286 A CN 202311055286A CN 117114712 A CN117114712 A CN 117114712A
- Authority
- CN
- China
- Prior art keywords
- data
- sample
- frequency
- electricity
- super
- 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.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 93
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 86
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 title claims abstract description 78
- 238000012549 training Methods 0.000 title claims abstract description 69
- 230000005611 electricity Effects 0.000 claims abstract description 180
- 238000011084 recovery Methods 0.000 claims abstract description 78
- 230000008447 perception Effects 0.000 claims description 41
- 238000004590 computer program Methods 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 22
- 238000005457 optimization Methods 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 11
- 238000012806 monitoring device Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012821 model calculation Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 239000000047 product Substances 0.000 description 3
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 239000004753 textile Substances 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 1
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 229910002092 carbon dioxide Inorganic materials 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000013507 mapping Methods 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
- 230000003287 optical effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000005406 washing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Public Health (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Finance (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The application relates to a power utilization side carbon monitoring method, a model training method, a device and computer equipment. The electricity side carbon monitoring method comprises the following steps: acquiring low-frequency electricity utilization data of a target monitoring electricity utilization side; inputting the low-frequency electricity data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through an initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result; and realizing carbon monitoring on the target monitoring electricity side based on the high-frequency electricity data. By adopting the method, accurate carbon monitoring can be carried out on the target monitoring power utilization side.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a power consumption side carbon monitoring method, a model training method, a device, and a computer device.
Background
In smart grids, accurate data acquisition is the basis for safe and economical operation of the whole system. With the continuous deepening of the fusion of information and physical systems, the requirements of tasks such as various big data applications, real-time control and the like on the acquisition of high-frequency data are continuously improved. However, increasing the data sampling frequency necessarily places a higher burden on the system for data communication and storage.
In the conventional technology, the low-frequency data is subjected to numerical filling by a conventional linear interpolation method after the low-frequency data is acquired, and the data accuracy is poor by the linear interpolation method although the data can be denser.
Disclosure of Invention
Based on the foregoing, it is necessary to provide a power-consumption-side carbon monitoring method, a model training device and a computer device capable of accurately recovering high-frequency data based on low-frequency data.
In a first aspect, the present application provides a method for power side carbon monitoring. The method comprises the following steps:
acquiring low-frequency electricity utilization data of a target monitoring electricity utilization side;
inputting the low-frequency electricity data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through an initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result;
and realizing carbon monitoring on the target monitoring electricity side based on the high-frequency electricity data.
In one embodiment, the training process of the target super-resolution perceptual model is as follows:
acquiring a power consumption sample data set; the electricity utilization sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity utilization side;
inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and carrying out parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and the recovery result.
In one embodiment, parameter optimization is performed on an initial super-resolution perceptual model based on a difference between sample high-frequency data belonging to the same sample power side as the sample low-frequency data and a recovery result, including:
constructing a loss function;
determining a loss function calculation value according to sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and a recovery result;
and updating model parameters of the initial super-resolution perception model according to the loss function calculated value.
In one embodiment, carbon monitoring is implemented on a target monitoring electricity side based on high frequency electricity data, comprising:
acquiring a unit electricity consumption carbon emission factor of a target monitoring electricity consumption side;
and determining the carbon emission amount of the target monitoring electricity side according to the high-frequency electricity consumption data and the unit electricity consumption carbon emission factor.
In a second aspect, the application further provides a training method of the super-resolution perception model. The method comprises the following steps:
acquiring a power consumption sample data set; the electricity utilization sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity utilization side;
inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and carrying out parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and the recovery result.
In a third aspect, the application further provides a power utilization side carbon monitoring device. The device comprises:
the low-frequency data acquisition module is used for acquiring low-frequency electricity utilization data of the target monitoring electricity utilization side;
the super-resolution sensing module is used for inputting the low-frequency electricity data into the trained target super-resolution sensing model, and recovering the data through the target super-resolution sensing model to obtain the high-frequency electricity data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through an initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result;
and the carbon monitoring module is used for realizing carbon monitoring on the target monitoring electricity utilization side based on the high-frequency electricity utilization data.
In a fourth aspect, the application further provides a training device of the super-resolution perception model. The device comprises:
the data set acquisition module is used for acquiring a power consumption sample data set; the electricity utilization sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity utilization side;
the model calculation module is used for inputting the electricity sample data set into the initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and the parameter optimization module is used for performing parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side and the recovery result.
In a fifth aspect, the present application further provides a computer device, where the computer device includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the above power consumption side carbon monitoring method or the steps of the above training method of the super resolution perception model when executing the computer program.
In a sixth aspect, the present application also provides a computer readable storage medium. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the above-described electricity-side carbon monitoring method, or the steps of the above-described training method of a super-resolution perceptual model.
In a seventh aspect, the present application also provides a computer program product. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the above-described power-side carbon monitoring method, or the steps of the above-described training method of a super-resolution perceptual model.
According to the electricity-consumption-side carbon monitoring method, the model training device and the computer equipment, the target super-resolution sensing model is obtained by training an untrained initial super-resolution sensing model through an electricity-consumption sample data set, the initial super-resolution sensing model is used for carrying out data recovery on sample low-frequency data to obtain a recovery result, and the model parameters are continuously optimized by carrying out multiple times of training based on the difference between the sample high-frequency data and the recovery result. After the low-frequency electricity consumption data of the target monitoring electricity consumption side are obtained, corresponding high-frequency electricity consumption data can be obtained through the trained target super-resolution perception model, and the accuracy of the electricity consumption data is improved, so that carbon monitoring is carried out on the target monitoring electricity consumption side based on the high-frequency electricity consumption data. Compared with the linear interpolation method in the traditional technology, the technical scheme provided by the application can accurately and effectively recover the lost information of the low-frequency data and improve the accuracy of practical application such as load identification and the like.
Drawings
FIG. 1 is a diagram of an application environment for a power side carbon monitoring method in one embodiment;
FIG. 2 is a flow chart of a method for power side carbon monitoring in one embodiment;
FIG. 3 is a flow chart of a training method of a super-resolution perception model in one embodiment;
FIG. 4 is a schematic diagram of a super-resolution perceptual model in an embodiment;
FIG. 5 is a schematic flow chart of a method for monitoring carbon on the electricity side in another embodiment;
FIG. 6 is a graph of super-resolution sensing results with a frequency of 10 Hz and a sensing multiple of 10 times in one embodiment;
FIG. 7 is a graph of super-resolution sensing results with a frequency of 100 Hz and a sensing multiple of 10 times in one embodiment;
FIG. 8 is a graph of super-resolution sensing results with a frequency of 10 Hz and a sensing multiple of 100 times in one embodiment;
FIG. 9 is a block diagram of an electrical side carbon monitoring device in one embodiment;
FIG. 10 is a block diagram of a training device for a super-resolution perception model in one embodiment;
FIG. 11 is an internal block diagram of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
The electricity side carbon monitoring method provided by the embodiment of the application can be applied to an application environment shown in figure 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The terminal can be smart grid equipment at the electricity utilization side, and can detect electric quantity data at the electricity utilization side and acquire low-frequency electricity utilization data at the electricity utilization side. The server 104 may receive the low-frequency power consumption data sent by the terminal, and perform data recovery based on the obtained low-frequency power consumption data, so as to obtain corresponding high-frequency power consumption data. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.
In one embodiment, as shown in fig. 2, a method for monitoring carbon on the electricity consumption side is provided, which is illustrated by taking the application of the method to the server 104 in fig. 1 as an example, and includes the following steps S202 to S206:
s202, acquiring low-frequency electricity utilization data of the target monitoring electricity utilization side.
In an electrical power system, the electricity consumption side refers to the side where the end user or the power consumption equipment of the power grid supply is located. It is the last path for electricity to travel from the power station to the end consumer. The electricity-using side includes various types of electrical loads such as industrial plants, commercial buildings, residential areas, public facilities, offices, and the like.
The target monitoring electricity side is the electricity side that is the monitoring target in the present embodiment.
The electricity consumption data are data obtained by sampling the electricity consumption of the electricity consumption side according to a certain sampling frequency through an electricity quantity statistics device.
The low-frequency electricity data is electricity data obtained by sampling the electricity side according to a lower sampling frequency, and the high-frequency electricity data is electricity data obtained according to a higher sampling frequency.
In the present application, the low-frequency power consumption data and the high-frequency power consumption data are relatively speaking, and are not absolute low frequencies and absolute high frequencies. In this embodiment, the high-frequency power consumption data means that the sampling frequency is higher than that of the low-frequency power consumption data.
S204, inputting the low-frequency electricity data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through the initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result.
The target super-resolution sensing model is used for recovering low-frequency power consumption data into high-frequency power consumption data, and is obtained by training an untrained initial super-resolution sensing model for multiple times through a power consumption sample data set.
Specifically, the power use sample data set includes sample low frequency data and sample high frequency data. And inputting the electricity sample data set into an untrained initial super-resolution sensing model for training, and firstly, carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result as an estimated value. At the moment, the accuracy of the recovery result is low, the sample high frequency data is used as a standard value, and the accuracy of the estimated value can be improved by training for multiple times based on the difference between the estimated value and the standard value, so that the target super-resolution perception model after training is obtained.
The target super-resolution sensing model obtained after training can output accurate high-frequency data based on the input low-frequency data, so that the low-frequency power consumption data of the target monitoring power consumption side is input into the target super-resolution sensing model for data recovery, and the accurate high-frequency power consumption data can be obtained.
S206, carbon monitoring is achieved on the target monitoring electricity utilization side based on the high-frequency electricity utilization data.
Carbon monitoring on the electricity side refers to the process of monitoring and assessing carbon emissions produced by end users of the power system or by electricity consumer equipment.
After the high-frequency electricity consumption data of the target monitoring electricity consumption side is determined, the carbon emission can be estimated based on the high-frequency electricity consumption data, so that carbon monitoring is realized.
In the electricity consumption side carbon monitoring method, the target super-resolution sensing model is obtained by training an untrained initial super-resolution sensing model through an electricity consumption sample data set, the initial super-resolution sensing model is used for carrying out data recovery on sample low-frequency data to obtain a recovery result, multiple times of training is carried out based on the difference between the sample high-frequency data and the recovery result, and model parameters are optimized continuously. After the low-frequency electricity consumption data of the target monitoring electricity consumption side are obtained, corresponding high-frequency electricity consumption data can be obtained through the trained target super-resolution perception model, and the accuracy of the electricity consumption data is improved, so that carbon monitoring is carried out on the target monitoring electricity consumption side based on the high-frequency electricity consumption data. Compared with the linear interpolation method in the traditional technology, the technical scheme provided by the application can accurately and effectively recover the lost information of the low-frequency data and improve the accuracy of practical application such as load identification and the like.
In one embodiment, as shown in fig. 3, the training process of the target super-resolution perceptual model comprises the following steps:
s302, acquiring a power consumption sample data set; the electricity consumption sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity consumption side.
The power sample dataset is a previously acquired dataset for training.
Illustratively, the power consumption sample data set may be an SRPD (Super Resolution Perception Dataset, super resolution sensing database) comprising multiple sets of time-series power consumption data pairs with different frequencies for the same power consumption side of the sample, wherein each set of data pairs at least comprises one low-frequency time-series dataAnd a high-frequency time series data h.
S304, inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result.
Illustratively, the initial super-resolution perceptual model may be a sprnn (Convolutional Neural Networks for Super Resolution, super-resolution perceptual convolutional neural network). As shown in fig. 4, the super-resolution perception model network structure in the present embodiment includes an input layer, a plurality of convolution layers, and an output layer. In FIG. 4, input represents data Input, output represents data Output, conv (f i ，n i ，c i ) Represents a layer of one-dimensional convolutional neural network, deConv (f) i ，n i ，c i ) Representing a layer of one-dimensional deconvolution neural network, where f i ，n i ，c i The filter size, the number of filters, and the number of feature vectors are represented, respectively, feature extraction represents feature extraction, information supplement represents information processing, and Reconstruction represents data Reconstruction.
The time sequence electricity utilization data with different frequencies for the same electricity utilization side are respectively recorded asAnd h, wherein low frequency data +.>The frequency of (2) is marked->Data contained inThe number of dots is denoted d, its specific value at time t is denoted +.>The frequency of the high frequency data h is denoted as f h The number of data points included is denoted as α×d ()>Is a super-resolution sensing multiple), and the specific numerical value of the super-resolution sensing multiple at time t is recorded as h [ t ]]According to the Bayes formula, the posterior probability of the high-frequency data under the condition of low-frequency data can be obtained as shown in formula (1):
wherein,is the similarity of the downsampled data, p (h) is the a priori probability of h, ++>For a given +.>Is a constant.
Thereby based on low frequency dataObtaining an estimated value h' that is most similar to the true high frequency data h can be solved by solving a maximum a posteriori estimated problem as shown in equation (2):
when the initial super-resolution sensing model is not trained, the accuracy is lower, and the sample low-frequency data is input into the initial super-resolution sensing model to obtain recovery data. The restored data at this time has a difference as an estimated value from the true value of the sample high-frequency data.
S306, performing parameter optimization on the initial super-resolution sensing model based on the difference between the sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and the recovery result.
The sample high frequency data is similar to a label in model training, and is used as a theoretical value to perform parameter optimization on the model, so that recovery data obtained based on the sample low frequency data is continuously close to the sample high frequency data, the accuracy of the initial super-resolution perception model is improved, and the trained target super-resolution perception model is obtained.
In this embodiment, the untrained initial super-resolution perceptual model is trained by using the electrical sample dataset. In the training process, recovery data is obtained based on sample low frequency data through an initial super-resolution perception model, the recovery data is used as an estimated value, then model parameters are corrected through the sample high frequency data, and the model parameters are adjusted, so that the obtained recovery data is more similar to the sample high frequency data. After multiple times of training, the difference between the recovery data obtained by the super-resolution perception model and the theoretical high-frequency data can be reduced, so that the accuracy of the model is improved, and the target super-resolution perception model after training is obtained.
In one embodiment, in S306, performing parameter optimization on the initial super-resolution perceptual model based on the difference between the sample high-frequency data belonging to the same sample power-using side as the sample low-frequency data and the recovery result, including: constructing a loss function; determining a loss function calculation value according to sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and a recovery result; and updating model parameters of the initial super-resolution perception model according to the loss function calculated value.
The constructed loss function is shown in equation (3),
wherein I II 2 Representing the L-2 distance between the vectors. Loss functionThe difference between the estimated value and the theoretical high frequency data is expressed.
The training of the proposed super-resolution perceptual model can thus be performed by solving an optimization problem as shown in equation (4):
wherein θ is a parameter of the neural network in the super-resolution perception model, and F is a mapping relationship represented by the neural network, so that the model parameter is corrected according to a calculated value of the loss function.
The super-resolution sensing of the power consumption data based on the trained super-resolution sensing model can be obtained by solving an optimization problem as in formula (5):
wherein a represents the downsampling matrix,representing a distortion measure under the assumption of gaussian noise, Ω (h) represents a regularized representation containing constant prior probability information. Further, because the downsampling matrix a is constant, the optimal approximation h' is expressed as shown in equation (6):
the purpose of the multiple training is to make the calculated value of the loss function as small as possible, indicating that the smaller the difference between the estimated value and the theoretical value. The model parameters are modified and updated based on the calculated value of the loss function. Schematically, a function threshold may be set, and when the calculated value of the loss function is smaller than the function threshold, training is considered to be completed, and model parameters at that time are updated, so as to obtain a target super-resolution perception model after training is completed.
In this embodiment, the super-resolution perceptual model is trained based on a loss function, the loss function shows the difference between the high-frequency data estimated value and the theoretical value, and the model is trained based on an optimization problem of minimizing the loss function, so that the accuracy of the model is improved.
In one embodiment, S206 implements carbon monitoring on the target monitoring power side based on the high frequency power usage data, including: acquiring a unit electricity consumption carbon emission factor of a target monitoring electricity consumption side; and determining the carbon emission amount of the target monitoring electricity side according to the high-frequency electricity consumption data and the unit electricity consumption carbon emission factor.
The unit electricity consumption carbon emission factor refers to the equivalent carbon dioxide emission amount of each unit energy consumption in the electricity consumption process of the electricity consumption side. It represents the amount of carbon emissions per unit of electricity used to evaluate the carbon footprint of energy usage and compare the environmental impact of different energy types.
After the unit electricity consumption carbon emission factor of the electricity consumption side is obtained, the carbon emission amount of the target monitoring electricity consumption side can be determined based on the product of the electricity consumption amount and the unit electricity consumption carbon emission factor, as shown in a formula (7), so that the carbon monitoring of the target monitoring electricity consumption side is realized:
E＝e·h′ (7)
where e is the unit electricity carbon emission factor of the target monitoring electricity side.
In this embodiment, after the carbon emission factor of the unit electricity consumption on the electricity consumption side of the target monitoring is obtained, the product of the high-frequency electricity consumption data output by the target super-resolution sensing model and the carbon emission factor of the unit electricity consumption is used as the carbon emission amount on the electricity consumption side of the target monitoring, so as to realize carbon monitoring.
In one embodiment, as shown in FIG. 5, a power side carbon monitoring method includes the steps of:
s502, acquiring low-frequency electricity utilization data of the target monitoring electricity utilization side.
S504, inputting the low-frequency electricity data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through the initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result.
S506, acquiring the unit electricity consumption carbon emission factor of the target monitoring electricity consumption side.
S508, determining the carbon emission amount of the target monitoring electricity side according to the high-frequency electricity consumption data and the unit electricity consumption carbon emission factor.
In this embodiment, first, low-frequency power consumption data of the target monitoring power consumption side is acquired, and based on the trained target super-resolution sensing model, the low-frequency power consumption data is subjected to data recovery, so as to obtain high-frequency power consumption data. After the unit electricity consumption carbon emission factor of the target monitoring electricity consumption side is obtained, the total carbon emission of the electricity consumption side can be determined based on the product of the electricity consumption and the unit electricity consumption carbon emission factor, so that the carbon monitoring of the target monitoring electricity consumption side is realized.
To characterize the feasibility of the solution provided by the present application, in some embodiments residential user electricity data, industrial user electricity data, and transmission line voltage data are selected for analysis. For resident user load data, the low frequency data frequency is 1 Hz, 5 Hz and 10 Hz respectively, and the corresponding super-resolution sensing multiples are 2, 5, 10 and 50. For industrial users, the low frequency data frequency is 1/50 Hz, and the corresponding super-resolution perception multiples are 2, 5, 10 and 50. For transmission line voltage data, the low frequency data frequency is 10 khz and 80 khz, and the corresponding super resolution sensing multiples are 2, 5, 10 and 50. Table 1 lists information set forth in the super-resolution perception experiments.
TABLE 1 super resolution perception experiment settings
The resident user data adopted in the application is based on the power data of 11 kinds of electric appliances such as air conditioners, compact fluorescent lamps, fans, refrigerators, blowers, electric heaters, incandescent lamps, notebook computers, microwave ovens, dust collectors, washing machines and the like, and household load data is generated according to the daily electricity utilization habits and activity conditions of resident users obtained through investigation. The resulting data was sampled at a frequency of 1,000 hertz for a time period of 80,000 minutes. For the electricity consumption data of the industrial user, a certain textile mill load data set is adopted, wherein the data comprises total power data with sampling frequency of 1 Hz and time length of 30 days, which are formed by four different types of machine tools and one textile machine.
Because the payload data is time-series data and the deep super-resolution perceptual model is adapted to handle data input of a fixed length, the time-series data needs to be preprocessed and split into sub-sequences of a fixed length. In particular, in this embodiment, different time lengths are used to divide the residential and industrial user power data and the transmission line voltage data, specifically 10 seconds, 5 minutes and 0.5 milliseconds, to form a set of sub-sequences X, which are used for training and testing the super-resolution perceptual model. In the experimental process, in order to perform appropriate processing on the low-frequency electricity data X, the low-frequency electricity data X is subjected to a scaling operation as shown in formula (8). The scaling process ensures a scaled resultAnd the method is kept non-negative, and is adjusted on the basis of ensuring the fluctuation of data so as to adapt to the requirements of a super-resolution perception model. The process is monotonous, and the consistency and stability of data are guaranteed, so that model training and testing are better supported.
After obtaining the scaled set of sub-sequencesThen, the data are firstly divided into a training set, a verification set and a test set according to the proportion of the data amount, and the training set, the verification set and the test set respectively account for 75%, 5% and 20% of the total data amount. Wherein the training set is used for training the super-resolution perception model, and the verification set is used for selecting the model structureAnd parameters, and the test set is used to evaluate the metrics to obtain the effect of the final model. And respectively carrying out downsampling operation on the three segmented parts to obtain target data with different frequencies. For residential user load data, 1 hz, 2 hz, 5 hz, 10 hz, 20 hz, 50 hz, 100 hz, and 500 hz were downsampled from an original frequency of 1,000 hz, respectively. For industrial user load data, the original frequency of 1 Hz is downsampled to 1/50 Hz. For transmission line voltage data, the original frequency was 40 mhz, and downsamples were 10 hz, 20 hz, 50 hz, 80 hz, 100 hz, 160 hz, 400 hz, 500 hz, 800 hz, and 4,000 hz.
According to the super-resolution sensing experiment settings in table 1, the low-frequency data of the corresponding frequencies and the corresponding high-frequency data are paired, so that the models can be comprehensively tested and evaluated under different frequencies. The number of the deep neural network layers and the number of hidden layer nodes are required to be determined through experiments, and two indexes of model training time consumption and average absolute percentage error (Mean Absolute Percentage Error, MAPE) on a verification set are simultaneously considered in the experiments to determine the network layers and the hidden layer nodes of the neural network model. The super resolution method (SRP) provided by the application is compared with the linear interpolation method (Linear Interpolation) and the cubic interpolation method (Cubic Interpolation), judgment is carried out based on real data (group Truth), and finally the obtained super resolution sensing result is shown in figures 6-8. Wherein, fig. 6 is a data result graph with a low frequency data frequency of 10 hz and a sensing multiple of 10 times, fig. 7 is a data result graph with a low frequency data frequency of 100 hz and a sensing multiple of 10 times, and fig. 8 is a data result graph with a low frequency data frequency of 10 hz and a sensing multiple of 100 times.
Based on the data results of fig. 6-8, table 2 shows the performance comparison of the super-resolution perceptual model provided by the present application with two types of conventional perceptual models.
Table 2 perception model performance comparison table
| Experiment | Linear Interpolation | Cubic Interpolation | SRP |
| f l ＝10Hz，α＝10 | 111.9504/461.0110 | 113.9308/458.9368 | 76.2802/0.0834 |
| f l ＝100Hz，α＝10 | 101.7538/301.6763 | 109.1351/300.7959 | 21.8203/0.0347 |
| f l ＝10Hz，α＝100 | 112.8678/423.6664 | 114.7332/423.6975 | 81.0092/0.2786 |
Parameters on both sides of the diagonal line in the table represent the ratio of RMSE (Root Mean Square Error ) to DTW (Dynamic Time Warping, dynamic time warping), respectively, both of which are lower and better. As can be seen from table 2, the super-resolution perceptual model provided by the present application can provide lower errors and higher accuracy than the linear interpolation method and the cubic interpolation method.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a power utilization side carbon monitoring device for realizing the power utilization side carbon monitoring method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiment of one or more power-consumption-side carbon monitoring devices provided below may be referred to the limitation of the power-consumption-side carbon monitoring method hereinabove, and will not be repeated herein.
In one embodiment, as shown in fig. 9, there is provided a power side carbon monitoring device 900 comprising: a low frequency data acquisition module 902, a super resolution perception module 904, and a carbon monitoring module 906, wherein:
the low frequency data acquisition module 902 is configured to acquire low frequency power consumption data of the target monitoring power consumption side;
the super-resolution sensing module 904 is configured to input the low-frequency power consumption data into a target super-resolution sensing model after training, and perform data recovery through the target super-resolution sensing model to obtain high-frequency power consumption data; the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through an initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result;
the carbon monitoring module 906 is configured to implement carbon monitoring on the target monitoring electricity consumption side based on the high-frequency electricity consumption data.
In one embodiment, the training process of the target super-resolution perceptual model is as follows: acquiring a power consumption sample data set; the electricity utilization sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity utilization side; inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result; and carrying out parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and the recovery result.
In one embodiment, parameter optimization is performed on an initial super-resolution perceptual model based on a difference between sample high-frequency data belonging to the same sample power side as the sample low-frequency data and a recovery result, including: constructing a loss function; determining a loss function calculation value according to sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and a recovery result; and updating model parameters of the initial super-resolution perception model according to the loss function calculated value.
In one embodiment, the carbon monitoring module 906 is specifically configured to: acquiring a unit electricity consumption carbon emission factor of a target monitoring electricity consumption side; and determining the carbon emission amount of the target monitoring electricity side according to the high-frequency electricity consumption data and the unit electricity consumption carbon emission factor.
The above-described modules in the electricity-side carbon monitoring device may be implemented in whole or in part by software, hardware, or a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
Based on the same inventive concept, the embodiment of the application also provides a training device for the super-resolution perception model, which is used for realizing the training method of the super-resolution perception model. The implementation scheme of the device for solving the problem is similar to that described in the above method, so the specific limitation in the embodiments of the training device for one or more super-resolution sensing models provided below may be referred to the limitation of the training method for the super-resolution sensing model hereinabove, and will not be repeated here.
In one embodiment, as shown in fig. 10, there is provided a training apparatus 1000 of a super-resolution perception model, including: a data set acquisition module 1002, a model calculation module 1004, and a parameter optimization module 1006, wherein:
a data set acquisition module 1002, configured to acquire a power consumption sample data set; the electricity utilization sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity utilization side;
the model calculation module 1004 is configured to input the power consumption sample data set into an initial super-resolution sensing model, and perform data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
the parameter optimization module 1006 is configured to perform parameter optimization on the initial super-resolution sensing model based on the difference between the sample high-frequency data belonging to the same sample power utilization side as the sample low-frequency data and the recovery result.
The above-mentioned modules in the training device of the super-resolution perception model may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 11. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used to store sample electricity usage data. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by the processor, implements a power-on-side carbon monitoring method, or a training method for a super-resolution perceptual model.
It will be appreciated by those skilled in the art that the structure shown in FIG. 11 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the power-side carbon monitoring method or the steps of the training method of the super-resolution perceptual model when the computer program is executed.
In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of the above-described power-side carbon monitoring method, or the steps of the above-described training method of the super-resolution perceptual model.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the above-described power-side carbon monitoring method, or the steps of the above-described training method of a super-resolution perceptual model.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.
Claims (10)
1. A method for electricity side carbon monitoring, the method comprising:
acquiring low-frequency electricity utilization data of a target monitoring electricity utilization side;
inputting the low-frequency electricity data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity data; the target super-resolution perception model is obtained by carrying out data recovery on the sample low-frequency data through an initial super-resolution perception model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and carrying out multiple training based on the sample high-frequency data and the recovery result;
and realizing carbon monitoring on the target monitoring electricity side based on the high-frequency electricity data.
2. The method according to claim 1, wherein the training process of the target super-resolution perceptual model is as follows:
acquiring a power consumption sample data set; the electricity consumption sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity consumption side;
inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and carrying out parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side with the sample low-frequency data and the recovery result.
3. The method according to claim 2, wherein the parameter optimizing the initial super-resolution perceptual model based on a difference between the sample high-frequency data belonging to the same sample power side as the sample low-frequency data and the recovery result comprises:
constructing a loss function;
determining a loss function calculation value according to the sample high-frequency data belonging to the same sample power utilization side with the sample low-frequency data and the recovery result;
and updating model parameters of the initial super-resolution perception model according to the loss function calculated value.
4. A method according to any one of claims 1 to 3, wherein the carbon monitoring of the target electricity side based on the high frequency electricity data comprises:
acquiring a unit electricity consumption carbon emission factor of the target monitoring electricity consumption side;
and determining the carbon emission amount of the target monitoring electricity utilization side according to the high-frequency electricity utilization data and the unit electricity utilization carbon emission factor.
5. A method for training a super-resolution perceptual model, the method comprising:
acquiring a power consumption sample data set; the electricity consumption sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity consumption side;
inputting the electricity sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and carrying out parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side with the sample low-frequency data and the recovery result.
6. An electricity side carbon monitoring device, the device comprising:
the low-frequency data acquisition module is used for acquiring low-frequency electricity utilization data of the target monitoring electricity utilization side;
the super-resolution sensing module is used for inputting the low-frequency electricity consumption data into a target super-resolution sensing model after training, and recovering the data through the target super-resolution sensing model to obtain high-frequency electricity consumption data;
the carbon monitoring module is used for realizing carbon monitoring on the target monitoring electricity side based on the high-frequency electricity data;
the target super-resolution sensing model is obtained by performing data recovery on the sample low-frequency data through an initial super-resolution sensing model based on a preset electricity sample data set comprising the sample low-frequency data and the sample high-frequency data, obtaining a recovery result and performing multiple training based on the sample high-frequency data and the recovery result.
7. A training device for a super-resolution perceptual model, the device comprising:
the data set acquisition module is used for acquiring a power consumption sample data set; the electricity consumption sample data set comprises a plurality of groups of sample low-frequency data and sample high-frequency data of the same sample electricity consumption side;
the model calculation module is used for inputting the electricity consumption sample data set into an initial super-resolution sensing model, and carrying out data recovery on the sample low-frequency data through the initial super-resolution sensing model to obtain a recovery result;
and the parameter optimization module is used for performing parameter optimization on the initial super-resolution perception model based on the difference between the sample high-frequency data belonging to the same sample power utilization side with the sample low-frequency data and the recovery result.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 5 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 5.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311055286.9A CN117114712A (en) | 2023-08-18 | 2023-08-18 | Power utilization side carbon monitoring method, model training method, device and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311055286.9A CN117114712A (en) | 2023-08-18 | 2023-08-18 | Power utilization side carbon monitoring method, model training method, device and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117114712A true CN117114712A (en) | 2023-11-24 |
Family
ID=88808544
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311055286.9A Pending CN117114712A (en) | 2023-08-18 | 2023-08-18 | Power utilization side carbon monitoring method, model training method, device and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117114712A (en) |
-
2023
- 2023-08-18 CN CN202311055286.9A patent/CN117114712A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110781332A (en) | Electric power resident user daily load curve clustering method based on composite clustering algorithm | |
| CN104516808A (en) | Data preprocessing device and method thereof | |
| CN115471098B (en) | Carbon emission amount acquisition method, carbon emission amount acquisition device, computer device, and storage medium | |
| CN117556369B (en) | Power theft detection method and system for dynamically generated residual error graph convolution neural network | |
| CN112381673A (en) | Park electricity utilization information analysis method and device based on digital twin | |
| CN115563477B (en) | Harmonic data identification method, device, computer equipment and storage medium | |
| CN112418545A (en) | Load characteristic and model fusion based electricity sales amount prediction method and system | |
| CN115600840A (en) | User group carbon emission dynamic portrait model construction method, device and equipment | |
| CN115100466A (en) | Non-invasive load monitoring method, device and medium | |
| CN118228069A (en) | Method, apparatus, device, medium and program product for predicting electric load | |
| US20240144004A1 (en) | Trained Models for Discovering Target Device Presence | |
| CN114545066A (en) | Non-invasive load monitoring model polymerization method and system | |
| CN117056776A (en) | Carbon emission monitoring method, apparatus, device, medium and computer program product | |
| CN117930012A (en) | Battery consistency assessment method and device, computer equipment and storage medium | |
| CN117252448A (en) | Wind power prediction method based on space-time feature extraction and secondary decomposition polymerization | |
| CN117253368A (en) | Traffic flow abnormality detection method, device, computer equipment and storage medium | |
| CN117114712A (en) | Power utilization side carbon monitoring method, model training method, device and computer equipment | |
| CN114745299B (en) | Non-invasive load monitoring method based on sequence delay reconstruction CSP convolutional neural network | |
| CN116737681A (en) | Real-time abnormal log detection method and device, computer equipment and storage medium | |
| CN116577552A (en) | Method and device for diagnosing electricity larceny of intelligent measuring switch | |
| CN116739867A (en) | Method and device for measuring carbon emission of electric power system and computer equipment | |
| CN118779752A (en) | Health state assessment method, apparatus, device, storage medium, and program product | |
| CN109038577A (en) | Power signal self-adapting reconstruction method in load decomposition | |
| Pokhariya et al. | Discretization-Agnostic Deep Self-Supervised 3D Surface Parameterization | |
| CN118627734A (en) | Urban electric power carbon emission related factor identification method, device, equipment, medium and product |
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 |