CN114037282B - Electricity consumption label system based on electricity consumption characteristics - Google Patents
Electricity consumption label system based on electricity consumption characteristics Download PDFInfo
- Publication number
- CN114037282B CN114037282B CN202111327202.3A CN202111327202A CN114037282B CN 114037282 B CN114037282 B CN 114037282B CN 202111327202 A CN202111327202 A CN 202111327202A CN 114037282 B CN114037282 B CN 114037282B
- Authority
- CN
- China
- Prior art keywords
- industry
- electricity
- fitting
- model
- power
- 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
- 230000005611 electricity Effects 0.000 title claims abstract description 225
- 238000009826 distribution Methods 0.000 claims abstract description 50
- 238000012549 training Methods 0.000 claims abstract description 12
- 238000004364 calculation method Methods 0.000 claims description 20
- 238000012216 screening Methods 0.000 claims description 10
- 238000000034 method Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 abstract description 2
- 238000013480 data collection Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000005612 types of electricity Effects 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- 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
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a power consumption label system based on power consumption characteristics, which is used for marking an industry label for a power consumption user, and comprises an industry model subsystem, an information fitting subsystem, a power consumption label subsystem and a power consumption information database; through the data collection of each dimension of the electricity utilization user, the primary matching of the industry label is carried out on the electricity utilization user in a model training mode, the situation that the industry label cannot be matched due to the fact that the data of the electricity utilization user is not full can be avoided, the marking of the industry label is not independent marking logic any more in a dynamic correction mode in two aspects of industry distribution deviation and concentration deviation, marking accuracy is improved, marking results of each industry label can influence the industry labels of other electricity utilization users (by affecting industry distribution and concentration), and possibility is really provided for the industry automatic labels of the electricity utilization user in a big data environment.
Description
Technical Field
The invention relates to an electricity consumption data acquisition system, in particular to an electricity consumption label system based on electricity consumption characteristics.
Background
The electric power quantity is an important index of an electric power system, the 'network power grid' system gathers massive electric quantity data, the electric quantity increase is closely related to social and economic development, the macroscopic economic situation and the electric power demand trend can be better mastered through the electric quantity and economic relationship, data support is provided for planning, electric power distribution and other businesses, and the aim of 'looking into economy from electric power' is achieved. With the increase of the power analysis demands, the granularity of the data classifying the power data is higher and higher, particularly the analysis of the industries is required to be more and more refined, but at present, although a power utilization network is established and registration is carried out on the power utilization network, the problems still exist that the data format is not uniform, the data cannot be quantized, part of the data is missing and missing, the efficiency of manual analysis is lower, the subjectivity is serious, and single data cannot correspond to the actual classification demands, so that the label of the power consumption data is difficult to realize.
Disclosure of Invention
In view of the foregoing, it is an object of the present invention to provide a power consumption tagging system based on power consumption characteristics.
In order to solve the technical problems, the technical scheme of the invention is as follows:
the electricity consumption label system based on the electricity consumption characteristics is used for marking an industry label for an electricity consumption user and comprises an industry model subsystem, an information fitting subsystem, an electricity consumption label subsystem and an electricity consumption information database;
the electricity consumption information database stores electricity consumption information of electricity consumption users in real time, wherein the electricity consumption information comprises electricity consumption characteristic data, user information data and electricity consumption type data, the electricity consumption characteristic data reflects electricity consumption conditions of the electricity consumption users, the user information data reflects registered user information of the electricity consumption users, and the electricity consumption type data reflects electricity consumption types registered by the electricity consumption users;
The industry model subsystem comprises a model configuration module, a model training module and a model generation module, wherein the model configuration module is used for configuring an industry model framework, the industry model framework comprises industry quantity and industry types, the model training module processes corresponding historical electricity utilization samples according to the industry types to generate an industry sub-model, the historical electricity utilization samples are electricity utilization information corresponding to electricity utilization users of known industry labels, and the model generation module generates an industry characteristic model according to the industry sub-model and the industry model framework;
The information fitting subsystem comprises an electricity utilization characteristic fitting module, a user information fitting module and an electricity utilization type fitting module, wherein the electricity utilization characteristic fitting module is configured with a first fitting strategy, the first fitting strategy obtains electricity utilization characteristic waveforms in electricity utilization characteristic data, and waveform characteristic fitting values corresponding to each industry sub-model are obtained according to calculation according to the fact that whether similar electricity utilization characteristic waveforms exist in historical electricity utilization characteristic waveforms corresponding to the industry sub-model or not, so as to generate a first fitting array; the user information fitting module is configured with a second fitting strategy, the second fitting strategy obtains industry feature words in the user information data, and corresponding feature matching values in each industry sub-model are obtained through calculation according to whether the same industry feature words exist in the industry sub-model or not so as to generate a second fitting array; the electricity utilization type fitting module is configured with a third fitting strategy, the third fitting strategy obtains scale values in the electricity utilization type data, and calculates the average number of the scale values corresponding to the industry sub-models to obtain electricity utilization scale difference values of each industry sub-model so as to generate a third fitting array;
the power consumption tag subsystem comprises a distribution deviation module, a concentration deviation module, a fitting deviation module and an industry tag module, wherein the distribution deviation module is configured with industry reference distribution proportion, the distribution deviation module calculates industry actual distribution proportion according to an industry tag of a power consumption user, and obtains industry distribution deviation values according to differences between the industry actual distribution proportion and the industry reference distribution proportion, the concentration deviation module is configured with corresponding concentration standards according to model items in each industry sub-model, the concentration standards comprise concentration ranges and corresponding concentration item values, and the concentration item values corresponding to the concentration ranges in which corresponding data fall in power consumption information are calculated according to the concentration deviation values; the fitting deviation module is configured with a matching scheme screening strategy, the matching scheme screening strategy generates a fitting array according to the first fitting array, the second fitting array and the third fitting array, calculates fitting deviation values of each item in the fitting array through a fitting equation, screens a preset number of matching schemes according to smaller fitting deviation values, and generates industry labels for application electric users, wherein the industry labels correspond to the industry submodels; the industry label module is configured with a preset label determining strategy, the label determining strategy calculates a corresponding industry distribution deviation value and a corresponding concentration deviation value according to each matching scheme, generates a matching deviation value according to the industry distribution deviation value and the concentration deviation value, determines a corresponding matching scheme as a marking scheme according to the size of the matching deviation value, and marks the electricity utilization user according to the marking scheme;
The power consumption tag subsystem comprises a preset first time period, and the power consumption information corresponding to the marked power consumption users is sequentially input into the fitting deviation module for re-marking every the first time period.
Further: the electricity utilization characteristic data comprise voltage, current, active power, reactive power and power factor; the user information data comprises an electricity user address, an electricity user operation range and an electricity user industry type; the electricity type data comprise the voltage grade of the electricity user, the electricity permission scale of the electricity user and the capacity of the electricity station corresponding to the electricity user.
Further: the power consumption characteristic fitting module obtains a power consumption characteristic waveform by matching the reference waveform characteristic from the power consumption characteristic data.
Further: the first fitting strategy includes:
s11, generating a change shape feature according to the power utilization characteristic waveform, wherein the change shape feature reflects the change shape of the power utilization characteristic waveform;
S12, screening waveforms conforming to the change shape characteristics from the historical electric waveforms to be used as first comparison waveforms;
S13, calculating the areas of the envelope areas of the first comparison waveform and the power utilization characteristic waveform, judging that the first comparison waveform is similar to the power utilization characteristic waveform if the areas are larger than a preset similarity threshold, and judging that the first comparison waveform is dissimilar to the power utilization characteristic waveform if the areas are smaller than the preset similarity threshold;
s14, calculating a waveform characteristic fitting value through a first fitting formula:
α=a 1+a2+...+an, where α is the waveform signature fit value and a n is the number of first comparison waveforms similar to the nth power usage signature.
Further: the second fitting strategy includes:
s21, providing an address word stock, an operation range word stock and an electricity industry type word stock;
s22, determining industry characteristic words from the user information data through an address word stock, an operation range word stock and an electricity industry type word stock respectively;
S23, calculating the number of the industrial feature words in each industrial sub-model to generate feature word numbers;
s24, calculating a feature matching value through a second fitting formula:
β=b 1*B1+b1*B2+b3*B3, where β is a feature matching value, B 1 is a preset address corresponding weight parameter, B 1 is a feature word number corresponding to the power consumption user address, B 2 is a preset operation range corresponding weight parameter, and B 2 is a feature word number corresponding to the power consumption user operation range; b 3 and B 3 are the feature word numbers corresponding to the industry types of the electricity users.
Further: the third fitting strategy includes:
s31, calculating an average value of voltage levels of the power utilization users, an average value of power utilization permission scales of the power utilization users and an average value of capacities of power utilization stations corresponding to the power utilization users, wherein the average value of the voltage levels of the power utilization users corresponds to each industry sub-model;
S32, digitizing the voltage level of the electricity user, the electricity use permission scale of the electricity user and the capacity of the electricity use station corresponding to the electricity user:
s33, calculating a power consumption scale difference value according to a third fitting formula:
χ＝|c1-C1|+|c2-C2|+|c3-C3|。
further: the industry distribution deviation value calculation formula is as follows:
Wherein X is an industry distribution deviation value, X is a preset industry deviation coefficient, S w is the total number of actual electricity utilization users, P m% is the standard distribution proportion of the corresponding industry, and S m is the number of electricity utilization users of the corresponding industry.
Further: the calculation formula of the concentration deviation value is y=y 1+y2+...+yn, Y is the concentration deviation value, and Y n is the concentration term value of the nth data in the electricity information of the electricity user.
Further: the fitting deviation value calculation formula is as follows;
Wherein Z is a fitting deviation value, alpha n is a waveform characteristic fitting value corresponding to an nth industry sub-model, beta n is a characteristic matching value corresponding to the nth industry sub-model, χ n is an electricity consumption scale difference value corresponding to the nth industry sub-model, a e is a preset waveform fitting parameter, b e is a characteristic matching parameter, and c e is an electricity consumption scale parameter.
Further: the calculation formula of the matching deviation score is as follows: s=x+y, where S is the match bias score.
The technical effects of the invention are mainly as follows: through the data collection of each dimension of the electricity utilization user, the preliminary matching of the industry label is carried out on the electricity utilization user in a model training mode, the situation that the industry label cannot be matched due to the fact that the data of the electricity utilization user is not full can be avoided, the marking of the industry label is not independent marking logic any more through the two aspects of industry distribution deviation and concentration deviation in a dynamic correction mode, marking precision is improved, marking results of each industry label can influence the industry labels of other electricity utilization users (by affecting industry distribution and concentration), and therefore complexity of an algorithm is improved, but under the influence of larger data quantity, the problem of accuracy can be solved to the greatest extent, accuracy of the industry label is guaranteed, and possibility is provided for the industry automatic label of the electricity utilization user in a large data environment.
Drawings
Fig. 1: the system architecture schematic diagram of the invention;
fig. 2: the invention relates to an industry model subsystem architecture diagram;
Fig. 3: the information fitting subsystem-electric label subsystem architecture diagram of the invention.
Reference numerals: 100. an industry model subsystem; 110. a model configuration module; 120. a model training module; 130. a model generation module; 200. an information fitting subsystem; 210. the power utilization characteristic fitting module; 220. a user information fitting module; 230. a power utilization type fitting module; 300. an electrical tag subsystem; 310. a distribution deviation module; 320. a concentration deviation module; 330. fitting a deviation module; 340. an industry label module; 400. an electricity consumption information database;
Detailed Description
The following detailed description of the invention is provided in connection with the accompanying drawings to facilitate understanding and grasping of the technical scheme of the invention.
An electricity consumption label system based on electricity consumption characteristics is used for marking industry labels for electricity consumption users, and comprises an industry model subsystem 100, an information fitting subsystem 200, an electricity consumption label subsystem 300 and an electricity consumption information database 400;
The electricity consumption database 400 stores the electricity consumption information of the electricity consumption user in real time, wherein the electricity consumption information comprises electricity consumption characteristic data, user information data and electricity consumption type data, the electricity consumption characteristic data reflects the electricity consumption condition of the electricity consumption user, the user information data reflects the registered user information of the electricity consumption user, and the electricity consumption type data reflects the registered electricity consumption type of the electricity consumption user; the electricity utilization characteristic data comprise voltage, current, active power, reactive power and power factor; the user information data comprises an electricity user address, an electricity user operation range and an electricity user industry type; the electricity type data comprise the voltage grade of the electricity user, the electricity permission scale of the electricity user and the capacity of the electricity station corresponding to the electricity user. Firstly, three types of data are described, electricity users can register and change the registration, meanwhile, the electricity consumption and the relation of electric equipment can also change, and the electricity consumption sources mainly comprise the following steps: user profile information, user operation data, user station profile information, user government data and the like, wherein the user profile information comprises data items such as user names, user numbers, metering point information, voltage levels, industry types, electricity utilization addresses and the like; the user operation data comprise voltage, current, active power, reactive power, power factor and the like; the subscriber station file information comprises data items such as subscriber station names, subscriber station numbers, associated metering point information, subscriber station capacity, voltage levels and the like; the user government affair data comprises data items such as enterprise unified credit codes, regular enterprises, small micro enterprises and the like; the data sources are different, but the data is mainly divided into three types of electricity utilization characteristic data and user information data from the data application level and the analysis level, the classification basis of the electricity utilization characteristic data is data which can directly reflect information through numerical change, such as voltage, current, electricity consumption and the like, the data is generally collected, stored and updated in real time, the user information data cannot be directly quantized, such as information of business registration operation range, address and the like, the electricity utilization type data can be quantized by a certain means, the data which can indirectly reflect information, such as scale, corresponding substation capacity and the like, a basis is provided for the subsequent analysis through the collection of the three types of data, and it is required to be explained that in the actual scene at present, the three types of data can be missing and wrong, so that the analysis cannot be implemented.
The industry model subsystem 100 includes a model configuration module 110, a model training module 120, and a model generating module 130, where the model configuration module 110 is configured to configure an industry model architecture, the industry model architecture includes an industry number and an industry category, and as a first link of industry model configuration, the industry architecture is determined first, and is determined according to analysis requirements, not according to any established industry classification standard in the past, but the service content of the company registration during business registration is not necessarily the same, and may be relatively wide, so that it is difficult to label the electricity users, so that the model firstly performs and divides the industry for the current requirements, establishes an architecture of the industry, determines the number of industries needing analysis, if the industry number changes, or needs to refine a certain industry, so that the reconfiguration of the industry model architecture can be completed, and after the configuration is completed, a plurality of empty mathematical containers are obtained.
The model training module 120 processes the corresponding historical electricity samples according to industry types to generate an industry sub-model, the historical electricity samples are the electricity information corresponding to electricity users of the known industry labels, the historical electricity samples are used for inputting information in the containers, the training model memorizes the corresponding characteristics, the industry sub-model can be obtained, the characteristics of the electricity characteristic data corresponding to each industry, the characteristics of the user information data and the characteristics of the electricity type data can be judged through the industry sub-model, and it is required to be stated that although two identical electricity information can exist in the samples, the two information are divided into different industry sub-models, but from a macroscopic view, a cluster effect can be generated, namely, the identical characteristics can be more, the difference is less, the model can be obtained through the filling mode of the historical electricity samples, more and more available samples can be generated along with the increase of the judgment quantity, more accurate judgment results can be obtained, but the two information can still have great influence on the characteristics of the industry according to the important characteristics, and the invention can still have great difficulty in the aspect of accurately achieving the characteristics of the invention, even if the actual characteristics of the industry are different, and the problem is solved, the problem that the actual characteristics of the invention has great difficulty is solved, and the difference of the characteristics of the industry is difficult to be greatly converged.
The model generation module 130 generates an industry feature model according to an industry sub-model and an industry model architecture; in the last step, the industry sub-model is obtained, and the model architecture is completed, so that the characteristic model of the industry is completed, and the model architecture comprises a relation among industries for converging the whole characteristic model and a quantitative proportion relation among different industries, wherein the information belongs to a preset value of a system, and the approximate proportion relation can be obtained according to the total production volume.
The information fitting subsystem 200 includes a power usage feature fitting module 210, a user information fitting module 220, a power usage type fitting module 230,
The electricity consumption characteristic fitting module 210 is configured with a first fitting strategy, the first fitting strategy obtains electricity consumption characteristic waveforms in the electricity consumption characteristic data, and obtains waveform characteristic fitting values corresponding to each industry sub-model according to calculation according to whether similar electricity consumption characteristic waveforms exist in historical electricity consumption waveforms corresponding to the industry sub-models to generate a first fitting array; also included is a power usage feature database storing a number of reference waveform features, the power usage feature fitting module 210 obtaining a power usage feature waveform by matching the reference waveform features from the power usage feature data. The first fitting strategy includes:
s11, generating a change shape feature according to the power utilization characteristic waveform, wherein the change shape feature reflects the change shape of the power utilization characteristic waveform;
S12, screening waveforms conforming to the change shape characteristics from the historical electric waveforms to be used as first comparison waveforms;
S13, calculating the areas of the envelope areas of the first comparison waveform and the power utilization characteristic waveform, judging that the first comparison waveform is similar to the power utilization characteristic waveform if the areas are larger than a preset similarity threshold, and judging that the first comparison waveform is dissimilar to the power utilization characteristic waveform if the areas are smaller than the preset similarity threshold;
s14, calculating a waveform characteristic fitting value through a first fitting formula:
α=a 1+a2+...+an, where α is the waveform signature fit value and a n is the number of first comparison waveforms similar to the nth power usage signature. Firstly, fitting electricity utilization characteristics, wherein each industry sub-model stores electricity utilization waveforms of each industry, for example, an unknown user needs to be marked at present, and judges which industry the electricity utilization waveform accords with, then the characteristics of the electricity utilization waveform actually generated by the user need to be extracted, and then whether similar characteristics exist in models of corresponding industries or not is judged in a similarity mode, the more the similar characteristics of the electricity utilization are, the more the electricity utilization user accords with the characteristics of the industry, for example, the general electricity utilization of the industry A has larger change, the one-time electricity utilization time is shorter, the electricity utilization of the industry B has larger electricity utilization in a specific two-three time period, the electricity utilization of the industry C follows the stepped rising of the electricity utilization quantity, then the stepped lowering process is completed, the collection and judgment of electricity utilization information can be completed through the above contents, if the characteristics exist in the electricity utilization user to be compared, the description possibility is higher, so that the corresponding fitting relation of the electricity utilization user on the electricity utilization waveform can be judged through calculation.
The user information fitting module is configured with a second fitting strategy, the second fitting strategy obtains industry feature words in the user information data, and corresponding feature matching values in each industry sub-model are obtained through calculation according to whether the same industry feature words exist in the industry sub-model or not so as to generate a second fitting array; the second fitting strategy includes:
s21, providing an address word stock, an operation range word stock and an electricity industry type word stock;
s22, determining industry characteristic words from the user information data through an address word stock, an operation range word stock and an electricity industry type word stock respectively;
S23, calculating the number of the industrial feature words in each industrial sub-model to generate feature word numbers;
s24, calculating a feature matching value through a second fitting formula:
β=b 1*B1+b1*B2+b3*B3, where β is a feature matching value, B 1 is a preset address corresponding weight parameter, B 1 is a feature word number corresponding to the power consumption user address, B 2 is a preset operation range corresponding weight parameter, and B 2 is a feature word number corresponding to the power consumption user operation range; b 3 and B 3 are the feature word numbers corresponding to the industry types of the electricity users. The key part of the second fitting strategy is that the information is unquantified, so that the comparison is required by a keyword recognition mode, the more the corresponding keywords are identical, the description is about likely to be in line with the industry, for example, a more enterprises of the A industry have keywords A1 in the operating range, have A2 province A21 market in the address, and have A3 in the electricity industry type, if the corresponding keywords exist in the user level, the electricity user is likely to belong to the industry, so that the corresponding range can be judged through the formula, the keywords in the electricity user information are completely recognized, so that whether the keywords belong to the industry is judged, the judgment can often obtain unexpected results, for example, the comparison tendency of the industry is in the operating range, the A1 and the B1 are registered at the same time, the B1 and the C1 are registered in the other industry, the two industries can be distinguished, the information is very difficult to obtain by human experience, but the problem can be well solved through big data analysis, whether the keywords are identical and the same, the quantity can be used as the judgment, and finally, the numerical value can be obtained through the characteristic.
The electricity utilization type fitting module is configured with a third fitting strategy, the third fitting strategy obtains scale values in the electricity utilization type data, and calculates the average number of the scale values corresponding to the industry sub-models to obtain electricity utilization scale difference values of each industry sub-model so as to generate a third fitting array; the third fitting strategy includes:
s31, calculating an average value of voltage levels of the power utilization users, an average value of power utilization permission scales of the power utilization users and an average value of capacities of power utilization stations corresponding to the power utilization users, wherein the average value of the voltage levels of the power utilization users corresponds to each industry sub-model;
S32, digitizing the voltage level of the electricity user, the electricity use permission scale of the electricity user and the capacity of the electricity use station corresponding to the electricity user:
s33, calculating a power consumption scale difference value according to a third fitting formula:
χ= |c 1-C1|+|c2-C2|+|c3-C3 |. And the electricity consumption type fitting strategy is used for fitting according to the scale value in the electricity consumption type data, and the corresponding result can be obtained by calculating the deviation of the value of the electricity consumption type fitting strategy and the average value corresponding to the model according to the factors such as the voltage level, the allowable scale, the capacity and the like of the grade, and after the electricity consumption scale difference value is obtained, the matching degree of the user in the quantifiable data can be reflected indirectly.
The matching numerical values of a user and all industry sub-models under three dimensions can be obtained through the models, and a mathematical array is obtained. Although the optimal result can be obtained through operation, it has been mentioned in the foregoing that if the matching degree judgment is directly performed, the result may be centralized and worse, so that the whole model cannot be converged, because the industry distribution rule is fuzzy, and only the matching degree judgment can cause more and more electricity users to be divided into the same industry for marking, so that the division is wrong, in order to solve the problem, we first put forward two assumptions, the first and industry distributions can be measured and predicted, and the deviation from the measurement prediction is not greatly exceeded; secondly, the industry will follow a feature concentration principle, that is, the higher the industry concentration, the more power users which are not in line with the industry features are likely to not belong to the industry, and based on the two assumptions, we propose the following algorithm to enable the model to converge so as to obtain more accurate dividing results.
The powered tagging subsystem 300 includes a distribution bias module 310, a concentration bias module 320, a fitting bias module 330 and an industry tagging module 340,
The distribution deviation module 310 is configured with an industry reference distribution ratio, the distribution deviation module 310 calculates an industry actual distribution ratio according to an industry label of an electricity user, and obtains an industry distribution deviation value according to a difference value between the industry actual distribution ratio and the industry reference distribution ratio, and the industry distribution deviation value has a calculation formula as follows:
Wherein X is an industry distribution deviation value, X is a preset industry deviation coefficient, S w is the total number of actual electricity utilization users, P m% is the standard distribution proportion of the corresponding industry, and S m is the number of electricity utilization users of the corresponding industry. Firstly, calculating industry deviation distribution, if the number of users in the A industry is judged to be 2% of the number of users in all industries in advance, if the number of users in the A industry is 2% at present, the larger the deviation is, so that through deviation calculation, if the deviation of the A industry is large enough, the electricity users which do not accord with the A industry model tend to be divided into other industries.
The concentration deviation module 320 is configured with a corresponding concentration standard according to model items in each industry sub-model, wherein the concentration standard comprises a concentration range and a corresponding concentration item value, and calculates a concentration deviation value according to the concentration item value corresponding to the concentration range in which corresponding data falls in the electricity consumption information; the calculation formula of the concentration deviation value is y=y 1+y2+...+yn, Y is the concentration deviation value, and Y n is the concentration term value of the nth data in the electricity information of the electricity user. On the other hand, by the concentration statistics, for example, if the a industry registers the electricity users with 90% of the business scope A1, the possibility that the A1 is not registered as the business scope, is reduced, so the concentration scope is defined according to the concentration standard, and then the deviation is recalculated according to the concentration scope, it is to be noted that if the concentration of one industry accords with the normal distribution, the concentration term value corresponding to the concentration scope of the edge is relatively high, if the term falls into the edge is also large, and if the term falls into the industry is uniformly distributed, no influence is exerted on the deviation value in each scope, the term is 0, so the determination of the concentration scope is judged according to the concentration of the known model, and the final concentration deviation value is obtained.
The fitting deviation module 3230 is configured with a matching scheme screening strategy, the matching scheme screening strategy generates a fitting array according to the first fitting array, the second fitting array and the third fitting array, calculates fitting deviation values of each item in the fitting array through a fitting equation, screens a preset number of matching schemes according to smaller fitting deviation values, and generates industry labels for application electric users, wherein the industry labels correspond to industry submodels; the fitting deviation value calculation formula is as follows;
Wherein Z is a fitting deviation value, alpha n is a waveform characteristic fitting value corresponding to an nth industry sub-model, beta n is a characteristic matching value corresponding to the nth industry sub-model, χ n is an electricity consumption scale difference value corresponding to the nth industry sub-model, a e is a preset waveform fitting parameter, b e is a characteristic matching parameter, and c e is an electricity consumption scale parameter. In the scheme screening strategy, a plurality of preferable schemes can be obtained by calculating the fitting deviation value, namely, if the matching degree is higher, the fitting deviation value is larger, and because the fitting deviation value relates to the optimal solution, the schemes are judged by two standards of industry distribution and concentration, so that the judging result can be ensured to be still in the range of alternative schemes, and the dividing result can be ensured under the training of a large number of users through adjustment.
The industry label module 340 is configured with a preset label determining policy, the label determining policy calculates a corresponding industry distribution deviation value and a corresponding concentration deviation value according to each matching scheme, generates a matching deviation value according to the industry distribution deviation value and the concentration deviation value, determines the corresponding matching scheme as a marking scheme according to the size of the matching deviation value, and marks the electricity consumer according to the marking scheme; the calculation formula of the matching deviation score is as follows: s=x+y, where S is the match bias score. By setting the matching bias score, the training model can be closed-loop.
The power consumption tag subsystem comprises a preset first time period, and the power consumption information corresponding to the marked power consumption users is sequentially input into the fitting deviation module for re-marking every the first time period. In this way, the marked data can be re-marked, so that the data tends to be reliable, and the model can be continuously optimized, thereby improving the reliability.
Of course, the above is only a typical example of the invention, and other embodiments of the invention are also possible, and all technical solutions formed by equivalent substitution or equivalent transformation fall within the scope of the invention claimed.
Claims (10)
1. The utility model provides a power consumption label system based on power consumption characteristic for to power consumption user mark trade label, its characterized in that: the system comprises an industry model subsystem, an information fitting subsystem, an electricity consumption tag subsystem and an electricity consumption information database;
the electricity consumption information database stores electricity consumption information of electricity consumption users in real time, wherein the electricity consumption information comprises electricity consumption characteristic data, user information data and electricity consumption type data, the electricity consumption characteristic data reflects electricity consumption conditions of the electricity consumption users, the user information data reflects registered user information of the electricity consumption users, and the electricity consumption type data reflects electricity consumption types registered by the electricity consumption users;
The industry model subsystem comprises a model configuration module, a model training module and a model generation module, wherein the model configuration module is used for configuring an industry model framework, the industry model framework comprises industry quantity and industry types, the model training module processes corresponding historical electricity utilization samples according to the industry types to generate an industry sub-model, the historical electricity utilization samples are electricity utilization information corresponding to electricity utilization users of known industry labels, and the model generation module generates an industry characteristic model according to the industry sub-model and the industry model framework;
The information fitting subsystem comprises an electricity utilization characteristic fitting module, a user information fitting module and an electricity utilization type fitting module, wherein the electricity utilization characteristic fitting module is configured with a first fitting strategy, the first fitting strategy obtains electricity utilization characteristic waveforms in electricity utilization characteristic data, and waveform characteristic fitting values corresponding to each industry sub-model are obtained according to calculation according to the fact that whether similar electricity utilization characteristic waveforms exist in historical electricity utilization characteristic waveforms corresponding to the industry sub-model or not, so as to generate a first fitting array; the user information fitting module is configured with a second fitting strategy, the second fitting strategy obtains industry feature words in the user information data, and corresponding feature matching values in each industry sub-model are obtained through calculation according to whether the same industry feature words exist in the industry sub-model or not so as to generate a second fitting array; the electricity utilization type fitting module is configured with a third fitting strategy, the third fitting strategy obtains scale values in the electricity utilization type data, and calculates the average number of the scale values corresponding to the industry sub-models to obtain electricity utilization scale difference values of each industry sub-model so as to generate a third fitting array;
the power consumption tag subsystem comprises a distribution deviation module, a concentration deviation module, a fitting deviation module and an industry tag module, wherein the distribution deviation module is configured with industry reference distribution proportion, the distribution deviation module calculates industry actual distribution proportion according to an industry tag of a power consumption user, and obtains industry distribution deviation values according to differences between the industry actual distribution proportion and the industry reference distribution proportion, the concentration deviation module is configured with corresponding concentration standards according to model items in each industry sub-model, the concentration standards comprise concentration ranges and corresponding concentration item values, and the concentration item values corresponding to the concentration ranges in which corresponding data fall in power consumption information are calculated according to the concentration deviation values; the fitting deviation module is configured with a matching scheme screening strategy, the matching scheme screening strategy generates a fitting array according to the first fitting array, the second fitting array and the third fitting array, calculates fitting deviation values of each item in the fitting array through a fitting equation, screens a preset number of matching schemes according to smaller fitting deviation values, and generates industry labels for application electric users, wherein the industry labels correspond to the industry submodels; the industry label module is configured with a preset label determining strategy, the label determining strategy calculates a corresponding industry distribution deviation value and a corresponding concentration deviation value according to each matching scheme, generates a matching deviation value according to the industry distribution deviation value and the concentration deviation value, determines a corresponding matching scheme as a marking scheme according to the size of the matching deviation value, and marks the electricity utilization user according to the marking scheme;
The power consumption tag subsystem comprises a preset first time period, and the power consumption information corresponding to the marked power consumption users is sequentially input into the fitting deviation module for re-marking every the first time period.
2. A power usage tagging system based on power usage characteristics as recited in claim 1, wherein: the electricity utilization characteristic data comprise voltage, current, active power, reactive power and power factor; the user information data comprises an electricity user address, an electricity user operation range and an electricity user industry type; the electricity type data comprise the voltage grade of the electricity user, the electricity permission scale of the electricity user and the capacity of the electricity station corresponding to the electricity user.
3. A power usage tagging system based on power usage characteristics as recited in claim 1, wherein: the power consumption characteristic fitting module obtains a power consumption characteristic waveform by matching the reference waveform characteristic from the power consumption characteristic data.
4. A power usage tagging system based on power usage characteristics as recited in claim 2, wherein: the first fitting strategy includes:
s11, generating a change shape feature according to the power utilization characteristic waveform, wherein the change shape feature reflects the change shape of the power utilization characteristic waveform;
S12, screening waveforms conforming to the change shape characteristics from the historical electric waveforms to be used as first comparison waveforms;
S13, calculating the areas of the envelope areas of the first comparison waveform and the power utilization characteristic waveform, judging that the first comparison waveform is similar to the power utilization characteristic waveform if the areas are larger than a preset similarity threshold, and judging that the first comparison waveform is dissimilar to the power utilization characteristic waveform if the areas are smaller than the preset similarity threshold;
s14, calculating a waveform characteristic fitting value through a first fitting formula:
α=a 1+a2+...+an, where α is the waveform signature fit value and a n is the number of first comparison waveforms similar to the nth power usage signature.
5. A power usage tagging system based on power usage characteristics as recited in claim 4, wherein: the second fitting strategy includes:
s21, providing an address word stock, an operation range word stock and an electricity industry type word stock;
s22, determining industry characteristic words from the user information data through an address word stock, an operation range word stock and an electricity industry type word stock respectively;
S23, calculating the number of the industrial feature words in each industrial sub-model to generate feature word numbers;
s24, calculating a feature matching value through a second fitting formula:
β=b 1*B1+b1*B2+b3*B3, where β is a feature matching value, B 1 is a preset address corresponding weight parameter, B 1 is a feature word number corresponding to the power consumption user address, B 2 is a preset operation range corresponding weight parameter, and B 2 is a feature word number corresponding to the power consumption user operation range; b 3 and B 3 are the feature word numbers corresponding to the industry types of the electricity users.
6. A power usage tagging system based on power usage characteristics as recited in claim 4, wherein: the third fitting strategy includes:
s31, calculating an average value of voltage levels of the power utilization users, an average value of power utilization permission scales of the power utilization users and an average value of capacities of power utilization stations corresponding to the power utilization users, wherein the average value of the voltage levels of the power utilization users corresponds to each industry sub-model;
S32, digitizing the voltage level of the electricity user, the electricity use permission scale of the electricity user and the capacity of the electricity use station corresponding to the electricity user:
s33, calculating a power consumption scale difference value according to a third fitting formula:
χ＝|c1-C1|+|c2-C2|+|c3-C3|。
7. A power usage tagging system based on power usage characteristics as recited in claim 6, wherein: the industry distribution deviation value calculation formula is as follows:
Wherein X is an industry distribution deviation value, X is a preset industry deviation coefficient, S w is the total number of actual electricity utilization users, P m% is the standard distribution proportion of the corresponding industry, and S m is the number of electricity utilization users of the corresponding industry.
8. A power usage tagging system based on power usage characteristics as recited in claim 7, wherein: the calculation formula of the concentration deviation value is y=y 1+y2+...+yn, Y is the concentration deviation value, and Y n is the concentration term value of the nth data in the electricity information of the electricity user.
9. A power usage tagging system based on power usage characteristics as recited in claim 8, wherein: the fitting deviation value calculation formula is as follows;
Wherein Z is a fitting deviation value, alpha n is a waveform characteristic fitting value corresponding to an nth industry sub-model, beta n is a characteristic matching value corresponding to the nth industry sub-model, χ n is an electricity consumption scale difference value corresponding to the nth industry sub-model, a e is a preset waveform fitting parameter, b e is a characteristic matching parameter, and c e is an electricity consumption scale parameter.
10. A power usage tagging system based on power usage characteristics as recited in claim 9, wherein: the calculation formula of the matching deviation score is as follows: s=x+y, where S is the match bias score.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111327202.3A CN114037282B (en) | 2021-11-10 | 2021-11-10 | Electricity consumption label system based on electricity consumption characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111327202.3A CN114037282B (en) | 2021-11-10 | 2021-11-10 | Electricity consumption label system based on electricity consumption characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114037282A CN114037282A (en) | 2022-02-11 |
| CN114037282B true CN114037282B (en) | 2024-06-21 |
Family
ID=80137108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111327202.3A Active CN114037282B (en) | 2021-11-10 | 2021-11-10 | Electricity consumption label system based on electricity consumption characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114037282B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113487448A (en) * | 2021-05-31 | 2021-10-08 | 国网上海市电力公司 | Power credit labeling method and system based on power big data |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006172027A (en) * | 2004-12-15 | 2006-06-29 | Hitachi Ltd | Rating system using financial index |
| CN111552923B (en) * | 2020-04-22 | 2021-05-04 | 国网江苏省电力有限公司苏州供电分公司 | Load prediction method and load prediction system based on universal distribution |
-
2021
- 2021-11-10 CN CN202111327202.3A patent/CN114037282B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113487448A (en) * | 2021-05-31 | 2021-10-08 | 国网上海市电力公司 | Power credit labeling method and system based on power big data |
Non-Patent Citations (1)
| Title |
|---|
| 电力用户行为画像构建技术研究;傅军;许鑫;罗迪;朱天博;刘霞;;电气应用;20180705(第13期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114037282A (en) | 2022-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106570778B (en) | A kind of method that data integration based on big data is calculated with line loss analyzing | |
| CN111369070B (en) | Multimode fusion photovoltaic power prediction method based on envelope clustering | |
| CN112199367B (en) | Long-time scale power balance check and big data analysis method and system | |
| CN108932557A (en) | A kind of Short-term Load Forecasting Model based on temperature cumulative effect and grey relational grade | |
| CN106682763B (en) | Power load optimization prediction method for large amount of sample data | |
| CN116031888B (en) | Dynamic load prediction-based power flow optimization method, system and storage medium | |
| CN110378510B (en) | Distribution network material demand prediction method based on time sequence and hierarchical aggregation | |
| CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN111553568A (en) | Line loss management method based on data mining technology | |
| CN115358437A (en) | Power supply load prediction method based on convolutional neural network | |
| Velasco et al. | Day-ahead base, intermediate, and peak load forecasting using k-means and artificial neural networks | |
| CN115860797A (en) | Electric quantity demand prediction method suitable for new electricity price reform situation | |
| CN110909786A (en) | New user load identification method based on characteristic index and decision tree model | |
| CN114037282B (en) | Electricity consumption label system based on electricity consumption characteristics | |
| CN111967684B (en) | Metering asset active distribution method based on big data analysis | |
| CN109858667A (en) | It is a kind of based on thunder and lightning weather to the short term clustering method of loading effects | |
| CN109583773A (en) | A kind of method, system and relevant apparatus that taxpaying credit integral is determining | |
| CN112508254A (en) | Method for determining investment prediction data of transformer substation engineering project | |
| CN112330030A (en) | System and method for predicting requirements of expansion materials | |
| CN116432437A (en) | Comprehensive evaluation method and system for photovoltaic power generation unit | |
| CN114372835B (en) | Comprehensive energy service potential customer identification method, system and computer equipment | |
| CN114676931B (en) | Electric quantity prediction system based on data center technology | |
| CN116799788A (en) | Peak clipping and valley filling electricity utilization method based on big data analysis | |
| Yanchao et al. | A proportion prediction model of terminal energy structure of IPS based on hidden Markov chain | |
| CN113077108A (en) | Data prediction system for power material configuration requirements |
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 |