CN114358359A - Electric charge recovery risk early warning method based on electric power market development situation perception - Google Patents
Electric charge recovery risk early warning method based on electric power market development situation perception Download PDFInfo
- Publication number
- CN114358359A CN114358359A CN202111120289.7A CN202111120289A CN114358359A CN 114358359 A CN114358359 A CN 114358359A CN 202111120289 A CN202111120289 A CN 202111120289A CN 114358359 A CN114358359 A CN 114358359A
- Authority
- CN
- China
- Prior art keywords
- user
- power
- trend
- industry
- electric charge
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an electric charge recovery risk early warning method based on electric power market development situation perception. According to the method, the power market development situation perception result of the power consumer is used as an influence factor to be added into the power charge recovery risk early warning modeling, so that the power charge recovery risk is not only influenced by microscopic behaviors such as user payment behaviors and business expansion changes, but also the macroscopic analysis result of the power market development situation of the user is combined, a multi-dimensional and multi-angle user power charge recovery risk early warning scheme is established, the future power charge risk of the user is identified more comprehensively, and data support is provided for an electric power company to actively deal with the marketing risk in the power market development process, to set the power charge recovery scheme in advance, to prevent and reduce the operation loss of the company.
Description
Technical Field
The invention relates to the technical field of electric power, in particular to an electric charge recovery risk early warning method based on electric power market development situation perception.
Background
The electric power market development situation awareness is an important basic work of an electric power company, accurate situation awareness and prediction can effectively support early warning and prevention of electric charge recovery risks, and the electric power market development situation awareness has very important significance for the electric power company to reasonably deal with the electric charge recovery risks and reduce operation loss. The electric charge is the main source of income of the electric power company, and the electric charge recovery rate is closely related to the normal production operation condition of the electric power company. Because the current external economic operation condition is complicated and complicated, most enterprises face transformation pressure, and in addition, a large number of users adopt an electricity fee payment mode of 'firstly using electricity and then paying electricity', so that the electricity fee overstock is easily caused, the effective development of electricity fee recovery work is not facilitated, and the normal production operation of a power company is greatly influenced. Therefore, based on the perception of the development situation of the power market, the industry development situation is mastered in advance, and the identification of the power utilization customers with the risk of recovering the power fee in the industry is an important guarantee for the company to actively deal with the marketing risk in the development process of the power market, prevent the marketing risk and reduce the operation loss.
The power demand change is a 'weather meter' and a 'wind vane' of social and economic operation, and the technology of big data, artificial intelligence and the like is used for controlling the power market development situation and identifying the marketing risk of the power company, so that the method has important significance for objectively knowing the national economic development situation and situation, reasonably making an operation plan to balance the power demands of all parties and preventing and reducing the production and operation loss of the power company. Due to the comprehensive influence of factors such as weather, economy, holidays, relevant policies and the like, the analysis difficulty of the development situation of the power market of the company is high, in addition, in recent years, the economic operation enters a new normal state, the economic transformation and upgrading, the environmental protection strength is increased, and the influence of the superposition of emergencies such as new crown epidemic situations and the like is caused, so that power consumers face the problems of yield reduction, difficult sales and the like, the identification of marketing risks such as untimely recovery of electric charges becomes more difficult, the operation loss of the power company is caused, and the normal production operation of a power grid enterprise is seriously influenced.
At present, the power market development situation perception of a company mainly depends on the experience of marketing business experts, analysis is carried out based on recent power selling amount and the change of the cocycle ratio of the acceleration of the power selling amount, the analysis conclusion is strong in subjectivity and difficult to quantify, and the analysis work mainly adopts a manual means, so that the work period is long, and a large amount of manpower is consumed; the electric charge recycling risk identification work mainly depends on expert experience to judge through setting simple rules for behavior information such as historical electricity utilization, payment and installation of users, technical means and analysis methods are simple, influences of development situations of the electric power market of the downstream industry of the current economic situation on production and operation of electricity utilization customers are not considered, and analysis conclusions are limited. Therefore, it is necessary to comprehensively consider the influence of internal and external factors on the development of the power market, study how to construct an analysis method for the development situation of the power market by using a big data technology, realize accurate analysis and timely perception of the development situation of the power market of a company, construct an automatic identification method for the risk of power charge recovery according to the influence of the development situation on the production and operation of a user, and combine the power consumption situation of the user, power charge payment, business expansion and installation and other behaviors to actively cope with the risk.
The conventional electric charge recovery risk identification and early warning work mainly comprises the following methods:
(1) experience screening method
The experience screening method generally sets certain judgment conditions according to experience for historical electricity consumption such as user electricity charge, electricity consumption safety, arrearage, default fund, power failure and the like and electricity charge payment conditions, for example, the times of electricity failure caused by arrearage, the times of electricity charge default fund generation in nearly 6 months, the electricity consumption increase conditions in nearly 3 months and the like occur in nearly 1 year, and screens out electricity consumption customer lists with different levels of electricity charge recycling risks according to different judgment conditions, and pays attention to the integrity and timeliness of future electricity charge payment.
(2) Internal data analysis method
The internal data analysis method generally establishes a model through a big data mining method and identifies the user electric charge recycling risk through data such as user industry category, operation capacity, contract capacity, electricity consumption, electric charge amount, real charge amount, release date, real charge date, default money and the like acquired by each information acquisition system in the power industry: firstly preprocessing data, then constructing input parameters more suitable for mathematical modeling according to actual business requirements, for example, calculating the electricity fee payment period of a user through the electricity fee release date and the real charge date, and then inputting the parameters into data mining algorithms such as classification and regression to identify and quantify the electricity fee recovery risk of the user, so as to provide data support for electricity fee recovery work and early warn the existing risk in advance.
The existing electric charge recovery risk early warning method mainly has the following problems: 1. setting judgment conditions for historical electricity consumption and electricity fee payment conditions of users according to business experience, screening an electricity fee recycling attention user list, and having a subjective judgment mode and lacking a quantitative method; 2. the internal data is used for establishing a model to identify the electric charge recycling risk, the influence of the electric power market development situation on the production and operation situation of the electricity consumer under the current economic situation is ignored, and the identification of the electric charge recycling risk of the user is insufficient.
Disclosure of Invention
The invention provides an electric charge recovery risk early warning method based on electric power market development situation perception, and aims to solve the problems of how to perceive and predict an electric power market development situation and how to identify a user electric charge recovery risk based on the electric power market development situation.
In a first aspect, the invention provides an electric charge recovery risk early warning method based on power market development situation awareness, which includes:
extracting trend characteristics of historical electricity sales development of the industry by using a seasonal adjustment algorithm;
analyzing the influence of each economic index on the fluctuation of the trend characteristics, and determining the economic index which needs to be considered when predicting the development trend of the power selling amount of each industry;
according to the economic indexes, a power market development situation prediction model is built through a regression algorithm, and the power market development situation of the user is perceived;
according to the perception result of the power market development situation of the user, the factors of the user power charge recycling risk to be considered are identified by combining the user payment behavior data of the user and the user business expansion, capacity increase, capacity reduction, account cancellation, suspension and capacity recovery data, wherein the user payment behavior data comprises: monthly electric charge amount, payment amount, arrearage amount, payment time, payment frequency and user payment behavior data under the condition of late payment;
and according to the factors needing to be considered for the user electric charge recovery risk, constructing a user electric charge recovery risk early warning model through a classification regression algorithm so as to quantify the probability of the risk of the user electric charge recovery and identify the users with the electric charge recovery risk.
Further, by utilizing a seasonal adjustment algorithm, extracting trend characteristics of industry historical electricity sales development, including:
the power selling amount is decomposed into three subsequences of trend item of power selling amount, season item of power selling amount and random item of power selling amount by using Bayes season regulation algorithm:
Q(i)＝Qt(i)+Qs(i)+Qr(i)；
wherein { Q (i) | i ∈ 1, 2.. and n } represents industry historical electricity selling amount data, and { Q | i ∈ 1, 2.. and n } represents industry historical electricity selling amount datat(i) I belongs to 1,2, the right, n is a trend item of electricity selling amount, { Q ∈ is sets(i) I belongs to 1,2, the right, n is a season item of electricity sale, { Q ∈ Qr(i) And | i belongs to 1,2,. and n } is a random term of electricity selling quantity.
Further, analyzing the influence of each economic index on the fluctuation of the trend characteristics, and determining the economic index which needs to be considered when predicting the development trend of the electricity sold in each industry, wherein the economic index comprises the following steps:
carrying out correlation analysis by using the power selling trend items of each industry and each economic index to obtain influence factors needing to be considered when predicting the power selling development trend of each industry, wherein the influence factors are instantaneity indexes and precursor indexes, the instantaneity indexes, namely the current influence factors, influence the current power selling trend development, and the precursor indexes, namely the current influence factors, influence the future power selling trend development;
and analyzing the leading relation between the leading indicators and the development of the power selling trend by adopting a dynamic time warping algorithm.
Further, the instantaneity index comprises everyone dominance income, second industry GDP acceleration rate, third industry GDP acceleration rate and industry added value acceleration rate; the leading indicators include industry net amplification capacity, manufacturing PMI, and non-manufacturing PMI.
Further, the factors that need to be considered for the user electric charge recycling risk are divided into: the system comprises model output factors, data statistics factors and data construction factors, wherein the model output factors comprise the development trend of an electric power market where a user is located, the data statistics factors comprise monthly payment frequency, historical generation times of late deposit, last generation time of late deposit and electric charge amount, and the data construction factors comprise payment duration and operation capacity change.
The invention has the following beneficial effects: according to the electric charge recovery risk early warning method based on electric power market development situation perception, an electric power market development situation perception result of an electricity consumer is used as an influence factor to be added into electric charge recovery risk early warning modeling, so that electric charge recovery risks are not influenced by microscopic behaviors such as user payment behaviors and business expansion changes, macroscopic analysis results of electric power market development situations of the user are combined, a multi-dimensional and multi-angle user electric charge recovery risk early warning scheme is established, future electric charge risks of the user are identified more comprehensively, and data support is provided for an electric power company to actively deal with marketing risks in the electric power market development process, make the electric charge recovery scheme in advance, and prevent and reduce operation loss of the company.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without any inventive exercise.
Fig. 1 is a flowchart of an electric charge recycling risk early warning method based on electric power market development situation awareness according to an embodiment of the present invention.
Fig. 2 is a schematic representation of the hard-epsilon band hyperplane.
FIG. 3 is a set of class-positive points D+And negative class point set D-Schematic diagram of the classification surface of (1).
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention. The technical solutions provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
Referring to fig. 1, the present invention provides an electric charge recycling risk early warning method based on power market development situation awareness, the method comprising:
and S101, extracting trend characteristics of historical industry electricity sales development by using a season adjustment algorithm.
Specifically, the industry historical monthly electricity sales amount is influenced by economic development, seasonal variation, temperature, holidays and other factors in a superposition mode, the development situation of the power market needs to be analyzed through the industry electricity sales amount, and the trend characteristics of the electricity sales amount influenced by the economic development need to be extracted. Extracting trend characteristics of industry historical electricity sales development by using a seasonal adjustment algorithm, wherein the trend characteristics comprise the following steps:
the power selling amount is decomposed into three subsequences of trend item of power selling amount, season item of power selling amount and random item of power selling amount by using Bayes season regulation algorithm:
Q(i)＝Qt(i)+Qs(i)+Qr(i)；
wherein { Q (i) | i ∈ 1, 2.. and n } represents industry historical electricity selling amount data, and { Q | i ∈ 1, 2.. and n } represents industry historical electricity selling amount datat(i) I belongs to 1,2, the right, n is a trend item of electricity selling amount, { Q ∈ is sets(i) I belongs to 1,2, the right, n is a season item of electricity sale, { Q ∈ Qr(i) And | i belongs to 1,2,. and n } is a random term of electricity selling quantity.
And S102, analyzing the influence of each economic index on the fluctuation of the trend characteristics, and determining the economic index which needs to be considered when the trend of the power selling development of each industry is predicted.
Specifically, analyzing the influence of each economic indicator on the fluctuation of the trend characteristics, and determining the economic indicators to be considered when predicting the development trend of the electricity sales in each industry, includes: carrying out correlation analysis by using the power selling trend items of each industry and each economic index to obtain influence factors needing to be considered when predicting the power selling development trend of each industry, wherein the influence factors are instantaneity indexes and precursor indexes, the instantaneity indexes, namely the current influence factors, influence the current power selling trend development, and the precursor indexes, namely the current influence factors, influence the future power selling trend development; and analyzing the leading relation between the leading indicators and the development of the power selling trend by adopting a dynamic time warping algorithm. Further, the instantaneity index comprises everyone dominance income, second industry GDP acceleration rate, third industry GDP acceleration rate and industry added value acceleration rate; the leading indicators include industry net amplification capacity, manufacturing PMI, and non-manufacturing PMI. The summary result of the instantaneity index and the leading index considered by the power sale development trend prediction is shown in table 1:
TABLE 1 sales electricity development trend prediction modeling consideration
The electric quantity selling trend item before the industry expansion net capacity trend item is calculated for 3 months, the electric quantity selling trend item before the manufacturing PMI is calculated for 3 months, and the electric quantity selling trend item before the non-manufacturing PMI is calculated for 6 months.
And S103, constructing a power market development situation prediction model through a regression algorithm according to the economic indexes, and perceiving the power market development situation of the user.
The monthly electricity selling trend item is mainly influenced by economic factors, and the steps for predicting the trend items of various industries are as follows:
the data sequence of the trend item of the electricity sales in the first step is { Qt(i)|i∈1,2,...,N}。
The second step obtains the corresponding real-time factor of the industry from table 1, and records as { E (i) | i ∈ 1,21(i)，E2(i) I ∈ 1, 2.., N }, such as an industrial instantaneity index E1(i) Showing the GDP acceleration of the second industry, E2(i) Indicating an increased value and speed of the industry.
The third step is taken from Table 1Obtaining a leader factor corresponding to the industry, and marking as { G (i) | i ∈ 1, 2.., N }, wherein if the leader factor includes two leader indexes of net amplification capacity and PMI, G (i) ═ G |, (i) ·1(i)，G2(i) I ∈ 1, 2., N }, as predicted by the industry trend term:
G1(i)＝{net(i-1)，net(i-2)，...，net(i-3)}；
G2(i)＝{PMI(i-1)，PMI(i-2)，PMI(i-3)}；
net (i) is the net capacity increase trend item of month i, PMI (i) is the PMI of month i, and other industry trend items are predicted in the same way.
And fourthly, performing index decorrelation. And obtaining a decorrelation result of the instantaneity index and the precursor index by adopting a Principal Component Analysis (PCA) algorithm.
P(i)＝pca(E(i)，G(i))，i＝1，2，...，N；
And fifthly, establishing a trend item prediction model, and predicting by adopting an SVM algorithm. Establishing a trend item prediction model as follows:
wherein
And (4) obtaining an electric power market development situation perception result which is an industry electric power selling trend prediction value in the ith month.
Step S104, according to the perception result of the power market development situation of the user, identifying the factors to be considered for the user power charge recycling risk by combining the user payment behavior data of the user and the user business expansion, capacity increase, capacity reduction, sale, pause and recovery capacity data, wherein the user payment behavior data comprises: monthly electric charge amount, payment amount, arrearage amount, payment time, payment frequency and user payment behavior data under the condition of late payment.
Selling electric quantity: the electricity selling quantity is the quantity of electricity sold by the electric power enterprise and can obtain the selling income according to the quantity of electricity, and comprises the quantity of electricity sold to the user for direct consumption, the quantity of electricity sold to other electric power enterprises in a wholesale way, and the quantity of electricity supplied by the electric power enterprise to non-electric power production, basic construction and non-production departments and the like of the enterprise. And E, business expansion and installation: the system is a general name of the service flow of the power supply department in the whole process from power customer application to actual power utilization. The method mainly comprises new installation, capacity increase, sales, capacity reduction, suspension recovery and the like. The net capacity increase is equal to the new capacity increase minus the sales capacity decrease. Risk of electric charge recovery: the electric charge recovery risk refers to the risk that the electric charge cannot be timely recovered and the like due to the reasons of shutdown, bankruptcy, recombination and conversion of the power utilization enterprise, poor user operation condition, shortage of user mobile funds, user re-lease, government removal, social stability and the like.
Specifically, the factors that need to be considered for the user electric charge recycling risk are divided into: model output factors, data statistics factors, and data construction factors.
The model output factors comprise the development trend of the power market where the user is located, the data statistical factors comprise monthly payment frequency, historical generation times of late fees, generation time of the last late fees and electric charge amount, and the data construction factors comprise payment duration and operation capacity change.
And after the power market development situation perception result of the user is obtained, the factors required by the early warning modeling of the user power charge recycling risk are constructed by using the internal power charge payment and business expansion data of the user.
The acquisition of user data items using the electricity marketing internal system is shown in table 2 below:
TABLE 2 electric charge payment and business expansion transaction data for electricity customers
(1) The electric charge payment data are generated when the user transacts the electric charge payment service each time, the power market development situation perception result takes the monthly degree as the frequency, the consistency of the data frequency is established when a unified algorithm is modeled, the times of generating the electric charge payment data are counted according to the month, and the monthly electric charge payment frequency factor f of the user is establishedm。
(2) For electricity customers with normal production operation and healthy fund turnover, the interval between the electricity fee issuing time and the electricity fee payment completion time is short, which indicates that the electricity fee payment is relatively timely. And constructing the user electric charge payment duration by utilizing the difference between the user electric charge payment time and the electric charge issuing time:
dp＝tp-tr；
(3) the electricity consumption client can handle the business expansion volume increase, reduction, pause, recovery and other businesses according to the operation requirement of the electricity consumption client, and represents that the operation state of the electricity consumption client needs to be changed and the production plan needs to be adjusted. And (3) constructing the monthly capacity variation of the user by utilizing the user business expansion capacity completion capacity, the business expansion capacity reduction completion capacity, the business expansion pause capacity and the business expansion recovery capacity:
Δc＝ca+cr-cs-cc；
(4) through the historical data of user's electric charge payment, count the generation times C of the historical electric charge late paymentf。
(5) The generation time t of the late electric charge late payment is counted through the historical data of the electric charge payment of the userl。
In summary, the factor data required for establishing the user electricity charge recycling risk early warning model can be divided into three types, namely model output, data statistics and data construction, according to the acquisition mode, as shown in table 3 below:
TABLE 3 early warning modeling consideration of power consumption and electricity charge recovery risk
And establishing a user electric charge recovery risk early warning model by an SVM regression algorithm by using the electric power market development situation perception result, the electric charge payment and the business expansion electric data of the user and combining factor data established according to the electric power market development situation perception result, the electric charge payment and the business expansion electric data as input and outputting whether the electric charge late payment is generated in the next month. The probability that the user generates the electric charge late fee in the ith month is as follows:
wherein C (i-1) is the amount of the electricity charge generated by the user in the (i-1) month.
And S105, constructing a user electric charge recovery risk early warning model through a classification regression algorithm according to the factors needing to be considered for the user electric charge recovery risk so as to quantify the probability of the risk of the user electric charge recovery and identify the users with the electric charge recovery risk.
The specific algorithm involved in the present invention is described as follows:
(1) bayesian season adjustment
Bayesian season adjustment is a classic method proposed by Akaike of Japanese economics in the eighties of the last century, has a sufficient theoretical basis, and practice proves that the season adjustment effect of the Bayesian season adjustment is superior to that of a traditional adjustment method based on moving average, and increasingly receives wide attention and application.
Suppose { YtT ═ 1,2, … N } is an observation time series and is suitable for the additive model as follows:
Yt＝Tt+St+It；
wherein T istIs a trend-cycle factor, StIs a seasonal factor, ItIs an irregular factor. StIs defined as P and assumes that N is MP (t is MP + j, j is 1,2, … P, M is 0,1, … M-1). The purpose of seasonal adjustment is to achieve a decomposition of the observation time series. Only the addition model needs to be discussed, and if the observation sequence is suitable for the multiplication model, the addition model can be formed after logarithmic processing.
Seasonal adjustments using classical regression methods are considered first. Usually to Tt、StAnd (3) fitting:
wherein f isk(t)、gk(t) needs to be chosen appropriately, the corresponding sum of squared residuals is:
minimizing Q determines the parameter ak、bk(k-1, 2, … N) to yield Tt、StFitting by the regression method described above is usually subject to f at the sequence endpointsk(·)、gk(. -) functional form constraints. The most unconstrained choice is to take K ═ N, defining the function:
namely Tt＝at，St＝btI.e. handle Tt、St(t ═ 1,2, … N) was directly targeted for estimation. Because the number of undetermined parameters is more than that of observed data, if the undetermined parameters are difficult to estimate by using a common least square method, a Bayesian method is introduced to TtAnd StAnd (6) directly estimating.
(2) SVM regression
SVM regression is the application of support vectors in the field of functional regression. The sample points of the SVM regression are of only one type, and the optimal hyperplane sought is to minimize the "total deviation" of all sample points from the hyperplane. At this time, the sample points are all between two boundary lines, and the calculation of the optimal regression hyperplane is also equivalent to the calculation of the maximum interval.
Given a data set D:
the essence of the regression problem is to find the function f (x) in order to infer the y value for any pattern x.
The SVM regression is determined by defining a hard-epsilon band hyperplane (as shown in fig. 2), i.e., all sample points in the data set D satisfy:
-ε≤yi-f(xi) ≦ ε, i ≦ 1,2, … N, convert the regression problem to find the optimal hard- ε band hyperplane, i.e.:
(3) dynamic time warping
Dynamic Time Warping (DTW) is a measure of two Time sequences of different lengths, X ═ X1,…,xi,…,xm) And Y ═ Y1,…,yj,…,yn) The method of similarity of (1). DTW calculates the similarity between two time series by extending and shortening the time series. Dynamic time warping, DTW, is a typical optimization problem.
And when the two templates are matched, accumulating the warping function D (i, j) corresponding to the minimum distance, wherein the warping function D (i, j) is expressed as the warping path distance between the point i on the sequence X and the point j on the sequence Y: d (i, j) — Dist (i, j) + min { D (i-1, j), D (i, j-1), D (i-1, j-1) }, where Dist (i, j) is the euclidean distance between point i on X and point j on Y.
According to the embodiment, the electric charge recovery risk early warning method based on electric power market development situation perception is provided, the influence of the upstream electric power market development situation on the production operation of the user under the current economic situation is considered, and the electric charge recovery risk early warning work is limited to the consideration of the internal electric power related factors. The invention provides the electric charge recovery risk early warning method which utilizes internal and external data for modeling and continuous learning to automatically quantify the electric charge recovery risk of the electricity customers, and overcomes the defects that the traditional method is subjective and lacks quantification means. The invention provides a method for separating and extracting the development trend of the power market by using a season adjustment algorithm, so that the analysis work can focus on the trend characteristics of the power market influenced by the development of the macro economy, the influence of other factors is eliminated, the influence analysis on each economic factor is more thorough, and the development situation perception result of the power market can more accurately reflect the economic production operation trend of the industry.
The embodiment of the invention also provides a storage medium, and the storage medium stores a computer program, and when the computer program is executed by a processor, the computer program realizes part or all of the steps of the electric charge recycling risk early warning method based on the electric power market development situation awareness. The storage medium may be a magnetic disk, an optical disk, a Read-only memory (ROM) or a Random Access Memory (RAM).
Those skilled in the art will readily appreciate that the techniques of the embodiments of the present invention may be implemented as software plus a required general purpose hardware platform. Based on such understanding, the technical solutions in the embodiments of the present invention may be essentially or partially implemented in the form of a software product, which may be stored in a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.
The above-described embodiments of the present invention should not be construed as limiting the scope of the present invention.
Claims (5)
1. A power market development situation awareness-based electric charge recovery risk early warning method is characterized by comprising the following steps:
extracting trend characteristics of historical electricity sales development of the industry by using a seasonal adjustment algorithm;
analyzing the influence of each economic index on the fluctuation of the trend characteristics, and determining the economic index which needs to be considered when predicting the development trend of the power selling amount of each industry;
according to the economic indexes, a power market development situation prediction model is built through a regression algorithm, and the power market development situation of the user is perceived;
according to the perception result of the power market development situation of the user, the factors of the user power charge recycling risk to be considered are identified by combining the user payment behavior data of the user and the user business expansion, capacity increase, capacity reduction, account cancellation, suspension and capacity recovery data, wherein the user payment behavior data comprises: monthly electric charge amount, payment amount, arrearage amount, payment time, payment frequency and user payment behavior data under the condition of late payment;
and according to the factors needing to be considered for the user electric charge recovery risk, constructing a user electric charge recovery risk early warning model through a classification regression algorithm so as to quantify the probability of the risk of the user electric charge recovery and identify the users with the electric charge recovery risk.
2. The method of claim 1, wherein extracting trend features of industry historical electricity sales development using a season adjustment algorithm comprises:
the power selling amount is decomposed into three subsequences of trend item of power selling amount, season item of power selling amount and random item of power selling amount by using Bayes season regulation algorithm:
Q(i)＝Qt(I)+Qs(i)+Qr(I)；
wherein { Q (i) | i ∈ 1, 2.. and n } represents industry historical electricity selling amount data, and { Q | i ∈ 1, 2.. and n } represents industry historical electricity selling amount datat(i) I belongs to 1,2, the right, n is a trend item of electricity selling amount, { Q ∈ is sets(i) I belongs to 1,2, the right, n is a season item of electricity sale, { Q ∈ Qr(i) And | i belongs to 1,2,. and n } is a random term of electricity selling quantity.
3. The method of claim 2, wherein analyzing the influence of each economic indicator on the fluctuation of the trend characteristic to determine the economic indicator to be considered in predicting the trend of the power sold in each industry comprises:
carrying out correlation analysis by using the power selling trend items of each industry and each economic index to obtain influence factors needing to be considered when predicting the power selling development trend of each industry, wherein the influence factors are instantaneity indexes and precursor indexes, the instantaneity indexes, namely the current influence factors, influence the current power selling trend development, and the precursor indexes, namely the current influence factors, influence the future power selling trend development;
and analyzing the leading relation between the leading indicators and the development of the power selling trend by adopting a dynamic time warping algorithm.
4. The method of claim 3, wherein the instantaneity indicators include average human availability revenue, second industry GDP acceleration, third industry GDP acceleration, and industry added value acceleration; the leading indicators include industry net amplification capacity, manufacturing PMI, and non-manufacturing PMI.
5. The method according to claim 4, wherein the factors to be considered for the risk of recovering the electric charge of the user are divided into: the system comprises model output factors, data statistics factors and data construction factors, wherein the model output factors comprise the development trend of an electric power market where a user is located, the data statistics factors comprise monthly payment frequency, historical generation times of late deposit, last generation time of late deposit and electric charge amount, and the data construction factors comprise payment duration and operation capacity change.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111120289.7A CN114358359A (en) | 2021-09-18 | 2021-09-18 | Electric charge recovery risk early warning method based on electric power market development situation perception |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111120289.7A CN114358359A (en) | 2021-09-18 | 2021-09-18 | Electric charge recovery risk early warning method based on electric power market development situation perception |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114358359A true CN114358359A (en) | 2022-04-15 |
Family
ID=81096183
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111120289.7A Pending CN114358359A (en) | 2021-09-18 | 2021-09-18 | Electric charge recovery risk early warning method based on electric power market development situation perception |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114358359A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113610409A (en) * | 2021-08-12 | 2021-11-05 | 北京中电普华信息技术有限公司 | Early warning method and device for electric charge recovery risk |
-
2021
- 2021-09-18 CN CN202111120289.7A patent/CN114358359A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113610409A (en) * | 2021-08-12 | 2021-11-05 | 北京中电普华信息技术有限公司 | Early warning method and device for electric charge recovery risk |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ribeiro et al. | Transfer learning with seasonal and trend adjustment for cross-building energy forecasting | |
| Guo et al. | The hybrid PROPHET-SVR approach for forecasting product time series demand with seasonality | |
| Khosravi et al. | A neural network-GARCH-based method for construction of Prediction Intervals | |
| TWI690865B (en) | Transaction volume prediction method and device | |
| Wang et al. | Using adaptive network-based fuzzy inference system to forecast automobile sales | |
| JP6837949B2 (en) | Prediction system and method | |
| CN109428344B (en) | Multi-power-supply investment planning method and device comprising wind power plant | |
| CN106779280B (en) | Decision-making determination method and system for secondary equipment major repair and technical modification | |
| CN114140176B (en) | Adjustable capacity prediction method and device for load aggregation platform | |
| WO2024001045A1 (en) | Carbon quota surplus and deficit prediction method, apparatus, and electronic device, and storage medium | |
| CN111709779A (en) | Trading variety optimization system for electric power spot market | |
| CN113256324A (en) | Data asset pricing method, device, computer equipment and storage medium | |
| Sudipa et al. | Trend Forecasting of the Top 3 Indonesian Bank Stocks Using the ARIMA Method | |
| CN114358359A (en) | Electric charge recovery risk early warning method based on electric power market development situation perception | |
| Wang et al. | Stock price prediction for new energy vehicle enterprises: An integrated method based on time series and cloud models | |
| CN114254916A (en) | Multi-dimensional supplier risk real-time early warning method with power grid service characteristics | |
| Boltürk et al. | Electricity consumption forecasting using fuzzy time series | |
| CN112884572A (en) | Multi-microgrid transaction optimization method and system under credit risk | |
| WO2023197502A1 (en) | Comprehensive power evaluation method and apparatus | |
| CN116777616A (en) | Probability density distribution-based future market new energy daily transaction decision method | |
| Reichelstein et al. | Cost-and price dynamics of solar PV modules | |
| CN114444883A (en) | Retail market risk monitoring, prevention and control early warning method for electricity selling company | |
| Chen et al. | Data Analytics in Power Markets | |
| CN110490480A (en) | A kind of newly-built transmission line of electricity economic benefit quantitative estimation method | |
| CN111027831A (en) | Investment decision method and system based on economic evaluation model |
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 |