CN111428745A - Clustering analysis-based low-voltage user electricity utilization feature extraction method - Google Patents
Clustering analysis-based low-voltage user electricity utilization feature extraction method Download PDFInfo
- Publication number
- CN111428745A CN111428745A CN202010004292.1A CN202010004292A CN111428745A CN 111428745 A CN111428745 A CN 111428745A CN 202010004292 A CN202010004292 A CN 202010004292A CN 111428745 A CN111428745 A CN 111428745A
- Authority
- CN
- China
- Prior art keywords
- electricity
- data
- user
- users
- adopting
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a clustering analysis-based low-voltage user electricity utilization feature extraction method, which comprises the following steps: classifying the measured original data by adopting a clustering algorithm, wherein the preprocessing mainly comprises the steps of cleaning the data and identifying and repairing the data; classifying the user electricity consumption data of different typical transformer areas by adopting an improved K-Means clustering algorithm, determining the clustering quantity by adopting an improved hill climbing method, analyzing the electricity consumption behavior characteristics of users, comparing the electricity consumption behavior characteristics with the low-voltage electricity consumption characteristics, and attributing the mortgage users to one of five typical transformer areas, namely a high-rise residential area, an old residential area, an urban and rural junction, an isolated small-sized settlement point and a rural area according to the acquired electricity metering characteristics; based on the distribution area information, the non-metering data and the non-electric metering data, extracting the parameters of the distribution line by adopting a compact extraction method, constructing a standard electricity utilization characteristic parameter matrix, determining the correlation between each influence factor and the electricity load characteristics of the user, and obtaining the characteristic parameters of each typical distribution area.
Description
Technical Field
The invention relates to the field of power consumer feature extraction, in particular to a low-voltage consumer power utilization feature extraction method based on cluster analysis.
Background
In the aspect of classification and characteristic extraction of power utilization behaviors of power users in a low-voltage distribution area, the current classification is mainly carried out on the users on the basis of user load characteristic indexes and historical data, for example, the classification is carried out on the basis of the traditional industry and the clustering analysis is carried out on the users, but the defects of the classification are that different power utilization modes among the users are omitted; the classification is carried out based on the user value embodiment, and the defect is that the division mode is too macroscopic; from the load of the transformer substation, a mean value clustering method is adopted to classify the power utilization of the users, or a K-means clustering algorithm is adopted to classify the power utilization of the users on the premise of a fuzzy C mean value clustering method.
The user characteristics are described as specific electrical behaviors presented when the user electric equipment operates, and the load characteristics can be divided into steady-state characteristics, transient-state characteristics and operation mode characteristics according to the operation state of the user electric equipment, wherein the steady-state characteristics are the electrical characteristics presented by the electric equipment in a stable operation state, are main characteristics of a load and are the key points of current power user electricity utilization analysis. The research on domestic and foreign classification methods is more, such as machine learning, pattern recognition and other intelligent algorithms, and the judgment of the intelligent algorithms on the electrical elements in the power system and the application research of applying the algorithms to the power system have attracted the attention of domestic and foreign researchers.
Disclosure of Invention
The invention provides a low-voltage user electricity utilization characteristic extraction method based on cluster analysis, which adopts a compact extraction method to extract parameters of a distribution line, can better reflect the low-voltage user electricity utilization characteristics, and is described in detail in the following description:
a low-voltage user electricity utilization feature extraction method based on cluster analysis comprises the following steps:
from the actual demand, the power utilization district is subjected to attribution division according to the hierarchical classification principle of a power grid company, a low-voltage district characteristic extraction index principle is constructed in the aspect of influencing the healthy operation of the district, and an index set is set;
determining main influence factors of the electrical load characteristics of users in the low-voltage distribution area according to the construction characteristics of the distribution area;
determining non-metering data, electrical metering data and non-electrical metering data to be collected according to the principles that index data are easy to obtain, an index system is feasible, the practicability of an evaluation model is strong and the like;
according to the operation state of the smart grid region, low-voltage user electricity metering data reflecting the operation state level are extracted, and the method comprises the following steps of: the maximum power utilization value, the minimum power utilization value, the average power utilization value and the power utilization fluctuation deviation rate;
classifying the measured original data by adopting a clustering algorithm, wherein the preprocessing mainly comprises the steps of cleaning the data and identifying and repairing the data;
classifying the user electricity consumption data of different typical transformer areas by adopting an improved K-Means clustering algorithm, determining the clustering quantity by adopting an improved hill climbing method, analyzing the electricity consumption behavior characteristics of users, comparing the electricity consumption behavior characteristics with the low-voltage electricity consumption characteristics, and attributing the mortgage users to one of five typical transformer areas, namely a high-rise residential area, an old residential area, an urban and rural junction, an isolated small-sized settlement point and a rural area according to the acquired electricity metering characteristics;
based on the distribution area information, the non-metering data and the non-electric metering data, extracting the parameters of the distribution line by adopting a compact extraction method, constructing a standard electricity utilization characteristic parameter matrix, determining the correlation between each influence factor and the electricity load characteristics of the user, and obtaining the characteristic parameters of each typical distribution area.
The method for constructing the standard electricity utilization characteristic parameter matrix specifically comprises the following steps:
for each electricity consumption characteristic quantity XjIs provided with Xj＝[xij]T(i ═ 1.,. multidot.m) is the total number of users participating in the power consumption behavior analysis, standard normalization processing is carried out on elements in the same column in the matrix, and a user standard power consumption characteristic quantity matrix is obtained for N power consumption characteristic quantities to be evaluated and M users participating in the power consumption behavior analysis;
for different characteristic quantities, calculating index weights of the characteristic quantities by adopting an entropy weight method;
and acquiring comprehensive power utilization information of the user.
The technical scheme provided by the invention has the beneficial effects that:
1) determining non-metering data, electric metering data and non-electric metering data which need to be collected, and defining the electricity utilization characteristic indexes of low-voltage residential users comprises the following steps: the voltage qualification rate, the current three-phase unbalance degree, the distribution transformer load rate, the file accuracy rate, the summary table acquisition success rate, the household table acquisition success rate, the metering equipment fault rate and the transformer area line loss rate are 8 sub-indexes, the considered factors are comprehensive, and the electricity utilization characteristics of low-voltage users can be better reflected;
2) the invention adopts a compact extraction method to extract the parameters of the distribution line, and overcomes the defects of a loose method: the propagation constant should be considered as an input parameter and a signal processing algorithm needs to be used to extract the eigenfrequency. Compared with other extraction algorithms, the algorithm has simpler program design and smaller memory occupation;
3) according to the electrical characteristic parameters of the samples, the improved K-Means clustering algorithm is adopted for classification, so that two main defects in the traditional clustering are overcome: firstly, before clustering starts, the category number k needs to be given in advance; secondly, the algorithm does not give the selection principle of the initial clustering center.
Drawings
FIG. 1 is a flow chart of a method for extracting electricity consumption characteristics of low-voltage users based on cluster analysis according to the present invention;
FIG. 2 is a flow chart of the pre-processing of the power consumption data of the users in the low-voltage distribution area according to the present invention;
FIG. 3 is a flow chart of clustering user load data based on an improved fuzzy mean algorithm of an improved hill climbing method provided by the invention;
FIG. 4 is a power consumption load clustering chart based on an improved K-means clustering algorithm provided by the invention;
FIG. 5 is a diagram of a neural network architecture provided by the present invention;
FIG. 6 is a graph of peak load (MW) versus temperature (deg.C) in summer according to the present invention;
fig. 7 is a graph showing the variation trend of peak load (MW) and temperature factors in holidays in summer.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in further detail below.
As shown in fig. 1, the parameter extraction module related to the present invention includes the following steps:
the invention provides a method for extracting electricity utilization characteristics of low-voltage residential users, which comprises the following specific steps of:
s1: from the actual demand, the power utilization district is subjected to attribution division according to the hierarchical classification principle of a power grid company, a low-voltage district characteristic extraction index principle is constructed in the aspect of influencing the healthy operation of the district, and an index set is set;
s2: determining main influence factors of the electrical load characteristics of users in the low-voltage distribution area according to the construction characteristics of the distribution area;
s3: determining non-metering data, electrical metering data and non-electrical metering data to be collected according to the principles that index data are easy to obtain, an index system is feasible, the practicability of an evaluation model is strong and the like;
s4: according to the operation state of the smart grid region, low-voltage user electricity metering data reflecting the operation state level are extracted, and the method comprises the following steps of: key indexes such as a maximum power utilization value, a minimum power utilization value, an average power utilization value and a power utilization fluctuation deviation rate;
s5: classifying the measured original data by adopting a clustering algorithm, wherein the preprocessing mainly comprises the steps of cleaning the data and identifying and repairing the data;
s6: classifying the electricity consumption data of users in different typical transformer areas by adopting an improved K-Means clustering algorithm, determining the clustering quantity by adopting an improved hill climbing method, analyzing the electricity consumption behavior characteristics of the users, comparing the electricity consumption behavior characteristics with the low-voltage electricity consumption characteristics, and attributing the mortgage users to one of five typical transformer areas, namely a high-rise residential area, an old residential area, an urban and rural junction, an isolated small-sized settlement point and a rural area according to the acquired electricity metering characteristics.
S7: based on the distribution area information, the non-metering data and the non-electric metering data, extracting the parameters of the distribution line by adopting a compact extraction method, constructing a standard electricity utilization characteristic parameter matrix, determining the correlation between each influence factor and the electricity load characteristics of the user, and obtaining the characteristic parameters of each typical distribution area.
The low-voltage platform feature extraction index principle in step S1 needs to follow the following 7 principles: the method comprises a reliability principle, an integrity principle, a real-time principle, an accuracy principle, an usability principle, a planning principle and a predictability principle.
(a) And (5) reliability principle. The information collection reliability principle means that collected information is generated by a real object or environment, information sources are guaranteed to be reliable, the collected information can reflect real conditions, and the reliability principle is the basis of information collection.
(b) Integrity principle. The integrity of information acquisition means that the acquired information must be complete in content, the information acquisition must acquire information reflecting the whole appearance of an object according to certain standard requirements, and the integrity principle is the basis of information utilization.
(c) And (5) a real-time principle. The real-time property of information acquisition refers to the ability to obtain the required information in time, and generally has three layers of meanings: the first is the time interval from the occurrence of the information to the acquisition, the shorter the interval is, the more timely the information is, and the fastest the information acquisition is synchronous with the information occurrence; secondly, when an enterprise or an organization executes a task and needs a certain information urgently, the information can be quickly acquired, which is called as timely; the third is the time taken for collecting all the information required by a certain task, and the less the time taken is, the faster the time taken is. The timeliness principle ensures the timeliness of information acquisition.
(d) And (5) accuracy principle. The principle of accuracy means that the relevance degree of the collected information with an application target and a work requirement is high, the expression of the collected information is correct, the collected information belongs to the category of the collection purpose, and the method has applicability and is valuable relative to enterprises or organizations. The higher the degree of association, the more adaptive and more accurate. The accuracy principle guarantees the value of information acquisition.
(e) Principle of easy use. The usability principle means that the collected information is convenient to use according to a certain representation form.
(f) And (4) planning principle. The acquired information not only needs to meet the current needs, but also needs to take care of future development; not only needs to widely spread information sources, but also needs to be kept constant and accumulated day by day; rather than being random, a more thorough and detailed acquisition plan and regulation can be made according to the task, the expense and the like of the unit.
(g) The principle of predictability. The information acquisition personnel need to master the development trends of society, economy and science and technology, and the acquired information needs to be good at grabbing the seedling head and grabbing the moving direction with certain advance and attention paid to the practical requirements. The future is known at any time, and predictive information which can guide the future development is collected.
In step S2, the main factors influencing the characteristics of the electrical load of the low-voltage transformer area users include:
(1) digital signature
Total daily electricity consumption: the index reflects the power utilization level of the whole platform area
Average daily load: the index reflects the average power consumption level of the region
The difference of day curve: the index shows the discrete degree of the electricity consumption of the station area user
(2) Index feature
Average acquisition success rate: the definition mode of the index calculation is to calculate how many users' power consumption in a distribution area is collected, then calculate the average number collected every day in a year, and finally determine the ratio of the average value to the total number of users in the distribution area as the average collection success rate of the distribution area.
Percentage of line loss of the transformer area: the calculation mode of the index is defined as the difference between the total meter and the sub-meters of the distribution area, and then the ratio of the difference to the total meter, and the index reflects the electric energy percentage wasted by the electric energy when the distribution area supplies power to the users.
Average number of values 0: the definition of the index is that the daily power consumption of the user is calculated, whether the power consumption is 0 or not is judged, the number of 0 in each day in one year is counted, then the average number of 0 power consumption in each day is obtained by averaging, and the index reflects the average number of unused power of the user in the distribution area, and reflects the population characteristics of the distribution area.
(3) User quantity feature
Month with maximum load of month: the index is defined as the month with the maximum month load, and the index shows the influence of the user on the month or the temperature on the user
Seasonal characteristics: the feature is defined as the number of users with the highest electric load per season, and shows the influence of seasons on the distribution area.
Festival and holiday characteristics: the feature is defined as whether the maximum load of the user occurs on holidays or weekday holidays, and shows how much the cell is affected by the holidays.
(1) Climate factors: along with the improvement of the living standard of residents and the popularization of household appliances, the cooling load is increased in summer, the heating load is increased in winter, and the influence of climate change on the load of a transformer area is more and more obvious. Temperature is the most important factor affecting the load, and other climatic factors, as well as humidity, wind speed, precipitation and duration of sunshine, also have an effect on the change of the load of the platform area.
(2) Time factor: the change of the load of the transformer area is influenced by the change of the day hours, the sunshine hours and the temperature in different seasons; because the working mode and the living habits of people are circularly alternated in work and rest, the load of the transformer area has day-week periodicity; legal holidays are accompanied by activities such as rest, tourism and the like, so that the phase change is obviously reduced compared with the ordinary time, and researches also show that the load before and after the holidays is slightly reduced by the accumulated influence of the holidays.
In the step S7, the constructed standard electricity utilization characteristic parameter matrix includes the following processes:
s71: in the K-Means clustering algorithm, abnormal values and special variables of user electricity utilization data have great influence on clustering yield, so that standard normalization processing needs to be performed on the abnormal values and the special variables to obtain a standard electricity utilization characteristic quantity matrix.
For each electricity consumption characteristic quantity XjIs provided with Xj＝[xij]T(i 1.., M) is the total number of users participating in the power usage behavior analysis. The elements of the same column in the matrix are subjected to a standard normalization process,
then, for N power utilization characteristic quantities to be evaluated and M users participating in power utilization behavior analysis, a user standard power utilization characteristic quantity matrix is obtained
Wherein x'ijAnd the j-th electricity utilization characteristic value to be evaluated of the ith user.
S72: setting of weight of different characteristic quantities
And for different characteristic quantities, calculating the index weight of the characteristic quantity by adopting an entropy weight method. Wherein the proportion of the ith user under the jth electricity utilization characteristic value to be evaluated
Entropy of j-th electricity utilization characteristic value to be evaluated
Thereby obtaining the entropy weight of the j power utilization characteristic value to be evaluated
Obtain a weight matrix W ═ W1…wj…wN]。
S73: user comprehensive electricity utilization information
Including (1) quantification of temperature data; (2) dividing and normalizing the date type; (3) quantification of weather factors, etc.
With regard to the date type, conventionally, people only divide it into weekdays (monday to friday) and bijourneys (saturday and sunday). Because the working day and the double-holiday people have different electricity utilization habits, the load difference is large. Of course, in addition to these conventional divisions, major holidays need to be divided separately. The responses to load for different date types are numerically characterized herein: the day type value for the working day was taken to be 1, saturday 0.5, sunday 0.6, and major holiday 0. The unified normalization of date types is accomplished by this quantization method.
The load is affected by the temperature and is also limited by the weather conditions. In the general case, the weather conditions cannot be quantitatively described I but the analysis of the load variation situation in various weather conditions can be found by W: the load changes steadily in sunny days, the load is increased under the influence of illumination in cloudy days, and the load is obviously increased in rainy days and snowy days. In order to reflect the influence degree, the relevant documents are referred to, and the influence degree of the weather on the load is divided into five grades of clear, cloudy, rain, snow and heavy snow in turn, and corresponding measurement values are respectively given as 0.1, 0.2, 0.4, 0.5 and 0.8.
S74: the standard electricity utilization characteristic quantity matrix is weighted and calculated to obtain a weighting matrix capable of reflecting the comprehensive electricity utilization information of the user,
as shown in fig. 3, in the step S6, for the processed comprehensive power consumption information of the user, the specific steps of clustering the user load data based on the improved fuzzy average algorithm of the improved hill climbing method are as follows:
s61: and determining a weighting index w according to the preprocessed user electricity utilization data, and iterating to terminate parameters.
S62: determining the number of clusters and the cluster center: correcting based on the weighting index, selecting a deviation value of a measured data sample and substituting the deviation value into a hill climbing function:
if it is
(
Is a certain sample of the sample set), the hill-climbing function takes the maximum value, and is preferably selected
Is the first cluster center, now with
Order to
When searching for other cluster centers, it is necessary to eliminate
The modified hill climbing function becomes:
solving the hill climbing function to obtain the maximum value of the hill climbing function of the second classification
And corresponding load samples
And order
β can be 4, 8, 16, etc.
Similarly, the mountain climbing function of the t-th time is as follows:
wherein the content of the first and second substances,
in order to be a new hill-climbing function,
is the hill-climbing function of the previous step,
is the maximum value of the hill climbing function of the previous step.
When in use
The process of finding the cluster center is ended, wherein a positive decimal number can be taken for the classified convergence coefficient. Judging the total times of iterative clustering before convergence as the classification number c of fuzzy clustering, and making the hill-climbing function value the maximum sample in each clustering process
For the corresponding initial cluster center vi。
S63: after obtaining the initial clustering center V0Then, based on equation (4), an initial membership matrix U can be obtained0。
S64: calculating a membership matrix U ═ Urj}，urjRepresenting the degree of membership of the jth sample with respect to the r cluster center.
S65: an objective function is calculated which is the weighted sum of squares of the distances of the individual samples to all cluster centers:
using formula (5) as an iterative formula, and calculating the iterative error delta J of the target function twice before and afterw(U, V) is less than a given positive number, clustering ends.
In step S7, the correlation analysis between each influencing factor and the user electrical load characteristic is specifically as follows:
(1) example of load time characteristic analysis
1) Annual load characteristic analysis
TABLE 2 Peak load duration time Table (unit: hour)
The comparison of table 2 shows that the maximum load and the minimum load of the residential users in the low-voltage distribution area are improved compared with those in the rural distribution area in 18 years, which indicates that the power supply demand is increased. However, the utilization hours of the peak loads of 80%, 90%, 95% and 97% of the first type of residential users are smaller than those of the second type of residential users, which indicates that the peak electricity prices are not well implemented. And carrying out power utilization scheduling at the peak power price in due time. The second category of residential users has the highest load utilization hours as follows: 5207.9 hours; the first category of residential users has the highest load utilization hours: 5051.0 hours, the number of hours of maximum load utilization by the first type of residential users is reduced compared to the second type of residential users.
2) Analysis of monthly load characteristics
The monthly load characteristic indexes comprise monthly load rate, monthly average daily load rate, monthly maximum peak-valley difference rate, monthly maximum, monthly minimum load, monthly average daily peak-valley difference and the like. The average daily load of each month, the maximum daily peak-to-valley difference and peak-to-valley difference rate of each month, and the load rate of each month and month are only analyzed here.
TABLE 3 average daily load rate per month for first and second types of residential users
As can be seen from Table 3, the average daily load rate of each month generally fluctuates about 80%, which is not less than 70% nor more than 90%, and from the viewpoint of fluctuation, the overall seasonal characteristic is high in summer and low in the rest of time; the load rate fluctuation is relatively smooth in view of the two types of stations.
(2) Example of load-temperature dependence analysis
And analyzing the relation between six temperature factors of highest temperature, lowest temperature, 2-point temperature, 8-point temperature, 14-point temperature and 20-point temperature of a certain low-voltage resident user and the daily highest load. By y1Simple correlation coefficient representing maximum temperature and maximum load, denoted by y2Simple correlation coefficient representing maximum load and minimum temperature, denoted by y3Simple correlation coefficient representing maximum load and 2-point temperature, denoted by y4Simple correlation coefficient representing maximum load and 8-point temperature, denoted by y5Simple correlation coefficient representing maximum load and 14-point temperature, denoted by y6A simple correlation coefficient representing the maximum load with the 20 point temperature.
To more clearly analyze the correlation between daily peak charge and temperature factor, the daily peak charge on working days was divided into two cases: the maximum load occurs in the morning (between 0 and 12 points) and in the evening (between 12 and 24 points). For the two different cases, the most relevant temperature factors are respectively found out for correlation analysis.
1) The maximum load occurs in the morning (between 0 and 12 points)
Y is obtained through calculation1＝0.88，y2＝0.82，y3＝0.79,y4＝0.84,y5＝0.84,y60.80. Namely, the correlation coefficient of the daily peak load and the highest temperature is 0.88, the correlation coefficient of the daily peak load and the temperature at the 8 morning point is 0.84, and the correlation coefficient of the daily peak load and the temperature at the 14 morning point is also 0.84, which are highly related to the three; the temperature correlation coefficient was 0.79, which was the smallest at 2 a.m.. The trend of peak load and the most relevant temperature factor in summer is shown in fig. 7.
2) The maximum load occurs late (between 12-24 points)
Through statistical calculation, y is obtained1＝0.85，y2＝0.76，y3＝0.76,y4＝0.85,y5＝0.85,y60.79. Namely, the correlation coefficients of the daily peak load and the highest temperature appearing in the late peak period, the temperature at 8 am and the temperature at 14 am are all 0.85, and are highly correlated with the three; the correlation coefficient with the lowest daily temperature and the temperature at 2 am was the smallest, and was 0.76.
(3) Correlation analysis of peak load and temperature factor of double holidays
Through statistical calculation, y is obtained1＝0.84，y2＝0.76，y3＝0.71,y4＝0.85,y5＝0.81,y60.74. Namely, the correlation coefficient of the daily peak load and the highest temperature is 0.84, the correlation coefficient of the daily peak load and the temperature at 8 morning is 0.85, and the correlation coefficient of the daily peak load and the temperature at 14 morning is 0.81, which are highly related to the three components; the temperature correlation coefficient was 0.71, which was the smallest at 2 a.m.. The peak load (MW) of the summer holiday in 2018 and the variation trend of each temperature factor are shown in fig. 7.
In the embodiment of the present invention, except for the specific description of the model of each device, the model of other devices is not limited, as long as the device can perform the above functions.
Those skilled in the art will appreciate that the drawings are only schematic illustrations of preferred embodiments, and the above-described embodiments of the present invention are merely provided for description and do not represent the merits of the embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (2)
1. A low-voltage user electricity utilization feature extraction method based on cluster analysis is characterized by comprising the following steps:
from the actual demand, the power utilization district is subjected to attribution division according to the hierarchical classification principle of a power grid company, a low-voltage district characteristic extraction index principle is constructed in the aspect of influencing the healthy operation of the district, and an index set is set;
determining influence factors of the electric load characteristics of users in the low-voltage distribution area according to the construction characteristics of the electric distribution area;
determining non-metering data, electric metering data and non-electric metering data which need to be collected according to the principle that index data are easy to obtain, an index system is feasible and the practicability of an evaluation model is strong;
according to the operation state of the smart grid region, low-voltage user electricity metering data reflecting the operation state level are extracted, and the method comprises the following steps of: the maximum power utilization value, the minimum power utilization value, the average power utilization value and the power utilization fluctuation deviation rate;
classifying the measured original data by adopting a clustering algorithm, wherein the preprocessing mainly comprises the steps of cleaning the data and identifying and repairing the data;
classifying the user electricity consumption data of different typical transformer areas by adopting an improved K-Means clustering algorithm, determining the clustering quantity by adopting an improved hill climbing method, analyzing the electricity consumption behavior characteristics of users, comparing the electricity consumption behavior characteristics with the low-voltage electricity consumption characteristics, and attributing the mortgage users to one of five typical transformer areas, namely a high-rise residential area, an old residential area, an urban and rural junction, an isolated small-sized settlement point and a rural area according to the acquired electricity metering characteristics;
based on the distribution area information, the non-metering data and the non-electric metering data, parameters of the distribution line are extracted by adopting compact extraction, a standard electricity utilization characteristic parameter matrix is constructed, the correlation between each influence factor and the electricity load characteristics of the user is determined, and the characteristic parameters of each typical distribution area are obtained.
2. The method for extracting the electricity utilization characteristics of the low-voltage users based on the cluster analysis according to claim 1, wherein the step of constructing the standard electricity utilization characteristic parameter matrix specifically comprises the steps of:
for each electricity consumption characteristic quantity XjIs provided with Xj＝[xij]T(i ═ 1.,. multidot.m) is the total number of users participating in the power consumption behavior analysis, standard normalization processing is carried out on elements in the same column in the matrix, and a user standard power consumption characteristic quantity matrix is obtained for N power consumption characteristic quantities to be evaluated and M users participating in the power consumption behavior analysis;
for different characteristic quantities, calculating index weights of the characteristic quantities by adopting an entropy weight method;
and acquiring comprehensive power utilization information of the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010004292.1A CN111428745A (en) | 2020-01-03 | 2020-01-03 | Clustering analysis-based low-voltage user electricity utilization feature extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010004292.1A CN111428745A (en) | 2020-01-03 | 2020-01-03 | Clustering analysis-based low-voltage user electricity utilization feature extraction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111428745A true CN111428745A (en) | 2020-07-17 |
Family
ID=71546976
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010004292.1A Pending CN111428745A (en) | 2020-01-03 | 2020-01-03 | Clustering analysis-based low-voltage user electricity utilization feature extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111428745A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112765826A (en) * | 2021-01-27 | 2021-05-07 | 长沙理工大学 | Indoor hemp planting resident user identification method based on power consumption frequency distribution relative entropy |
| CN113434690A (en) * | 2021-08-25 | 2021-09-24 | 广东电网有限责任公司惠州供电局 | Clustering algorithm-based electricity utilization prediction evaluation method, device, system and medium |
| CN117113159A (en) * | 2023-10-23 | 2023-11-24 | 国网山西省电力公司营销服务中心 | Deep learning-based power consumer side load classification method and system |
-
2020
- 2020-01-03 CN CN202010004292.1A patent/CN111428745A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112765826A (en) * | 2021-01-27 | 2021-05-07 | 长沙理工大学 | Indoor hemp planting resident user identification method based on power consumption frequency distribution relative entropy |
| CN113434690A (en) * | 2021-08-25 | 2021-09-24 | 广东电网有限责任公司惠州供电局 | Clustering algorithm-based electricity utilization prediction evaluation method, device, system and medium |
| CN113434690B (en) * | 2021-08-25 | 2022-02-08 | 广东电网有限责任公司惠州供电局 | Clustering algorithm-based electricity utilization prediction evaluation method, device, system and medium |
| CN117113159A (en) * | 2023-10-23 | 2023-11-24 | 国网山西省电力公司营销服务中心 | Deep learning-based power consumer side load classification method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Forecasting residential energy consumption: Single household perspective | |
| CN109359389B (en) | Urban electric vehicle charging decision method based on typical load dynamic game | |
| US11043808B2 (en) | Method for identifying pattern of load cycle | |
| Chicco et al. | Emergent electricity customer classification | |
| CN111428745A (en) | Clustering analysis-based low-voltage user electricity utilization feature extraction method | |
| CN111932402A (en) | Short-term power load bidirectional combined prediction method based on similar day and LSTM | |
| Gerbec et al. | An approach to customers daily load profile determination | |
| CN101398919A (en) | Electric power requirement analytic system for utilizing mode analysis and method thereof | |
| CN110109971A (en) | A kind of low-voltage platform area user power utilization Load Characteristic Analysis method | |
| CN110033134A (en) | A kind of short-term load forecasting algorithm of segmentation day by day considering meteorologic factor | |
| Bobric et al. | Clustering techniques in load profile analysis for distribution stations | |
| JP2006011715A (en) | Estimation method for resource consumption, and device | |
| CN115660225A (en) | Electricity load prediction management method and system based on ammeter communication module | |
| CN115796915A (en) | Electricity price prediction method and system for electricity trading market | |
| Chicco et al. | Load profiling revisited: prosumer profiling for local energy markets | |
| Sun et al. | Impacts of solar penetration on short-term net load forecasting at the distribution level | |
| Colley et al. | Queensland load profiling by using clustering techniques | |
| CN114048200A (en) | User electricity consumption behavior analysis method considering missing data completion | |
| Parra et al. | Customers' demand clustering analysis—A case study using smart meter data | |
| CN109149644B (en) | Light-storage integrated online strategy matching and collaborative optimization method based on big data analysis | |
| Ogihara et al. | A study of risk reduction for daily peak load demand forecasting | |
| Grigoras et al. | Knowlegde extraction from Smart Meters for consumer classification | |
| Zhang et al. | Load characteristics analysis based on improved k-means clustering algorithm | |
| Li et al. | Sensitivity analysis and forecast of power load characteristics based on meteorological feature information | |
| Liao | Research on load loss prediction of distribution network outage based on hybrid neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |