CN114819530A - A demand-side flexible resource adjustment potential forecasting method and system - Google Patents

Abstract

The invention provides a method and system for predicting the adjustable potential of flexible resources on the demand side, including: clustering the load data of multiple users in a selected area in turn from the time scale and the industry scale, to obtain typical users of each industry in each season. Daily load curve: Based on the typical daily load curve of users in various industries in each season, the differential autoregressive moving average model is used to predict the load value of users in various industries in the future; The present invention uses the differential autoregressive moving average model to predict the load value of users in various industries in the future, which can effectively deal with the possible instability of the input data, effectively reduce the impact of the data instability on the prediction results, and based on the future moment. The load value of users in various industries can be used to calculate the adjustable potential of flexible resources on the demand side. It is not necessary to establish a complex equipment load model, which effectively simplifies the model for calculating the adjustable potential.

Description

A demand-side flexible resource adjustment potential forecasting method and system

technical field

The invention belongs to the technical field of an adjustable load potential calculation method, and in particular relates to a demand-side flexible resource adjustable potential prediction method and system.

Background technique

With the development of society and the adjustment of industrial structure, the power system is facing new challenges, such as the continuous growth of peak-to-valley difference and the consumption of clean energy brought about by the gradual increase in the proportion of new energy power generation. Due to its flexibility and adjustability, demand-side resources are conducive to solving a series of problems faced by the power system. Therefore, researching and predicting the regulation potential of adjustable loads plays an important role in the optimal operation and control of power systems, as well as in power demand side management and load dispatching.

In recent years, with the development of new power electronic technology and control methods, the load characteristics have undergone fundamental changes. On the one hand, users can change the original power consumption mode according to market prices or incentive mechanisms, and actively participate in the control of power grid operation; on the other hand, electric vehicles, energy storage equipment, etc. have the ability to interact with the power grid in two directions, and have the function of peak shaving and valley filling. , which provides a new means for power grid regulation. The research on load participation in grid interactive dispatch has become the focus of attention. The adjustable load on the demand side can be used as an important means to reduce the peak load and balance the power supply gap. The research methods for load characteristics mainly include statistical synthesis method, overall identification method, fault simulation method, etc., which provide data support for the study of adjustable potential. At present, the prediction methods for tunable potential are mainly based on equipment models, which require the establishment of a large number of equipment models. The parameters required by the equipment model, such as equipment type, quantity, and operating characteristics, are various and inappropriate to obtain, which becomes a major problem in predicting the tunable potential.

SUMMARY OF THE INVENTION

In order to overcome the above-mentioned deficiencies of the prior art, the present invention proposes a demand-side flexible resource adjustment potential prediction method, including:

The load data of multiple users in the selected area are clustered in turn from the time scale and the industry scale, and the typical daily load curve of users in each industry in each season is obtained;

Based on the typical daily load curve of users in various industries in each season, the differential autoregressive moving average model is used to predict the load value of users in various industries in the future;

According to the load value of users in various industries in the future, the adjustable potential of flexible resources on the demand side is calculated.

Preferably, the load data of multiple users in the selected area are clustered sequentially from the time scale and the industry scale to obtain typical daily load curves of users in each industry in each season, including:

For the load data of multiple users in the selected area, the Canopy-Kmeans clustering algorithm is used to cluster from the time scale, and the typical daily load curve of each user in each season is obtained;

For the typical daily load curve of each user in each season, the Canopy-Kmeans clustering algorithm is used to cluster from the industry scale, and the typical daily load curve of each industry user in each season is obtained.

Preferably, the load data of multiple users in the selected area is clustered from the time scale using the Canopy-Kmeans clustering algorithm to obtain the typical daily load curve of each user in each season, including:

Based on the load data of multiple users in the selected area, calculate the load rate, peak-to-valley difference and average load of each user;

Based on the load rate, peak-to-valley difference and average load of each user, the Canopy clustering algorithm is used to find multiple time clustering center points from the load data of multiple users;

Based on the time clustering center point and the load rate, peak-to-valley difference and average load of each user, the Kmeans clustering algorithm is used to cluster the load data of multiple users from the time scale, and the typical day of each user in each season is obtained. load curve.

Preferably, the typical daily load curve of each user in each season is clustered from the industry scale using the Canopy-Kmeans clustering algorithm to obtain the typical daily load curve of each industry user in each season, including:

Based on the load rate, peak-valley difference and average load of each user, the Canopy clustering algorithm is used to find multiple industry cluster center points from the typical daily load curve of each user in each season;

Based on the industry clustering center point and the load rate, peak-to-valley difference and average load of each user, the Kmeans clustering algorithm is used to cluster the typical daily load curve of each user in each season from the industry scale, and obtain each industry in each season. Typical daily load profile of the user.

Preferably, based on the typical daily load curves of users in various industries in each season, a differential autoregressive moving average model is used to predict the load values of users in various industries in the future, including:

The unit root test method is used to test the stationarity of the typical daily load curves of users in each industry in each season: if the test results are stable, the difference order of the typical daily load curve is set to the preset value, otherwise the typical daily load The difference operation is performed on the curve to obtain the difference order;

Calculate the autocorrelation coefficient and partial autocorrelation coefficient of the typical daily load curve of users in each industry in each season, and determine the autoregressive model order and the moving average model order according to the autocorrelation coefficient and the partial autocorrelation coefficient;

Based on the difference order, autoregressive model order and moving average model order of the typical daily load curves of users in various industries in each season, the difference autoregressive moving average model is constructed;

The corresponding differential autoregressive moving average models are used to predict the load value of users in various industries in the future.

Preferably, according to the load value of users in various industries in the future, calculating the adjustable potential of flexible resources on the demand side, including:

For users in various industries, count the load value distribution of users in the industry at the same historical moment within the specified time range before the future time, and calculate the mean and variance of the load according to the distribution of load values;

Based on the mean value and variance of the load, the maximum and minimum values satisfying the Three Sigma principle are taken as the maximum baseline load and the minimum baseline load of the corresponding industry user in the future;

Based on the load value, maximum baseline load and minimum baseline load of users in various industries in the future, the adjustable potential of flexible resources on the demand side is calculated.

Preferably, based on the load value, the maximum baseline load and the minimum baseline load of users in various industries in the future, the calculation of the demand-side flexible resource adjustment potential includes:

Based on the load value and the maximum baseline load of users in various industries in the future, calculate the potential of demand-side load increase, and based on the load value and minimum baseline load of users in various industries in the future, calculate the potential of demand-side load reduction;

The demand-side load increasing potential and the demand-side load decreasing potential are used as the demand-side flexible resource adjustment potential.

Preferably, the calculation formula of the demand-side load upward adjustment potential is as follows:

η _1,j =M _j ×(L _1,j -H _MN,j )

In the formula, η _1,j is the demand-side load increase potential of industry j, M _j is the number of users in industry j, L _1,j is the maximum baseline load of industry j, and H _MN,j is the load of users in industry j in the future value;

The calculation formula of the demand-side load reduction potential is as follows:

η _2,j =M _j ×(H _MN,j -L _2,j )

where η _2,j is the demand-side load reduction potential of industry j, and L _2,j is the minimum baseline load of industry j.

Based on the same inventive concept, the present invention also provides a demand-side flexible resource adjustable potential prediction system, including: a load clustering module, a load prediction module and an adjustable potential module;

The load clustering module is used for clustering the load data of multiple users in the selected area in turn from the time scale and the industry scale to obtain the typical daily load curves of users in each industry in each season;

The load prediction module is used to predict the load value of users in various industries in the future by using a differential autoregressive moving average model based on the typical daily load curves of users in various industries in each season;

The adjustable potential module is used to calculate the adjustable potential of demand-side flexible resources according to the load values of users in various industries in the future.

Preferably, the load clustering module is specifically used for:

For the load data of multiple users in the selected area, the Canopy-Kmeans clustering algorithm is used to cluster from the time scale, and the typical daily load curve of each user in each season is obtained;

For the typical daily load curve of each user in each season, the Canopy-Kmeans clustering algorithm is used to cluster from the industry scale, and the typical daily load curve of each industry user in each season is obtained.

Preferably, the load prediction module is specifically used for:

The unit root test method is used to test the stationarity of the typical daily load curves of users in each industry in each season: if the test results are stable, the difference order of the typical daily load curve is set to the preset value, otherwise the typical daily load The difference operation is performed on the curve to obtain the difference order;

Calculate the autocorrelation coefficient and partial autocorrelation coefficient of the typical daily load curve of users in each industry in each season, and determine the autoregressive model order and the moving average model order according to the autocorrelation coefficient and the partial autocorrelation coefficient;

Based on the difference order, autoregressive model order and moving average model order of the typical daily load curves of users in various industries in each season, the difference autoregressive moving average model is constructed;

The corresponding differential autoregressive moving average models are used to predict the load value of users in various industries in the future.

Preferably, the adjustable potential module is specifically used for:

For users in various industries, count the load value distribution of users in the industry at the same historical moment within the specified time range before the future time, and calculate the mean and variance of the load according to the distribution of load values;

Based on the mean value and variance of the load, the maximum and minimum values satisfying the Three Sigma principle are taken as the maximum baseline load and the minimum baseline load of the corresponding industry user in the future;

Based on the load value, maximum baseline load and minimum baseline load of users in various industries in the future, the adjustable potential of flexible resources on the demand side is calculated.

The present invention also provides a computer device, comprising: one or more processors;

memory for storing one or more programs;

When the one or more programs are executed by the one or more processors, the aforementioned method for predicting the adjustable potential of demand-side flexible resources is implemented.

The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed, the aforementioned method for predicting the adjustable potential of flexible resources on the demand side is implemented.

Compared with the closest prior art, the present invention has the following beneficial effects:

The invention provides a method and system for predicting the adjustable potential of flexible resources on the demand side, including: clustering the load data of multiple users in a selected area in turn from the time scale and the industry scale, to obtain the typical users of each industry in each season. Daily load curve: Based on the typical daily load curve of users in various industries in each season, the differential autoregressive moving average model is used to predict the load value of users in various industries in the future; The present invention uses the differential autoregressive moving average model to predict the load value of users in various industries in the future, which can effectively deal with the possible instability of the input data, effectively reduce the impact of the data instability on the prediction results, and based on the future moment. The load value of users in various industries can be used to calculate the adjustable potential of flexible resources on the demand side. It is not necessary to establish a complex equipment load model, which effectively simplifies the model for calculating the adjustable potential.

The invention uses the Canopy-Kmeans clustering algorithm to classify a large amount of data on the time and industry scales, which provides a classification basis for analyzing the user's adjustable potential by season and industry, and the typical daily load curve obtained is used as the classification basis. The calculation basis of the load simplifies the calculation process.

Description of drawings

1 is a schematic flowchart of a method for predicting the adjustable potential of demand-side flexible resources provided by the present invention;

2 is a schematic flowchart of the Canopy-Kmeans clustering algorithm in the embodiment of the present invention;

3 is a schematic flowchart of an ARIMA prediction model algorithm in an embodiment of the present invention;

Figure 4(a) is a schematic diagram of a typical load of Spring and Autumn users after two clustering in an embodiment of the present invention;

Figure 4(b) is a schematic diagram of a typical load of summer users after two clusterings in an embodiment of the present invention;

Fig. 4(c) is a schematic diagram of typical load of winter users after two clustering in the embodiment of the present invention;

Figure 5(a) is a schematic diagram of typical daily load curves of various types of users in Spring and Autumn in the embodiment of the present invention;

Figure 5(b) is a schematic diagram of typical daily load curves of various types of users in summer in the embodiment of the present invention;

Figure 5(c) is a schematic diagram of typical daily load curves of various users in winter in the embodiment of the present invention;

Fig. 6 (a) is the spring and autumn user 1 load predicted by ARIMA algorithm in the embodiment of the present invention;

Fig. 6(b) is the spring and autumn user 2 load predicted by the ARIMA algorithm in the embodiment of the present invention;

Fig. 6 (c) is the spring and autumn user 3 load predicted by ARIMA algorithm in the embodiment of the present invention;

Fig. 7(a) is the summer user 1 load predicted by the ARIMA algorithm in the embodiment of the present invention;

Fig. 7(b) is the summer load of user 2 predicted by the ARIMA algorithm in the embodiment of the present invention;

Fig. 8(a) is the winter load of user 1 predicted by the ARIMA algorithm in the embodiment of the present invention;

Fig. 8(b) is the winter load of user 2 predicted by the ARIMA algorithm in the embodiment of the present invention;

Figure 9(a) is a schematic diagram of the load up-regulation potential and down-regulation potential of user 1 in Spring and Autumn in the embodiment of the present invention;

Fig. 9(b) is a schematic diagram of the load up-regulation potential and down-regulation potential of Spring and Autumn User 2 in the embodiment of the present invention;

Figure 9(c) is a schematic diagram of the load up-regulation potential and down-regulation potential of Spring and Autumn User 3 in the embodiment of the present invention;

Figure 10(a) is a schematic diagram of the load up-regulation potential and down-regulation potential of summer user 1 in the embodiment of the present invention;

Figure 10(b) is a schematic diagram of the load up-regulation potential and down-regulation potential of summer user 2 in the embodiment of the present invention;

Fig. 11(a) is a schematic diagram of the load up-regulation potential and down-regulation potential of winter user 1 in the embodiment of the present invention;

Figure 11(b) is a schematic diagram of the load up-regulation potential and down-regulation potential of winter user 2 in the embodiment of the present invention;

FIG. 12 is a schematic structural diagram of a demand-side flexible resource adjustment potential prediction system provided by the present invention.

Detailed ways

The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

Example 1:

A schematic flowchart of a method for predicting the adjustable potential of demand-side flexible resources provided by the present invention is shown in FIG. 1 , including:

Step 1: Cluster the load data of multiple users in the selected area in turn from the time scale and the industry scale to obtain the typical daily load curves of users in each industry in each season;

Step 2: Based on the typical daily load curves of users in various industries in each season, the differential autoregressive moving average model is used to predict the load value of users in various industries in the future;

Step 3: Calculate the adjustable potential of flexible resources on the demand side according to the load values of users in various industries in the future.

The method for predicting the adjustable potential of demand-side flexible resources adopted in the present invention can be used for short-term adjustable potential prediction of demand-side flexible resources.

Specifically, step 1 includes:

Step 1.1, use the Canopy clustering algorithm to find K _N cluster center points, including:

Step 1.1.1, select N user load data, randomly arrange the historical load data of the same user for one year respectively, and set up initial data sets S ₁ , S ₂ , . . . , S _N respectively. Since the load rate, peak-to-valley difference, and average load can effectively reflect the adjustability and adjustability of user load, a central point P _N is selected from the initial clustering sample set according to these three indicators, among which,

Load factor:

为日负荷曲线的平均值，G_m为日负荷曲线的最大值。in,

is the average value of the daily load curve, and G _m is the maximum value of the daily load curve.

Peak-to-valley difference:

Δ=G _m -G _n

Among them, G _n is the minimum value of the daily load curve.

Step 1.1.2, select the distance closest to the center point as the distance threshold T2 _-N (ie T2 in Figure 2), and the farthest distance from the center point is the distance threshold T1 _-N (ie T1 in Figure 2) , and T _1-N >T _2-N ;

Step 1.1.3, take _PN as the cluster center point of the first cluster, and remove the point _PN from the initial clustering sample set _SN ;

Step 1.1.4, randomly select a point Q _N from the remaining data sample set S _N , calculate the distance from Q _N to all known cluster center points, and examine the minimum distance D _N : if T _{2-N ≤} D _N ≤T _1-N , record Q _N with a weak mark, indicating that point Q _N belongs to the cluster, and add Q _N to it; if D _N ≤T _2-N , record point Q _N with a strong mark, indicating point Q _N belongs to this cluster, and Q _N is deleted from the data sample set S _N ; if D _N >T _1-N , then Q _N forms a new cluster, and Q _N is deleted from the data sample set S _N ;

Step 1.1.5, repeat step 1.4 until the number of elements in the set S _N is zero, and finally obtain K _N cluster center points.

Step 1.2, use the Kmeans clustering algorithm to classify the original data set (that is, the user's load data), including:

Step 1.2.1, according to step 1.1, the S _N data sets generate a total of K _N cluster centers, which are respectively k _1N , k _2N , ..., k _kN , and the load rate, peak-to-valley difference, and average load are used as the cluster centers. class evaluation index;

Step 1.2.2, calculate the similarity between the typical daily load of each user and the cluster evaluation index of the cluster center point;

Step 1.2.3, add the user to the cluster with the highest similarity with the center point, and update the cluster center point;

Step 1.2.4, iterate steps 1.2.2 to 1.2.3 until the number of iteration steps is 500;

Step 1.2.5, obtain the typical daily load curve of users in different seasons.

The flowchart of the Canopy-Kmeans clustering algorithm of step 1.1-step 1.2 is shown in Figure 2.

Step 1.3: Perform secondary clustering on the clustering results obtained in step 1.2 according to the industry scale. The clustering method is the same to obtain the typical daily load curves of users in different industries in different seasons. The historical data of users in the cluster center are used as differential autoregression. Input data for the moving average model (aka ARIMA forecast model) algorithm.

Specifically, in step 1.3, perform secondary clustering on the K _N typical daily load curves of N users obtained in step 1.2 according to the clustering method in steps 1.1 to 1.2, to obtain K _N typical users in K _M industries . The daily load curve of typical users in different seasons is analyzed, and the collected load data of typical users is used as the input data of the ARIMA prediction model.

Between step 1 and step 2, data preprocessing may also be performed on the data of the clustering result.

Specifically, the input data of different categories obtained by the clustering in step 1 are processed, wherein the default value is replaced by the mean value of the feature where it is located, and the over-limit value and singular value are replaced by the mean value of the two times before and after the feature where it is located.

Step 2 specifically includes:

Step 2.1, build the original sequence:

X _t =α ₁ X _t-1 +α ₂ X _t-2 +…α _p X _tp +ε _t +β ₁ ε _t-1 +β ₂ ε _t-2 +…β _q ε _tq

In the formula, p and q are the order of autoregression and moving average respectively; α ₁ ,α ₂ ,…,α _p are autoregressive coefficients; β ₁ ,β ₂ ,…,β _q are moving average coefficients, X is the use of The steady-state part of the original typical user load data construction sequence, ε is the white noise sequence in the original sequence. The subscripts of X and ε represent time, respectively.

In step 2.2, since the time series needs to be guaranteed to be stationary when the ARIMA algorithm is performed, the ADF test (ie, unit root test) method is used to test the stationarity in this step;

Step 2.3, if the original sequence is non-stationary after the test in step 2.2, this step is to perform a difference operation on the non-stationary time series to obtain the difference order d, so that the sequence is stationary; if the original sequence is stationary after the test in step 4.2 , the difference order is set to the default value, usually the default value is set to 0;

Step 2.4, find the autocorrelation coefficient ACF and the partial autocorrelation coefficient PACF, where the calculation formula of the autocorrelation coefficient ACF is:

Partial autocorrelation coefficient PACF calculation formula:

为协方差，EX_t为期望值，

为方差。in,

is the covariance, EX _t is the expected value,

is the variance.

The autocorrelation coefficient ACF is used to determine the correlation between the daily load at time t and the load at this time in the past week. The partial correlation coefficient is used to determine the correlation between the time t in the daily load and the load at this time in the past week under the influence of factors such as temperature changes.

Step 2.5: Determine the model parameters according to the above autocorrelation coefficient and partial autocorrelation coefficient, that is, the AR autoregressive model order p and the MA moving average model order q, where the p-order AR autoregressive model:

X _t =α ₁ X _t-1 +α ₂ X _t-2 +...α _p X _tp +u _t

q-order MA moving average model:

X _t =ε _t +β ₁ ε _t-1 +β ₂ ε _t-2 +…β _q ε _tq

Among them, ε _t , ε _t-1 ,…,ε _tq are white noise sequences at time t, t-1,…, tq.

Step 2.6, combine the autoregressive model (AR), the moving average model (MA) and the difference method to obtain the difference autoregressive moving average model ARIMA (p, d, q);

Step 2.7, verify and optimize the model, compare and analyze the predicted value obtained by using the model and the actual value of the predicted period to verify the accuracy of the method, and can use this method to predict the load in an unknown period. The predicted value H _MN obtained by the ARIMA algorithm is used as the actual load value when calculating the adjustable potential.

A schematic flowchart of the ARIMA prediction model algorithm is shown in Figure 3, which also includes the input data processing process between Step 1 and Step 2.

Step 3 specifically includes:

取满足

(即三西格玛)原则中的最大值L₁和最小值L₂作为t时刻的最大基线负荷和最小基线负荷；Step 3.1, according to the clustering results of various industries in different seasons obtained in step 1, the calculation method of the baseline load of each industry in different seasons is the same, that is, the distribution of loads at time t within the specified time range (for example, a week) before the statistical forecast date, and calculate their Mean _Δf and variance

take satisfaction

The maximum value L ₁ and the minimum value L ₂ in the three-sigma principle are used as the maximum baseline load and the minimum baseline load at time t;

Step 3.2, according to the predicted value H _MN obtained in step 2.7 and the maximum baseline load L ₁ and the minimum baseline load L ₂ obtained in step 3.1, establish a calculation model of adjustable potential:

Single-day load escalation potential:

η _2,j =M _j ×(H _MN,j -L _2,j )

Single-day load reduction potential:

η ₂ =M _j ×(H _MN -L ₂ )

where, η _1,j is the demand-side load increase potential of industry j, M _j is the number of users in industry j, L _1,j is the maximum baseline load of industry j, and H _MN,j is the user of industry j in the future The load value of ; η _2,j is the demand-side load reduction potential of industry j, and L _2,j is the minimum baseline load of industry j.

Step 3.3, take the demand-side load increase potential and demand-side load decrease potential as the demand-side flexible resource adjustment potential.

Example 2:

A specific calculation example of the method for predicting the adjustable potential of demand-side flexible resources is given below. According to the sample data of 38 user loads in a certain place, the adjustable load of users in different seasons and industries in the region is predicted; among them, there are 38 users in this region. Users, the sampling point is 24 per day.

Figures 4(a)-(c) are the results of two clusterings. It can be seen from the figure that according to the seasonal characteristics, it can be known that the time scale can be divided into three seasons: spring and autumn, summer and winter, and according to the industry characteristics, it can be divided into industrial users , commercial users, and residential users. Figures 5(a)-(c) are typical daily load curves of various users in different seasons. According to the analysis of the industry characteristics of the region, the typical user 1 load and the typical user 2 load are commercial users and industrial users, respectively, and there are two peak periods. The load is larger in winter and summer, and the typical user 3 load is the residential user load. The curve is relatively smooth.

Figures 6(a) to 8(b) are the prediction results of the ARIMA algorithm, and the prediction curves of three industries in different seasons are obtained. According to the analysis of the industry characteristics in the region, it is obtained that user 1 is a commercial user, user 2 is an industrial user, and user 3 is a Resident users. It can be seen from the curve that there is little error with the actual value, which proves that the predicted value can be used.

Figures 9(a) to 11(b) are schematic diagrams of the load up-regulation potential and down-regulation potential of typical users in the three industries of industrial, commercial and civil. In Figures 9(a) to 11(b), there are three curves of user predicted load, minimum baseline load and maximum baseline load, respectively. The difference between the maximum baseline load and the user's predicted load is the load increase potential, and the user's predicted load and the minimum baseline load The time is the load reduction potential. According to the analysis of industry characteristics in this area, it is obtained that user 1 is a commercial user, user 2 is an industrial user, and user 3 is a residential user. Multiply by the number of users in the industry to get the daily adjustable potential of the industry. According to this method, the adjustable potential of different industries in different seasons in a certain region can be predicted, and technical support for power demand side management and load dispatching can be provided.

Example 3:

Based on the same inventive concept, the present invention also provides a demand-side flexible resource adjustment potential prediction system, the system structure is shown in Figure 12, including: a load clustering module, a load prediction module and an adjustable potential module;

Among them, the load clustering module is used to cluster the load data of multiple users in the selected area in turn from the time scale and the industry scale, and obtain the typical daily load curve of users in each industry in each season;

The load forecasting module is used to predict the load value of users in various industries in the future by using the differential autoregressive moving average model based on the typical daily load curves of users in various industries in each season;

The adjustable potential module is used to calculate the adjustable potential of flexible resources on the demand side according to the load value of users in various industries in the future.

Among them, the load clustering module is specifically used for:

For the load data of multiple users in the selected area, the Canopy-Kmeans clustering algorithm is used to cluster from the time scale, and the typical daily load curve of each user in each season is obtained;

For the typical daily load curve of each user in each season, the Canopy-Kmeans clustering algorithm is used to cluster from the industry scale, and the typical daily load curve of each industry user in each season is obtained.

Among them, for the load data of multiple users in the selected area, the Canopy-Kmeans clustering algorithm is used to cluster from the time scale, and the typical daily load curve of each user in each season is obtained, including:

Based on the load data of multiple users in the selected area, calculate the load rate, peak-to-valley difference and average load of each user;

Based on the load rate, peak-to-valley difference and average load of each user, the Canopy clustering algorithm is used to find multiple time clustering center points from the load data of multiple users;

Based on the time clustering center point and the load rate, peak-to-valley difference and average load of each user, the Kmeans clustering algorithm is used to cluster the load data of multiple users from the time scale, and the typical daily load curve of each user in each season is obtained. .

Among them, for the typical daily load curve of each user in each season, the Canopy-Kmeans clustering algorithm is used to cluster from the industry scale, and the typical daily load curve of each industry user in each season is obtained, including:

Based on the load rate, peak-valley difference and average load of each user, the Canopy clustering algorithm is used to find multiple industry cluster center points from the typical daily load curve of each user in each season;

Based on the industry clustering center point and the load rate, peak-to-valley difference and average load of each user, the Kmeans clustering algorithm is used to cluster the typical daily load curve of each user in each season from the industry scale, and the average daily load curve of each user in each season is obtained. Typical daily load curve.

Among them, the load forecasting module is specifically used for:

The unit root test method is used to test the stationarity of the typical daily load curves of users in each industry in each season: if the test results are stable, the difference order of the typical daily load curve is set to the preset value, otherwise the typical daily load The difference operation is performed on the curve to obtain the difference order;

Calculate the autocorrelation coefficient and partial autocorrelation coefficient of the typical daily load curve of users in various industries in each season, and determine the autoregressive model order and moving average model order according to the autocorrelation coefficient and the partial autocorrelation coefficient;

Based on the difference order, autoregressive model order and moving average model order of the typical daily load curves of users in various industries in each season, the difference autoregressive moving average model is constructed;

The corresponding differential autoregressive moving average models are used to predict the load value of users in various industries in the future.

Among them, the adjustable potential module is specifically used for:

For users in various industries, count the load value distribution of industry users at the same historical moment within the specified time range before the future time, and calculate the mean and variance of the load according to the distribution of load values;

Based on the mean and variance of the load, take the maximum and minimum values that satisfy the Three Sigma principle as the maximum baseline load and minimum baseline load of the corresponding industry users in the future;

Based on the load value, maximum baseline load and minimum baseline load of users in various industries in the future, the adjustable potential of flexible resources on the demand side is calculated.

Among them, based on the load value, maximum baseline load and minimum baseline load of users in various industries in the future, the adjustable potential of flexible resources on the demand side is calculated, including:

Based on the load value and the maximum baseline load of users in various industries in the future, calculate the potential of demand-side load increase, and based on the load value and minimum baseline load of users in various industries in the future, calculate the potential of demand-side load reduction;

The demand-side load-up potential and demand-side load-down potential are taken as the demand-side flexible resource adjustment potential.

Among them, the calculation formula of the demand-side load increase potential is as follows:

η _1,j =M _j ×(L _1,j -H _MN,j )

In the formula, η _1,j is the demand-side load increase potential of industry j, M _j is the number of users in industry j, L _1,j is the maximum baseline load of industry j, and H _MN,j is the load of users in industry j in the future value;

The calculation formula of the demand-side load reduction potential is as follows:

η _2,j =M _j ×(H _MN,j -L _2,j )

where η _2,j is the demand-side load reduction potential of industry j, and L _2,j is the minimum baseline load of industry j.

Example 4:

Based on the same inventive concept, the present invention also provides a computer device, the computer device includes a processor and a memory, the memory is used for storing a computer program, the computer program includes program instructions, and the processor is used for executing the Program instructions stored by the computer storage medium. The processor may be a central processing unit (Central Processing Unit, CPU), or other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array ( Field-Programmable GateArray, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, specifically suitable for loading And execute one or more instructions in the computer storage medium to realize the corresponding method process or corresponding function, so as to realize the steps of the method for predicting the adjustable potential of demand-side flexible resources in the above embodiment.

Example 5:

Based on the same inventive concept, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It can be understood that, the computer-readable storage medium here may include both a built-in storage medium in a computer device, and certainly also an extended storage medium supported by the computer device. The computer-readable storage medium provides storage space in which the operating system of the terminal is stored. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions may be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the steps of the method for predicting the adjustable potential of demand-side flexible resources in the above embodiment.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the scope of its protection. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand: Those skilled in the art can still make various changes, modifications or equivalent replacements to the specific embodiments of the application after reading the present disclosure, but these changes, modifications or equivalent replacements are all within the protection scope of the pending claims of the application.

Claims (14)

1. A demand side flexible resource adjustable potential prediction method is characterized by comprising the following steps:

clustering the load data of a plurality of users in the selected area from the time scale and the industry scale in sequence to obtain a typical daily load curve of each industry user in each season;

based on typical daily load curves of users in various industries in various seasons, predicting load values of the users in various industries at a future moment by adopting a differential autoregressive moving average model;

and calculating the flexible resource adjustable potential of the demand side according to the load values of users in various industries at a future moment.

2. The method of claim 1, wherein the clustering load data of a plurality of users in a selected area sequentially from a time scale and an industry scale to obtain a typical daily load curve for each industry user in each season comprises:

clustering load data of a plurality of users in the selected area from a time scale by using a Canopy-Kmeans clustering algorithm to obtain a typical daily load curve of each user in each season;

and clustering the typical daily load curve of each user in each season by using a Canopy-Kmeans clustering algorithm from the industry scale to obtain the typical daily load curve of each industry user in each season.

3. The method of claim 2, wherein the clustering load data of a plurality of users in the selected area from the time scale using a Canopy-Kmeans clustering algorithm to obtain a typical daily load curve for each user in each season comprises:

calculating the load rate, the peak-valley difference and the average load of each user based on the load data of a plurality of users in the selected area;

based on the load rate, the peak-valley difference and the average load of each user, adopting a Canopy clustering algorithm to find a plurality of time clustering center points from the load data of a plurality of users;

and based on the time clustering center point and the load rate, the peak-valley difference and the average load of each user, clustering the load data of a plurality of users from a time scale by using a Kmeans clustering algorithm to obtain a typical daily load curve of each user in each season.

4. The method of claim 3, wherein the clustering the typical daily load curve of each user in each season from the industry scale by using a Canopy-Kmeans clustering algorithm to obtain the typical daily load curve of each industry user in each season comprises:

based on the load rate, the peak-valley difference and the average load of each user, a Canopy clustering algorithm is adopted to find a plurality of industry clustering center points from typical daily load curves of each season of each user;

and based on the industry clustering center point and the load rate, the peak-valley difference and the average load of each user, clustering the typical daily load curve of each user in each season by using a Kmeans clustering algorithm from the industry scale to obtain the typical daily load curve of each industry user in each season.

5. The method of claim 1, wherein the predicting the load value of each industry user at a future time by using a differential autoregressive moving average model based on a typical daily load curve of each industry user in each season comprises:

and (3) carrying out stability test on the typical daily load curves of users in various industries in various seasons by using a unit root test method: if the test result is stable, setting the difference order of the typical daily load curve as a preset value, otherwise, carrying out difference operation on the typical daily load curve to obtain the difference order;

respectively calculating autocorrelation coefficients and partial autocorrelation coefficients of typical daily load curves of users of various industries in various seasons, and determining an autoregressive model order and a moving average model order according to the autocorrelation coefficients and the partial autocorrelation coefficients;

constructing a differential autoregressive moving average model based on the differential order, the autoregressive model order and the moving average model order of the typical daily load curve of each industry user in each season;

and respectively adopting the corresponding difference autoregressive moving average model to predict the load value of each industry user at the future time.

6. The method of claim 1, wherein calculating the demand-side flexible resource tunable potential according to the load values of the industry users at the future time comprises:

counting load value distribution of the industry users at the same historical moment in a specified duration range before the future moment aiming at the users of each industry, and calculating the mean value and the variance of the load according to the load value distribution;

based on the mean value and the variance of the load, taking the maximum value and the minimum value which meet the three-sigma principle as the maximum baseline load and the minimum baseline load of the corresponding industry user at the future moment;

and calculating the flexible resource adjustable potential of the demand side based on the load value, the maximum baseline load and the minimum baseline load of each industry user at the future moment.

7. The method of claim 6, wherein calculating the demand-side flexible resource tunable potential based on the load values, the maximum baseline loads, and the minimum baseline loads of the industry users at the future time comprises:

calculating the load up-regulation potential of the demand side based on the load values and the maximum baseline loads of all industry users at the future time, and calculating the load down-regulation potential of the demand side based on the load values and the minimum baseline loads of all industry users at the future time;

and taking the demand side load up-regulation potential and the demand side load down-regulation potential as demand side flexible resource adjustable potential.

8. The method of claim 7, wherein the demand side load up-regulation potential is calculated as follows:

η _1,j ＝M _j ×(L _1,j -H _M-N,j )

in the formula eta _1,j For demand side load up-regulation potential, M, of industry j _j Number of users of industry j, L _1,j Is the maximum baseline load, H, of industry j _M-N,j The load value of the industry j user at the future moment;

the calculation formula of the demand side load down-regulation potential is as follows:

η _2,j ＝M _j ×(H _M-N,j -L _2,j )

in the formula eta _2,j For demand side load turndown potential, L, of industry j _2,j The minimum baseline load for industry j.

9. A demand-side flexible resource tunable potential prediction system, comprising: the system comprises a load clustering module, a load forecasting module and an adjustable potential module;

the load clustering module is used for sequentially clustering the load data of a plurality of users in a selected area from a time scale and an industry scale to obtain a typical daily load curve of each industry user in each season;

the load prediction module is used for predicting the load value of each industry user at a future moment by adopting a differential autoregressive moving average model based on the typical daily load curve of each industry user in each season;

and the adjustable potential module is used for calculating the flexible resource adjustable potential of the demand side according to the load values of users in various industries at a future moment.

10. The system of claim 9, wherein the load clustering module is specifically configured to:

clustering load data of a plurality of users in the selected area from a time scale by using a Canopy-Kmeans clustering algorithm to obtain a typical daily load curve of each user in each season;

and clustering the typical daily load curve of each user in each season by using a Canopy-Kmeans clustering algorithm from the industry scale to obtain the typical daily load curve of each industry user in each season.

11. The system of claim 9, wherein the load prediction module is specifically configured to:

and (3) carrying out stability test on the typical daily load curves of users in various industries in various seasons by using a unit root test method: if the test result is stable, setting the difference order of the typical daily load curve as a preset value, otherwise, carrying out difference operation on the typical daily load curve to obtain the difference order;

respectively calculating autocorrelation coefficients and partial autocorrelation coefficients of typical daily load curves of users of various industries in various seasons, and determining an autoregressive model order and a moving average model order according to the autocorrelation coefficients and the partial autocorrelation coefficients;

constructing a differential autoregressive moving average model based on the differential order, the autoregressive model order and the moving average model order of the typical daily load curve of each industry user in each season;

and respectively adopting corresponding difference autoregressive moving average models to predict the load value of each industry user at the future moment.

12. The system of claim 9, wherein the adjustable potential module is specifically configured to:

counting load value distribution of the industry users at the same historical moment in a specified duration range before the future moment aiming at the users of each industry, and calculating the mean value and the variance of the load according to the load value distribution;

based on the mean value and the variance of the load, taking the maximum value and the minimum value which meet the three-sigma principle as the maximum baseline load and the minimum baseline load of the corresponding industry user at the future moment;

and calculating the flexible resource adjustable potential of the demand side based on the load value, the maximum baseline load and the minimum baseline load of each industry user at the future moment.

13. A computer device, comprising: one or more processors;

a memory for storing one or more programs;

the one or more programs, when executed by the one or more processors, implement the demand-side flexible resource tunable potential prediction method of any one of claims 1 to 8.

14. A computer-readable storage medium having stored thereon a computer program which, when executed, implements the demand-side flexible resource tunable potential prediction method of any one of claims 1 to 8.

Images