CN115275977A - Power load prediction method and device - Google Patents
Power load prediction method and device Download PDFInfo
- Publication number
- CN115275977A CN115275977A CN202210746657.7A CN202210746657A CN115275977A CN 115275977 A CN115275977 A CN 115275977A CN 202210746657 A CN202210746657 A CN 202210746657A CN 115275977 A CN115275977 A CN 115275977A
- Authority
- CN
- China
- Prior art keywords
- influence
- sample
- factors
- influence factors
- factor
- 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method and a device for predicting an electric load, comprising the following steps: preliminarily screening the influence factors to obtain the power load under the relevant influence factors; calculating the influence weight of the related influence factors and sequencing to obtain the power load under the optimal influence factor; carrying out data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained; inputting a sample to be trained into a neural network for training to obtain an electric load prediction model; and predicting the electrical load at the current moment by using the electrical load prediction model. Compared with the prior art, the method and the device have the advantages that the optimal influence factors are obtained by screening and sorting the influence factors of the power load, and then the power load prediction model is obtained based on the optimal influence factors, so that the dimensionality of data is reduced, and the training speed and precision of the neural network are improved.
Description
Technical Field
The invention relates to the technical field of power load prediction, in particular to a power load prediction method and a power load prediction device.
Background
Accurate prediction of the power load is of great significance to safe operation of the power system. However, due to the complexity and variability of the power load, it is difficult to build an accurate model to predict the power load. With the rapid development of the energy internet, more and more electric equipment are connected to the internet, a large amount of electricity utilization data must be generated in the process, and how to utilize the massive data to deeply dig out the load change characteristics of various electric equipment from the massive data can improve the electricity utilization economy of a user side, is more beneficial to a power grid dispatching mechanism to make a power generation plan so as to realize supply and demand balance, and is beneficial to greatly improving the safety and economy of the operation of a power system.
The current prediction method of the power load mainly comprises a regression analysis method. The main principle of the regression analysis method is to construct a model with a certain periodic rule through the relationship between a dependent variable and one or more independent variables, and the model is used as a theoretical basis to analyze the change condition of the load predicted by the current data. The regression analysis method has the obvious disadvantages that the method depends on sample data excessively, a large amount of sample data is needed to support training, the training process is complicated, the method is sensitive to changes caused by environmental factors, and the anti-interference performance is weak.
Disclosure of Invention
To solve the above problems, embodiments of the present invention provide a method and an apparatus for predicting an electrical load.
An electrical load prediction method, comprising:
step 1: acquiring a historical training sample; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and regions;
and 2, step: preliminarily screening the influence factors to obtain the power load under the relevant influence factors;
and 3, step 3: calculating the influence weight of the related influence factors and sequencing the influence weight to obtain the power load under the optimal influence factor;
and 4, step 4: carrying out data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained;
and 5: inputting the sample to be trained into a neural network for training to obtain an electrical load prediction model;
step 6: and predicting the electric load at the current moment by using the electric load prediction model.
Preferably, the step 2: preliminarily screening the influence factors to obtain the power load under the relevant influence factors, wherein the power load comprises the following steps:
step 2.1: extracting corresponding electric loads under different influence factors according to the time sequence to form an influence factor time sequence and an electric load time sequence;
step 2.2: performing difference processing on the influence factor time sequence and the power load time sequence to obtain a difference sequence;
step 2.3: obtaining a distance coefficient of an influence factor at any time according to the difference sequence;
step 2.4: obtaining the correlation degree of corresponding influence factors according to the distance coefficient;
step 2.5: and taking the influence factors with the correlation degree larger than a preset threshold value as relevant influence factors.
Preferably, the calculation formula of the distance coefficient of the influencing factor at any time is as follows:
wherein r is0j(k) Distance coefficient, Δ, representing the influencing factor at time kj(k)=|x0(k)-xj(k)|,j=1,2,...,n,x0(k) A value representing the influencing factor at time k, xj(k) A value, delta, representing the electrical load at time kj(k) The k-th numerical value in the difference sequence is represented, M represents the minimum value in the difference sequence, M represents the maximum value in the difference sequence, and xi represents a preset value.
Preferably, the step 2.4: obtaining the correlation of the corresponding influence factors according to the distance coefficient, including:
the formula is adopted:
obtaining the correlation of the influence factors; where θ represents the degree of correlation of the influencing factors.
Preferably, the step 3: calculating the influence weight of the related influence factors and sequencing the influence weights to obtain the power load under the optimal influence factor, wherein the method comprises the following steps:
step 3.1: constructing a sample matrix according to the relevant influence factors and the corresponding electric loads;
step 3.2: constructing an influence weight matrix according to the sample matrix;
step 3.3: calculating the influence weight of the relevant influence factors according to the characteristic value of the influence weight matrix and sequencing;
step 3.4: and selecting the relevant influence factor corresponding to the influence weight with the highest ranking as the optimal influence factor.
Preferably, the sample matrix is:
wherein x isnpIndicating the value of the p-th relevant influencing factor at the time n.
Preferably, the step 3.2: constructing an influence weight matrix according to the sample matrix, comprising:
step 3.2.1: calculating the variance and covariance of each column of elements in the sample matrix;
step 3.2.2: constructing an influence weight matrix according to the variance and covariance of each column of elements; the construction formula of the influence weight matrix is as follows:
wherein r isNMRepresenting the values of the elements of the influence weight matrix in the Nth row and the Mth column, cov (i, j) representing the covariance of the ith column and the jth column in the sample matrix, var (i) representing the variance of the ith column in the sample matrix, var (j) representing the variance of the jth column in the sample matrix, E (j) representing the mean of the elements of the ith column in the sample matrix, x (j) representing the mean of the elements of the ith column in the sample matrixiRepresents the first in the sample matrixi the value of the column element.
Preferably, the step 4: and performing data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained, wherein the data cleaning comprises the following steps:
step 4.1: arranging the optimal influence factors at each moment in a mode of sequentially increasing the electric load to obtain an influence parameter sequence;
step 4.2: sequentially calculating similarity coefficients of the current influence parameter sequence and a previous group of influence parameter sequences;
step 4.3: judging whether the similarity coefficient is in a preset range or not;
step 4.4: if the similarity coefficient is not in the preset range, removing the corresponding influence parameter sequence;
step 4.5: if the similarity coefficient is within the preset range, the corresponding influence parameter sequence is reserved until all the influence parameter sequences are traversed, and the sample to be trained is obtained.
Preferably, the similarity coefficient calculation formula is as follows:
wherein p isX,YCov (X, Y) represents the covariance, σ, between the current influencing parameter sequence X and the previous influencing parameter sequence Y for the similarity coefficientXRepresenting the variance, σ, of the current influencing parameter sequence XYRepresenting the variance of the previous set of influencing parameter sequences Y.
The invention also provides an electrical load prediction device, comprising:
the sample acquisition module is used for acquiring historical training samples; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and region;
the influence factor screening module is used for primarily screening the influence factors to obtain the power load under the relevant influence factors;
the influence weight calculation module is used for calculating the influence weights of the relevant influence factors and sequencing the influence weights to obtain the power load under the optimal influence factor;
the data cleaning module is used for carrying out data cleaning on the electricity utilization load under the optimal influence factor to obtain a sample to be trained;
the training module is used for inputting the sample to be trained into a neural network for training to obtain an electric load prediction model;
and the prediction module is used for predicting the electric load at the current moment by using the electric load prediction model.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention relates to a method and a device for predicting an electric load, comprising the following steps: preliminarily screening the influence factors to obtain the power load under the relevant influence factors; calculating the influence weight of the related influence factors and sequencing to obtain the power load under the optimal influence factor; carrying out data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained; inputting a sample to be trained into a neural network for training to obtain an electric load prediction model; and predicting the electrical load at the current moment by using the electrical load prediction model. Compared with the prior art, the method and the device have the advantages that the optimal influence factors are obtained by screening and sorting the influence factors of the power load, and then the power load prediction model is obtained based on the optimal influence factors, so that the dimensionality of data is reduced, and the training speed and precision of the neural network are improved.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of an electrical load prediction method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of an electrical load prediction apparatus according to an embodiment of the present invention.
Detailed Description
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", and the like, indicate orientations and positional relationships based on those shown in the drawings, and are used only for convenience of description and simplicity of description, and do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be considered as limiting the present invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to implicitly indicate the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood according to specific situations by those of ordinary skill in the art.
The embodiment of the invention aims to provide an electric load prediction method and an electric load prediction device, which are used for predicting an electric load in real time.
Referring to fig. 1, a method for predicting an electrical load includes:
step 1: acquiring a historical training sample; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and regions;
step 2: preliminarily screening the influence factors to obtain the power load under the relevant influence factors;
in actual production activities, the electrical load is often associated with different influence factors, for example, the resident population of a certain community is increased, and the electrical load is increased accordingly. Therefore, if the electrical load is to be predicted, the influence factors of the electrical load are first analyzed and summarized.
Further, the step 2 comprises:
step 2.1: extracting corresponding power loads under different influence factors according to the time sequence to form an influence factor time sequence and a power load time sequence;
step 2.2: performing difference processing on the influence factor time sequence and the power load time sequence to obtain a difference sequence;
step 2.3: obtaining a distance coefficient of an influence factor at any time according to the difference sequence;
in the invention, the calculation formula of the distance coefficient of the influencing factor at any time is as follows:
in the formula, r0j(k) Distance coefficient, Δ, representing the influencing factor at time kj(k)=|x0(k)-xj(k)|,j=1,2,...,n,x0(k) Values representing the influencing factors at time k, xj(k) A value, delta, representing the electrical load at time kj(k) The k-th numerical value in the difference sequence is represented, M represents the minimum value in the difference sequence, M represents the maximum value in the difference sequence, and xi represents a preset value.
Step 2.4: obtaining the correlation degree of the corresponding influence factor according to the distance coefficient;
specifically, step 2.4 includes:
the formula is adopted:
obtaining the correlation of the influence factors; where θ represents the degree of correlation of the influencing factors.
Step 2.5: and taking the influence factors with the correlation degree larger than a preset threshold value as correlation influence factors.
According to the method, the relevance of each influence factor is quantified based on the distance coefficient, and the influence factors with high relevance to the power load are preliminarily screened out based on the distance coefficient, so that the relationship among the influence factors is balanced, and the influence factors with low relevance are prevented from influencing the final prediction effect.
And step 3: calculating the influence weight of the related influence factors and sequencing to obtain the power load under the optimal influence factor;
although the influence factors after being screened can reflect some information of the researched problem to different degrees, the influence factors have certain correlation with each other, so that the information reflected by the screened data is overlapped to a certain degree. For example, seasons generally have correlation with temperature, so the method needs to analyze the screened influence factors and sequence the influence factors with the maximum correlation with the power load.
Further, step 3 comprises:
step 3.1: constructing a sample matrix according to the relevant influence factors and the corresponding electric loads;
in the embodiment of the present invention, the sample matrix is:
wherein x isnpIndicating the value of the p-th relevant influencing factor at the time n.
Step 3.2: constructing an influence weight matrix according to the sample matrix;
specifically, step 3.2 includes:
step 3.2.1: calculating the variance and covariance of each column of elements in the sample matrix;
step 3.2.2: constructing an influence weight matrix according to the variance and covariance of each column of elements; the construction formula of the influence weight matrix is as follows:
wherein r isNMRepresenting the values of the elements of the influence weight matrix in the Nth row and the Mth column, cov (i, j) representing the covariance of the ith column and the jth column in the sample matrix, var (i) representing the variance of the ith column in the sample matrix, var (j) representing the variance of the jth column in the sample matrix, E (j) representing the mean of the elements of the ith column in the sample matrix, x (j) representing the mean of the elements of the ith column in the sample matrixiRepresenting the value of the ith column element in the sample matrix.
Step 3.3: calculating the influence weights of the relevant influence factors according to the characteristic values of the influence weight matrix and sorting the influence weights;
step 3.4: and selecting the related influence factor corresponding to the influence weight with the highest rank as the optimal influence factor.
According to the method, the optimal influence factors are selected based on the characteristic values of the influence weight matrix, so that the modeling process of a subsequent power load prediction model is simplified, and the training speed and the training precision are improved.
In data modeling, the accuracy of the model depends on the number of hidden layers of the neural network, the number of neurons per layer, the activation function, the training period, and the like, but the accuracy of modeling is more affected by the precision of the data. The data modeling itself learns and finds rules from historical data, and if learning is done from erroneous data, the accuracy of the modeling is not mentioned. In the data acquisition process, the sensor is difficult to avoid being influenced by unexpected factors, so that data at a certain moment deviates from a true value, a coarse error is generated, and abnormal peak-valley fluctuation occurs. Or the sensor fails and data is not acquired within a certain period of time, so that data is lost. Therefore, the data is processed before the model is trained.
And 4, step 4: carrying out data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained;
further, step 4 comprises:
step 4.1: arranging the optimal influence factors at each moment in a mode of sequentially increasing the electric load to obtain an influence parameter sequence;
step 4.2: sequentially calculating similarity coefficients of the current influence parameter sequence and the previous group of influence parameter sequences; wherein, the similarity coefficient calculation formula is as follows:
wherein p isX,YCov (X, Y) represents the covariance, σ, between the current influencing parameter sequence X and the previous influencing parameter sequence Y for the similarity coefficientXRepresenting the variance, σ, of the current influencing parameter sequence XYRepresenting the variance of the previous set of influencing parameter sequences Y.
Step 4.3: judging whether the similarity coefficient is in a preset range or not;
step 4.4: if the similarity coefficient is not in the preset range, removing the corresponding influence parameter sequence;
step 4.5: if the similarity coefficient is within the preset range, the corresponding influence parameter sequence is reserved until all the influence parameter sequences are traversed, and the sample to be trained is obtained.
According to the method, the similarity coefficient calculation formula is constructed through the covariance, and then the influence parameter sequences which do not meet the requirements are removed on the basis of the similarity coefficient calculation formula, so that the authenticity of data can be ensured.
And 5: inputting the sample to be trained into a neural network for training to obtain an electrical load prediction model;
it should be noted that the neural network of the present invention may be a radial basis function neural network, and the electrical load prediction model of the present invention is obtained by training with the electrical load as an output, with the temperature value at the current time (i.e., the temperature value is selected as the optimal influencing factor in the present invention, and may be set as another value in practical application) as an input. In addition, the activation function of the radial basis function neural network in the training process is as follows:
wherein x isT=[x1,x2,…xn]TInput vector, x, representing a neural networkmRepresenting the nth input vector of the neural network, ciRepresenting the output, σ, of the ith node of the hidden layeriRepresenting the width of the ith basis function.
Step 6: and predicting the electrical load at the current moment by using the electrical load prediction model.
According to the invention, the optimal influence factors are obtained by screening and sequencing the influence factors of the electric load, and then the electric load prediction model is obtained based on the optimal influence factors, so that the dimensionality of data is reduced, and the training speed and precision of the neural network are improved.
Referring to fig. 2, the present invention further provides an electrical load prediction apparatus, including:
the sample acquisition module is used for acquiring historical training samples; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and regions;
the influence factor screening module is used for primarily screening the influence factors to obtain the power load under the relevant influence factors;
the influence weight calculation module is used for calculating the influence weights of the related influence factors and sequencing the calculated influence weights to obtain the power load under the optimal influence factor;
the data cleaning module is used for carrying out data cleaning on the electricity utilization load under the optimal influence factor to obtain a sample to be trained;
the training module is used for inputting the sample to be trained into a neural network for training to obtain an electric load prediction model;
and the prediction module is used for predicting the electric load at the current moment by using the electric load prediction model.
Compared with the prior art, the beneficial effects of the power load prediction device provided by the invention are the same as those of the power load prediction method in the technical scheme, and are not repeated herein.
The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of changes or substitutions within the technical scope of the present invention, and the present invention shall be covered by the claims. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (10)
1. An electrical load prediction method, comprising:
step 1: acquiring a historical training sample; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and regions;
step 2: preliminarily screening the influence factors to obtain the power load under the relevant influence factors;
and step 3: calculating the influence weight of the related influence factors and sequencing to obtain the power load under the optimal influence factor;
and 4, step 4: carrying out data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained;
and 5: inputting the sample to be trained into a neural network for training to obtain an electric load prediction model;
and 6: and predicting the electrical load at the current moment by using the electrical load prediction model.
2. The method for predicting the electrical load according to claim 1, wherein the step 2: preliminarily screening the influence factors to obtain the power load under the relevant influence factors, wherein the power load comprises the following steps:
step 2.1: extracting corresponding electric loads under different influence factors according to the time sequence to form an influence factor time sequence and an electric load time sequence;
step 2.2: performing difference processing on the influence factor time sequence and the power load time sequence to obtain a difference sequence;
step 2.3: obtaining a distance coefficient of an influence factor at any time according to the difference sequence;
step 2.4: obtaining the correlation degree of the corresponding influence factor according to the distance coefficient;
step 2.5: and taking the influence factors with the correlation degree larger than a preset threshold value as correlation influence factors.
3. The method according to claim 2, wherein the distance coefficient of the influencing factor at any time is calculated as:
wherein r is0j(k) Distance coefficient, Δ, representing the influencing factor at time kj(k)=|x0(k)-xj(k)|,j=1,2,...,n,x0(k) A value representing the influencing factor at time k, xj(k) A value, delta, representing the electrical load at time kj(k) The k-th numerical value in the difference sequence is represented, M represents the minimum value in the difference sequence, M represents the maximum value in the difference sequence, and xi represents a preset value.
4. A method according to claim 3, wherein said step 2.4: obtaining the correlation of the corresponding influence factors according to the distance coefficient, including:
the formula is adopted:
obtaining the correlation of the influence factors; where θ represents the degree of correlation of the influencing factors.
5. The method for predicting the electrical load according to claim 1, wherein the step 3: calculating the influence weight of the related influence factors and sequencing the influence weights to obtain the power load under the optimal influence factor, wherein the method comprises the following steps:
step 3.1: constructing a sample matrix according to the relevant influence factors and the corresponding electric loads;
step 3.2: constructing an influence weight matrix according to the sample matrix;
step 3.3: calculating the influence weight of the relevant influence factors according to the characteristic value of the influence weight matrix and sequencing;
step 3.4: and selecting the relevant influence factor corresponding to the influence weight with the highest ranking as the optimal influence factor.
7. The method for forecasting the electric load according to claim 6, wherein the step 3.2: constructing an influence weight matrix according to the sample matrix, comprising:
step 3.2.1: calculating the variance and covariance of each column of elements in the sample matrix;
step 3.2.2: constructing an influence weight matrix according to the variance and covariance of each column of elements; the construction formula of the influence weight matrix is as follows:
wherein r isNMRepresenting the values of the elements of the influence weight matrix at the Nth row and the Mth column, cov (i, j) representing the covariance of the ith column and the jth column in the sample matrix, var (i) representing the variance of the ith column in the sample matrix, var (j) representing the variance of the jth column in the sample matrix, E (j) representing the mean of the ith column elements in the sample matrix, and xiRepresenting the value of the ith column element in the sample matrix.
8. The method for predicting the electrical load according to claim 7, wherein the step 4: and performing data cleaning on the power utilization load under the optimal influence factor to obtain a sample to be trained, wherein the data cleaning comprises the following steps:
step 4.1: arranging the optimal influence factors at each moment in a mode of sequentially increasing the power load to obtain an influence parameter sequence;
and 4.2: sequentially calculating similarity coefficients of the current influence parameter sequence and the previous group of influence parameter sequences;
step 4.3: judging whether the similarity coefficient is in a preset range or not;
step 4.4: if the similarity coefficient is not in the preset range, removing the corresponding influence parameter sequence;
step 4.5: if the similarity coefficient is within the preset range, the corresponding influence parameter sequence is reserved until all the influence parameter sequences are traversed, and the sample to be trained is obtained.
9. The method for predicting the electrical load according to claim 8, wherein the similarity coefficient calculation formula is as follows:
wherein p isX,YCov (X, Y) represents the covariance, σ, between the current influencing parameter sequence X and the previous influencing parameter sequence Y for the similarity coefficientXRepresenting the variance, σ, of the current influencing parameter sequence XYRepresenting the variance of the previous set of influencing parameter sequences Y.
10. An electrical load prediction apparatus, comprising:
the sample acquisition module is used for acquiring historical training samples; the historical training sample is the power load of the object to be predicted under different influence factors in a preset time period; the influencing factors comprise electricity price, population, temperature, humidity, season, electric appliance quantity, economic income and region;
the influence factor screening module is used for primarily screening the influence factors to obtain the power load under the relevant influence factors;
the influence weight calculation module is used for calculating the influence weights of the related influence factors and sequencing the calculated influence weights to obtain the power load under the optimal influence factor;
the data cleaning module is used for carrying out data cleaning on the electricity utilization load under the optimal influence factor to obtain a sample to be trained;
the training module is used for inputting the sample to be trained into a neural network for training to obtain an electric load prediction model;
and the prediction module is used for predicting the electric load at the current moment by using the electric load prediction model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210746657.7A CN115275977A (en) | 2022-06-28 | 2022-06-28 | Power load prediction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210746657.7A CN115275977A (en) | 2022-06-28 | 2022-06-28 | Power load prediction method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115275977A true CN115275977A (en) | 2022-11-01 |
Family
ID=83762859
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210746657.7A Pending CN115275977A (en) | 2022-06-28 | 2022-06-28 | Power load prediction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115275977A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116702978A (en) * | 2023-06-07 | 2023-09-05 | 西安理工大学 | Electric vehicle charging load prediction method and device considering emergency characteristics |
-
2022
- 2022-06-28 CN CN202210746657.7A patent/CN115275977A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116702978A (en) * | 2023-06-07 | 2023-09-05 | 西安理工大学 | Electric vehicle charging load prediction method and device considering emergency characteristics |
CN116702978B (en) * | 2023-06-07 | 2024-02-13 | 西安理工大学 | Electric vehicle charging load prediction method and device considering emergency characteristics |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110597240B (en) | Hydroelectric generating set fault diagnosis method based on deep learning | |
Lawrence | Using neural networks to forecast stock market prices | |
Wang et al. | Data-driven mode identification and unsupervised fault detection for nonlinear multimode processes | |
CN109359665B (en) | Household appliance load identification method and device based on support vector machine | |
CN113505923B (en) | Regional power grid short-term load prediction method and system | |
CN116380445B (en) | Equipment state diagnosis method and related device based on vibration waveform | |
CN114676742A (en) | Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network | |
CN112836604A (en) | Rolling bearing fault diagnosis and classification method, system and equipment based on VMD-SSAE and storage medium thereof | |
CN116167527B (en) | Pure data-driven power system static safety operation risk online assessment method | |
CN115841278B (en) | Method, system, equipment and medium for evaluating running error state of electric energy metering device | |
CN113447879A (en) | Electric energy meter metering error prediction method, system and device based on multi-stress influence | |
CN111126868A (en) | Road traffic accident occurrence risk determination method and system | |
CN112414715A (en) | Bearing fault diagnosis method based on mixed feature and improved gray level co-occurrence algorithm | |
CN116169670A (en) | Short-term non-resident load prediction method and system based on improved neural network | |
CN115275977A (en) | Power load prediction method and device | |
CN117473411A (en) | Bearing life prediction method based on improved transducer model | |
CN113151842B (en) | Method and device for determining conversion efficiency of wind-solar complementary water electrolysis hydrogen production | |
CN113837434A (en) | Solar photovoltaic power generation prediction method and device, electronic equipment and storage medium | |
CN116776245A (en) | Three-phase inverter equipment fault diagnosis method based on machine learning | |
CN115034307A (en) | Vibration data self-confirmation method, system and terminal for multi-parameter self-adaptive fusion | |
CN113112166A (en) | Equipment state variable selection method and equipment based on gray fuzzy hierarchical analysis | |
CN116776134A (en) | Photovoltaic output prediction method based on PCA-SFFS-BiGRU | |
JAYAGOWRI et al. | ENHANCEMENT OF FCM FOR FORECASTING AND DECISION SUPPORT SYSTEM | |
CN112699920A (en) | Method and system for determining main influence factors of line loss rate of passive station area | |
CN115660154A (en) | LSTM-CGAN-based power load data prediction method |
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 |