Electrical load prediction method and system based on carbon emission factors
By dividing the area to be predicted into sub-regions, building a characteristic matrix of electricity load and carbon emission data and inputting it into a prediction model, the problem that carbon emission factors in the prior art are not considered is solved, and the accuracy of electricity load prediction is improved.
Patent Information
- Application Number
- CN202410995176.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-12
AI Technical Summary
The existing power load prediction methods lack the impact of carbon emission factors on power load prediction, resulting in low prediction accuracy.
By dividing the area to be predicted into multiple sub-regions, collecting electricity load data and carbon emission data, building a feature matrix and inputting it into the electricity load prediction model, establishing a nonlinear mapping relationship between electricity load and carbon emission data. Forecasting the power load in the future time period.
The prediction accuracy of electricity load is improved, and the accuracy of prediction is enhanced by considering carbon emission factors.
Smart Images

Figure CN118944074A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electricity load forecasting, and more specifically, to a method and system for forecasting electricity load based on a carbon emission factor. Background Art
[0002] Electricity load forecasting is a core component of power system planning, operation and management. Its purpose is to estimate the power demand of the power grid in the future time period based on the existing power consumption data, ensure the balance of power supply and demand, and formulate reasonable energy policies. Existing power load forecasting methods mainly include: time series analysis and regression analysis; the time series analysis method constructs a mathematical model based on historical power load data to mine the periodic change trend within the power load data, so as to predict the power demand of the power grid in the future time period; the regression analysis method takes the power load as the dependent variable, selects factors that have a significant impact on the power load (weather conditions, holidays, population density, etc.) as independent variables, and mines the nonlinear mapping relationship between the two by establishing a mathematical model, so as to predict the power demand of the power grid in the future time period.
[0003] The carbon emission factor refers to the greenhouse gas emissions corresponding to unit energy consumption. It reflects the carbon emission intensity of electricity production and use of different energy types (such as thermal power, hydropower, wind power, photovoltaics, etc.), and can evaluate the impact of different energy types on the environment. In addition, the carbon emission factor can also indirectly reflect users' electricity consumption behavior and carbon emission policies.
[0004] However, existing electricity load forecasting methods usually lack consideration of the impact of carbon emission factors on electricity load forecasting, resulting in low prediction accuracy. Summary of the invention
[0005] The present invention provides a method and system for predicting electricity load based on carbon emission factors, which solves the technical problem that the existing electricity load prediction methods in the related art usually lack consideration of the impact of carbon emission factors on electricity load prediction, resulting in low prediction accuracy.
[0006] The present invention provides a method for predicting electricity load based on carbon emission factors, comprising the following steps:
[0007] Step S101, dividing the area to be predicted into M sub-areas;
[0008] Step S102, within a preset time period T, collecting power load data and carbon emission data of M sub-areas at preset time intervals t;
[0009] The power load data includes: power consumption, power type, maximum voltage value, minimum voltage value, average voltage value, maximum power consumption, minimum power consumption, average power consumption, maximum temperature value, minimum temperature value and average temperature value;
[0010] Carbon emission data include: energy type and carbon emission amount;
[0011] Step S103, preprocessing the power load data and carbon emission data in step S102, and constructing a feature matrix; the i-th row and j-th column of the feature matrix represent the preprocessed power load data and carbon emission data of the i-th sub-region at the j-th time point, where 1≤i≤M, 1≤j≤N, N represents the total number of time points, and N=T / t;
[0012] Step S104, inputting the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
[0013] Furthermore, M, T, t and G are all custom parameters.
[0014] Furthermore, the electricity consumption types include: residential electricity, commercial electricity and industrial electricity, and the electricity consumption types are represented by positive integer codes; the energy types include: coal, natural gas and renewable energy, and the energy types are represented by positive integer codes.
[0015] Furthermore, the carbon emission is equal to the electricity consumption multiplied by the carbon emission coefficient corresponding to the energy type, where the carbon emission coefficient is obtained by querying a relevant database.
[0016] Further, the preprocessing comprises the following steps:
[0017] Step S201, when there are missing values in the power load data and the carbon emission data, the average value of the previous element value and the next element value of the missing value is taken for filling;
[0018] Step S202, deleting abnormal values in the power load data and carbon emission data;
[0019] Step S203: normalize each element value in the power load data and the carbon emission data.
[0020] Furthermore, the calculation formula for normalization processing is as follows:
[0021]
[0022] Where 1≤u≤R, 1≤v≤M*N, R represents the total number of elements of electricity load data and carbon emission data, that is, R is equal to 13, Indicates the normalized value of the vth element of the uth element. Indicates the value of the vth element of the uth element, min u and max u Respectively represent the minimum and maximum value of the u-th element.
[0023] Further, the power load prediction model includes N hidden layers, each hidden layer includes a first hidden unit and a second hidden unit;
[0024] The first hidden unit includes M channels;
[0025] The i-th channel of the first hidden unit of the j-th hidden layer inputs the element value of the i-th row of the j-th column of the feature matrix, that is, the preprocessed power load data and carbon emission data of the i-th sub-region at the j-th time point, and the i-th channel of the first hidden unit of the j-th hidden layer outputs the intermediate feature;
[0026] The first hidden unit of the jth hidden layer outputs the first feature;
[0027] The second hidden unit of the j-th hidden layer inputs the first feature output by the first hidden unit of the j-th hidden layer, and the second hidden unit of the j-th hidden layer outputs the second feature;
[0028] The second feature output by the second hidden unit of the Nth hidden layer is input into the classifier, and the classification space of the classifier represents the power load of the area to be predicted in the future time period G.
[0029] Furthermore, the calculation formula of the jth hidden layer of the power load forecasting model includes:
[0030] The calculation formula for the first hidden unit of the jth hidden layer includes:
[0031]
[0032] Among them, H j represents the first feature of the output of the first hidden unit of the jth hidden layer, represents the intermediate feature of the ith channel output of the first hidden unit of the jth hidden layer, Represents the element value of the j-th column and the i-th row of the feature matrix of the i-th channel input to the first hidden unit of the j-th hidden layer, and They represent the weight parameter and bias parameter of the ith channel of the first hidden unit of the jth hidden layer, respectively, and concat M It means to concatenate the intermediate features output by M channels, and ReLU means the ReLU activation function;
[0033] The calculation formula for the second hidden unit of the jth hidden layer includes:
[0034]
[0035] where F j ,I j , C j , O j and denote the forget gate, input gate, intermediate state, intermediate hidden state, output gate and second feature of the second hidden unit of the jth hidden layer, respectively. f1 , W f2 and b f denote the first weight parameter, the second weight parameter and the bias parameter of the forget gate of the second hidden unit of the jth hidden layer, respectively, and W I1 , W I2 and b I denote the first weight parameter, the second weight parameter and the bias parameter of the input gate of the second hidden unit of the jth hidden layer, respectively, and W C1 , W C2 and b C denote the first weight parameter, the second weight parameter and the bias parameter of the intermediate state of the second hidden unit of the jth hidden layer, respectively, and W O1 , W O2 and b O denote the first weight parameter, the second weight parameter and the bias parameter of the output gate of the second hidden unit of the jth hidden layer, respectively. j represents the first feature of the output of the first hidden unit of the jth hidden layer, represents the second feature of the output of the second hidden unit of the jth hidden layer, C j-1 represents the intermediate hidden state of the j-1th hidden layer, ⊙ represents point-by-point multiplication, sigmoid represents the sigmoid activation function, tanh represents the hyperbolic tangent activation function, C0=0.
[0036] Furthermore, the historical power load of the area to be predicted within the time period G is collected as sample labels of training samples for training the power load prediction model.
[0037] The present invention provides a power load forecasting system based on carbon emission factors, comprising:
[0038] A sub-region division module, which is used to divide the area to be predicted into M sub-regions;
[0039] A data collection module, which is used to collect power load data and carbon emission data of M sub-areas within a preset time period T and at a preset time interval t;
[0040] A feature matrix construction module is used to preprocess the electricity load data and carbon emission data and construct a feature matrix;
[0041] The power load prediction module is used to input the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
[0042] The beneficial effect of the present invention is that the present invention improves the prediction accuracy of the power load by collecting power load data and carbon emission data of multiple sub-areas of the area to be predicted and establishing a nonlinear mapping relationship between the power load and the power load through a power load prediction model. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 It is a flow chart of a method for predicting electricity load based on carbon emission factors of the present invention;
[0044] Figure 2 is a flow chart of the pre-processing of the present invention;
[0045] Figure 3 It is a schematic diagram of an electricity load forecasting system based on carbon emission factors of the present invention.
[0046] In the figure: sub-area division module 301, data collection module 302, feature matrix construction module 303, power load prediction module 304. DETAILED DESCRIPTION
[0047] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.
[0048] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0049] like Figure 1-Figure 3 As shown, a method for predicting electricity load based on carbon emission factors includes the following steps:
[0050] Step S101, dividing the area to be predicted into M sub-areas;
[0051] Step S102, within a preset time period T, collecting power load data and carbon emission data of M sub-areas at preset time intervals t;
[0052] The power load data includes: power consumption, power type, maximum voltage value, minimum voltage value, average voltage value, maximum power consumption, minimum power consumption, average power consumption, maximum temperature value, minimum temperature value and average temperature value;
[0053] Carbon emission data include: energy type and carbon emission amount;
[0054] Step S103, preprocessing the power load data and carbon emission data in step S102, and constructing a feature matrix;
[0055] The i-th row and j-th column of the feature matrix represent the preprocessed electricity load data and carbon emission data of the i-th sub-region at the j-th time point, where 1≤i≤M, 1≤j≤N, N represents the total number of time points, N=T / t;
[0056] Step S104, inputting the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
[0057] In one embodiment of the present invention, M, T, t and G are all custom parameters. Preferably, the default value of M is set to 20, the default value of T is set to 7 days, the default value of t is set to 1 day, and the default value of G is set to 1 day.
[0058] In one embodiment of the present invention, the power load data may further include: maximum wind speed, minimum wind speed, average wind speed, rainfall, whether it is a holiday and population size, etc.
[0059] In one embodiment of the present invention, electricity types include: residential electricity, commercial electricity and industrial electricity, and the electricity types are represented by positive integer codes, for example, the residential electricity type is represented by the number 1; energy types include: coal, natural gas and renewable energy, and the energy types are represented by positive integer codes, for example, the natural gas energy type is represented by the number 2.
[0060] In one embodiment of the present invention, carbon emissions are equal to electricity consumption multiplied by the carbon emission coefficient corresponding to the energy type, where the carbon emission coefficient is obtained by querying relevant databases, such as the China Carbon Accounting Database, the National Greenhouse Gas Inventory Guidelines, etc.
[0061] In one embodiment of the present invention, Figure 2 As shown, the preprocessing includes the following steps:
[0062] Step S201, when there are missing values in the power load data and the carbon emission data, the average value of the previous element value and the next element value of the missing value is taken for filling;
[0063] Step S202, deleting abnormal values in the power load data and carbon emission data;
[0064] Step S203: normalize each element value in the power load data and the carbon emission data.
[0065] In one embodiment of the present invention, the calculation formula for normalization processing is as follows:
[0066]
[0067] Where 1≤u≤R, 1≤v≤M*N, R represents the total number of elements of electricity load data and carbon emission data, that is, R is equal to 13, Indicates the normalized value of the vth element of the uth element. Indicates the value of the vth element of the uth element, min u and max u Respectively represent the minimum and maximum value of the u-th element.
[0068] In one embodiment of the present invention, the power load prediction model includes N hidden layers, each hidden layer includes a first hidden unit and a second hidden unit;
[0069] The first hidden unit includes M channels;
[0070] The i-th channel of the first hidden unit of the j-th hidden layer inputs the element value of the i-th row of the j-th column of the feature matrix, that is, the preprocessed power load data and carbon emission data of the i-th sub-region at the j-th time point, and the i-th channel of the first hidden unit of the j-th hidden layer outputs the intermediate feature;
[0071] The first hidden unit of the jth hidden layer outputs the first feature;
[0072] The second hidden unit of the j-th hidden layer inputs the first feature output by the first hidden unit of the j-th hidden layer, and the second hidden unit of the j-th hidden layer outputs the second feature;
[0073] The second feature output by the second hidden unit of the Nth hidden layer is input into the classifier, and the classification space of the classifier represents the power load of the area to be predicted in the future time period G.
[0074] In one embodiment of the present invention, the calculation formula of the jth hidden layer of the power load prediction model includes:
[0075] The calculation formula for the first hidden unit of the jth hidden layer includes:
[0076]
[0077] Where Hj represents the first feature output by the first hidden unit of the jth hidden layer, represents the intermediate feature of the ith channel output of the first hidden unit of the jth hidden layer, Represents the element value of the j-th column and the i-th row of the feature matrix of the i-th channel input to the first hidden unit of the j-th hidden layer, and They represent the weight parameter and bias parameter of the ith channel of the first hidden unit of the jth hidden layer, respectively, and concat M It means to concatenate the intermediate features output by M channels, and ReLU means the ReLU activation function;
[0078] The calculation formula for the second hidden unit of the jth hidden layer includes:
[0079]
[0080] where F j ,I j , C j , O j and denote the forget gate, input gate, intermediate state, intermediate hidden state, output gate and second feature of the second hidden unit of the jth hidden layer, respectively. f1 , Wf2 and b f denote the first weight parameter, the second weight parameter and the bias parameter of the forget gate of the second hidden unit of the jth hidden layer, respectively, and W I1 , W I2 and b I denote the first weight parameter, the second weight parameter and the bias parameter of the input gate of the second hidden unit of the jth hidden layer, respectively, and W C1 , W C2 and b C denote the first weight parameter, the second weight parameter and the bias parameter of the intermediate state of the second hidden unit of the jth hidden layer, respectively, and W O1 , W O2 and b O denote the first weight parameter, the second weight parameter and the bias parameter of the output gate of the second hidden unit of the jth hidden layer, respectively. j represents the first feature of the output of the first hidden unit of the jth hidden layer, represents the second feature of the output of the second hidden unit of the jth hidden layer, C j-1 represents the intermediate hidden state of the j-1th hidden layer, ⊙ represents point-by-point multiplication, sigmoid represents the sigmoid activation function, tanh represents the hyperbolic tangent activation function, C0=0.
[0081] In one embodiment of the present invention, the historical power load of the area to be predicted within the time period G is collected as sample labels of training samples for training the power load prediction model.
[0082] It should be noted that the difference between the output value of the electricity load prediction model and the sample label is specified as the loss function. The gradient information of the weight parameters and bias parameters of the electricity load prediction model is first calculated by the back propagation algorithm, and then the weight parameters and bias parameters of the electricity load prediction model are updated according to the gradient information by the gradient descent method, so as to iteratively optimize the model until the loss function is minimized. The training of the neural network model is a conventional technical means and will not be elaborated here.
[0083] In one embodiment of the present invention, the output power of the power grid generator set can be adjusted according to the power load of the area to be predicted in the future time period G to ensure that the power grid frequency is maintained at 50 Hz.
[0084] In one embodiment of the present invention, Figure 3 As shown, a power load forecasting system based on carbon emission factors includes:
[0085] A sub-region division module 301 is used to divide the region to be predicted into M sub-regions;
[0086] A data collection module 302 is used to collect power load data and carbon emission data of M sub-areas within a preset time period T and at a preset time interval t;
[0087] A feature matrix construction module 303 is used to pre-process the power load data and the carbon emission data and construct a feature matrix;
[0088] The power load prediction module 304 is used to input the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
[0089] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.
Claims
1. A method for predicting electricity load based on carbon emission factors, characterized in that: The following steps are involved: Step S101, dividing the area to be predicted into M sub-areas; Step S102, within a preset time period T, collecting power load data and carbon emission data of M sub-areas at preset time intervals t; The power load data includes: power consumption, power type, maximum voltage value, minimum voltage value, average voltage value, maximum power consumption, minimum power consumption, average power consumption, maximum temperature value, minimum temperature value and average temperature value; Carbon emission data include: energy type and carbon emission amount; Step S103, preprocessing the power load data and carbon emission data in step S102, and constructing a feature matrix; the i-th row and j-th column of the feature matrix represent the preprocessed power load data and carbon emission data of the i-th sub-region at the j-th time point, where 1≤i≤M, 1≤j≤N, N represents the total number of time points, and N=T / t; Step S104, inputting the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
2. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: M, T, t, and G are custom parameters.
3. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: Electricity consumption types include: residential electricity, commercial electricity and industrial electricity, and electricity consumption types are represented by positive integer codes; energy types include: coal, natural gas and renewable energy, and energy types are represented by positive integer codes.
4. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: Carbon emissions are equal to electricity consumption multiplied by the carbon emission coefficient corresponding to the energy type, where the carbon emission coefficient is obtained by querying the relevant database.
5. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: The preprocessing includes the following steps: Step S201, when there are missing values in the power load data and the carbon emission data, the average value of the previous element value and the next element value of the missing value is taken for filling; Step S202, deleting abnormal values in the power load data and carbon emission data; Step S203: normalize each element value in the power load data and the carbon emission data.
6. The method for predicting electricity load based on carbon emission factor according to claim 5, characterized in that: The calculation formula for normalization is as follows: Where 1≤u≤R, 1≤v≤M*N, R represents the total number of elements of electricity load data and carbon emission data, that is, R is equal to 13, Indicates the normalized value of the vth element of the uth element. Indicates the value of the vth element of the uth element, min u and max u Respectively represent the minimum and maximum value of the u-th element.
7. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: The power load prediction model includes N hidden layers, each hidden layer includes a first hidden unit and a second hidden unit; The first hidden unit includes M channels; The i-th channel of the first hidden unit of the j-th hidden layer inputs the element value of the i-th row of the j-th column of the feature matrix, that is, the preprocessed power load data and carbon emission data of the i-th sub-region at the j-th time point, and the i-th channel of the first hidden unit of the j-th hidden layer outputs the intermediate feature; The first hidden unit of the jth hidden layer outputs the first feature; The second hidden unit of the j-th hidden layer inputs the first feature output by the first hidden unit of the j-th hidden layer, and the second hidden unit of the j-th hidden layer outputs the second feature; The second feature output by the second hidden unit of the Nth hidden layer is input into the classifier, and the classification space of the classifier represents the power load of the area to be predicted in the future time period G.
8. The method for predicting electricity load based on carbon emission factor according to claim 7, characterized in that: The calculation formula of the jth hidden layer of the power load forecasting model includes: The calculation formula for the first hidden unit of the jth hidden layer includes: Where Hj represents the first feature output by the first hidden unit of the jth hidden layer, represents the intermediate feature of the ith channel output of the first hidden unit of the jth hidden layer, Represents the element value of the j-th column and the i-th row of the feature matrix of the i-th channel input to the first hidden unit of the j-th hidden layer, and They represent the weight parameter and bias parameter of the ith channel of the first hidden unit of the jth hidden layer, respectively, and concat M It means to concatenate the intermediate features output by M channels, and ReLU means the ReLU activation function; The calculation formula for the second hidden unit of the jth hidden layer includes: Among them, F j ,I j , C j , O j and denote the forget gate, input gate, intermediate state, intermediate hidden state, output gate and second feature of the second hidden unit of the jth hidden layer, respectively. f1 , W f2 and b f denote the first weight parameter, the second weight parameter and the bias parameter of the forget gate of the second hidden unit of the jth hidden layer, respectively, and W I1 , W I2 and b I denote the first weight parameter, the second weight parameter and the bias parameter of the input gate of the second hidden unit of the jth hidden layer, respectively, and W C1 , W C2 and b C denote the first weight parameter, the second weight parameter and the bias parameter of the intermediate state of the second hidden unit of the jth hidden layer, respectively, and W O1 , W O2 and b O denote the first weight parameter, the second weight parameter and the bias parameter of the output gate of the second hidden unit of the jth hidden layer, respectively. j represents the first feature of the output of the first hidden unit of the jth hidden layer, represents the second feature of the output of the second hidden unit of the jth hidden layer, C j-1 represents the intermediate hidden state of the j-1th hidden layer, ⊙ represents point-by-point multiplication, sigmoid represents the sigmoid activation function, tanh represents the hyperbolic tangent activation function, C0=0.
9. The method for predicting electricity load based on carbon emission factor according to claim 1, characterized in that: The historical power load of the area to be predicted within the time period G is collected as sample labels of training samples for training the power load prediction model.
10. A power load forecasting system based on carbon emission factors, using a power load forecasting method based on carbon emission factors as claimed in any one of claims 1 to 9, characterized in that: include: A sub-region division module, which is used to divide the area to be predicted into M sub-regions; A data collection module, which is used to collect power load data and carbon emission data of M sub-areas within a preset time period T and at a preset time interval t; A feature matrix construction module is used to preprocess the electricity load data and carbon emission data and construct a feature matrix; The power load prediction module is used to input the characteristic matrix into the power load prediction model, and the output value represents the power load of the area to be predicted in the future time period G.
Citation Information
Cited By
Power load prediction method, device and equipment based on energy-carbon double control, and storage medium
CN121529551A
Power load prediction method and device based on energy and carbon double control, equipment and storage medium
CN121529551B
