CN115796023A - Power system scheduling method, device and equipment based on carbon quota - Google Patents
Power system scheduling method, device and equipment based on carbon quota Download PDFInfo
- Publication number
- CN115796023A CN115796023A CN202211478786.9A CN202211478786A CN115796023A CN 115796023 A CN115796023 A CN 115796023A CN 202211478786 A CN202211478786 A CN 202211478786A CN 115796023 A CN115796023 A CN 115796023A
- Authority
- CN
- China
- Prior art keywords
- data
- power
- power supply
- supply area
- neural network
- 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
Abstract
The invention discloses a power system scheduling method based on carbon quota, which comprises the following steps: the method comprises the steps that a to-be-used carbon quota and generated energy corresponding to the to-be-used carbon quota of a power plant in a future period of time of a power supply area are obtained, and the to-be-used carbon quota and the corresponding generated energy are stored in a cloud or a database in advance; acquiring power supply area data, and predicting the use electric quantity of the power supply area in a future period of time according to the power supply area data; and scheduling the power system according to the power generation amount corresponding to the power plant and the used electric quantity of the power supply area. The method and the device realize the prediction of the used electric quantity of the power supply area, and dispatch the power system according to the prediction result, thereby avoiding the occurrence of insufficient power supply.
Description
Technical Field
The invention belongs to the field of power system scheduling, and particularly relates to a power system scheduling method, device and equipment based on carbon quota.
Background
Under the carbon market environment, thermal power enterprises need to calculate and calculate the carbon quota surplus and shortage. Therefore, the carbon quota excess and deficiency analysis can be carried out, and the carbon quota of the thermal power plant is in the national regulation allowable range through operation optimization and blending of coal entering the plant. After the carbon quota of the power plant is fixed in a future period of time, the power generation amount of the power plant is also limited, and the power consumption of the power supply area of the power plant in the future period of time changes, so that the power generation amount of the power plant may be lower than the power consumption of the corresponding power supply area, and a problem of insufficient power supply is caused.
Disclosure of Invention
In order to overcome the defects in the prior art, the method, the device and the equipment for scheduling the power system based on the carbon quota solve the problems in the prior art.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that: a power system scheduling method based on carbon quota includes:
the method comprises the steps that a to-be-used carbon quota and generated energy corresponding to the to-be-used carbon quota of a power plant in a future period of time of a power supply area are obtained, and the to-be-used carbon quota and the corresponding generated energy are stored in a cloud or a database in advance;
acquiring power supply area data, and predicting the use electric quantity of the power supply area in a future period of time according to the power supply area data;
and scheduling the power system according to the power generation amount corresponding to the power plant and the used electric quantity of the power supply area.
Further, collecting power supply area data and predicting the power consumption of the power supply area in a future period of time according to the power supply area data comprises:
collecting a plurality of sample data, and preprocessing the sample data to obtain preprocessed sample data;
constructing a neural network model, and training the neural network model by adopting the preprocessed sample data to obtain a power supply area electricity consumption prediction model;
and acquiring power supply area data, and predicting the power consumption of the power supply area in a future period by taking the power supply area data as the input of a power consumption prediction model of the power supply area, wherein the sample data and the power supply area data are the same type of data.
Further, the sample data comprises seasonal data, weather data, temperature data, date type data and corresponding used electricity quantity of the power supply area at continuous time points, the seasonal data comprises spring, summer, autumn or winter, the weather data comprises rainy days, sunny days, cloudy days or snow days, and the date type data comprises working days or non-working days;
the continuous time points represent time points corresponding to a plurality of consecutive days, and the time points represent integral values of each day.
Further, preprocessing the sample data to obtain preprocessed sample data, including:
judging whether missing values exist in the seasonal data, the weather data, the temperature data, the date type data and the corresponding used electric quantity in the sample data, if so, completing the missing values, and performing normalization processing on the completed sample data to obtain the preprocessed sample data, otherwise, directly performing normalization processing on the sample data to obtain the preprocessed sample data.
Further, constructing a neural network model, and training the neural network model by using the preprocessed sample data to obtain a power consumption prediction model of a power supply area, including:
A. constructing an RBF neural network, and taking the RBF neural network as a neural network model;
B. randomly initializing a network center, a network width and a network weight of the neural network model, constructing a vector by using the network center, the network width and the network weight, taking the vector as a primary chromosome, and repeating the steps to obtain two primary chromosomes;
C. adding the two parent chromosomes into a population, wherein individuals in the population are used for representing chromosomes;
D. acquiring the fitness of each individual in the population by adopting the preprocessed sample data;
E. judging whether the fitness of the individuals in the population is greater than a preset fitness threshold value, if so, taking the individual with the maximum fitness as an optimal individual to obtain an optimal network center, an optimal network width and an optimal network weight, and entering a step G, otherwise, entering a step F;
F. carrying out variation and cross operation on individuals of the population to obtain offspring chromosomes, adding the offspring chromosomes into the population, and returning to the step D;
G. and performing secondary training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width and the optimal network weight, and taking the trained neural network model as a power consumption prediction model of a power supply area.
Further, acquiring the fitness of each individual in the population by using the preprocessed sample data includes:
taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model;
taking the corresponding use electric quantity in the preprocessed sample data as expected output data of the neural network model, and obtaining the error corresponding to the neural network model according to the output data and the expected output data as follows:
wherein E represents the corresponding error of the neural network model, D i Indicating the power consumption Y in the ith preprocessed sample data i The method comprises the steps of obtaining a neural network model according to ith preprocessed sample data, wherein the output data of the neural network model obtained according to the ith preprocessed sample data is represented by I =1,2, \ 8230;
according to the error corresponding to the neural network model, obtaining the fitness of the individual as follows:
wherein, F represents the fitness of the individual;
and traversing all the individuals to obtain the fitness of each individual in the population.
Further, performing secondary training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width and the optimal network weight, and using the trained neural network model as a power consumption prediction model of a power supply area, including:
taking the optimal network center, the optimal network width and the optimal network weight as an initial network center, an initial network width and an initial network weight of the neural network model in a secondary training process;
taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model;
taking the corresponding used electric quantity in the preprocessed sample data as expected output data of the neural network model, and acquiring a residual signal corresponding to the neural network model according to the output data and the expected output data as follows:
e n ＝d n -y n
wherein e is n Representing a residual signal corresponding to the neural network model; y is n Indicating the input nth sample data x n Output data of the neural network model; d n Representing output data y n Corresponding expected output data, N =1,2, \8230, N, representing the total number of input sample data;
adjusting the network center, the network width and the network weight of the neural network model according to the residual signal corresponding to the neural network model, wherein the adjusting of the network center, the network width and the network weight of the neural network model comprises the following steps:
wherein the content of the first and second substances,
representing the adjusted network weight, namely the network weight corresponding to the jth neuron in the hidden layer during the t +1 training;
representing the network weight before adjustment, namely the network weight corresponding to the jth neuron in the hidden layer during the tth training; eta w Representing a first learning rate;
representing the residual signal e corresponding to the neural network model during the t-th training n ；
Representing sample data x input by jth neuron pair of hidden layer in the neural network model during the t training n (ii) a response; beta is a w Represents a first operation parameter value corresponding to the network weight, and beta w Has a value range of [0,1 ]]；
Representing the adjustment quantity corresponding to the network weight value during the t-th training;
representing the network weight corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity is increased,
representing an optimal network weight; j =1,2, \8230j, J representing the total number of neurons in the neural network model;
representing the adjusted network center, namely the network center corresponding to the jth neuron in the hidden layer during the t +1 th training;
representing a network center before adjustment, namely the network center corresponding to the jth neuron in the hidden layer during the tth training; eta c Representing a second learning rate; beta is a c Represents a second operation parameter value corresponding to the network center, and beta c Has a value range of [0,1 ]]；
The adjustment amount corresponding to the network center during the t training is shown,
representing the network center corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity is increased,
representing an optimal hub;
the adjusted network width is represented, namely the network width corresponding to the jth neuron in the hidden layer during the t +1 th training;
representing the network width before adjustment, namely the network width corresponding to the jth neuron in the hidden layer during the tth training; eta σ Represents the third learning rate, β c Represents a third operation parameter value corresponding to the network width, and beta c Has a value range of [0,1 ]]；
The adjustment quantity corresponding to the network width during the t-th training is shown;
representing the network width corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal network width; residual signal e corresponding to the neural network model n Obtaining the sum of squares E of the total training error e Comprises the following steps:
determining the sum of squares E of the total training error e And if the current network center, the network width and the network weight are respectively used as the final network center, the network width and the network weight of the neural network model to obtain the trained neural network model, the trained neural network model is used as a power supply area power consumption prediction model, otherwise, N sample data are reselected, and the next adjustment of the network center, the network width and the network weight is carried out on the basis of the reselected N sample data until the training is finished.
Furthermore, according to the power generation amount corresponding to the power plant and the used power amount of the power supply area, the method for scheduling the power system comprises the following steps:
judging whether the generated energy corresponding to the power plant is larger than the used electric quantity of the power supply area, if so, supplying power only by the power plant corresponding to the power supply area in a period of time in the future, otherwise, acquiring redundant electric quantity of other power plants in the power system, wherein the redundant electric quantity represents the extra electric quantity after the other power plants supply power for the corresponding power supply area;
judging whether the surplus electric quantity of other power plants is larger than or equal to the electric quantity difference of the current power supply area or not within the distance threshold range, if so, scheduling the surplus electric quantity of the other power plants closest to the current power supply area, otherwise, scheduling the surplus electric quantities of at least two other power plants closest to the current power supply area to complete scheduling of the power system, wherein the electric quantity difference is the difference value obtained by subtracting the generated energy corresponding to the current power plant from the used electric quantity of the current power supply area, the surplus electric quantities of the at least two other power plants closest to the current power plant are larger than or equal to the electric quantity difference, and the distance between the current power plant and the other power plants is stored in a database or a cloud in advance.
In a second aspect, an embodiment of the present application provides a power system scheduling apparatus based on a carbon quota, including an electric power generation amount obtaining module, an electric power usage amount predicting module, and a scheduling module;
the power generation amount acquisition module is used for acquiring a carbon quota to be used of a power plant corresponding to a power supply area in a future period of time and power generation amount corresponding to the carbon quota to be used, and the carbon quota to be used and the corresponding power generation amount are stored in a cloud end or a database in advance;
the power consumption prediction module is used for acquiring data of a power supply area and predicting the power consumption of the power supply area in a future period of time according to the data of the power supply area;
and the scheduling module is used for scheduling the power system according to the power generation amount corresponding to the power plant and the use electric quantity of the power supply area.
In a third aspect, an embodiment of the present application provides a power system scheduling device based on a carbon quota, including a memory and a processor, where the memory and the processor are connected to each other through a bus;
the memory stores computer-executable instructions;
the processor executes the memory-stored computer-executable instructions to cause the processor to perform the carbon quota-based power system scheduling method as described in the first aspect.
The invention has the beneficial effects that:
(1) The invention provides a method, a device and equipment for scheduling a power system based on carbon quota, which realize the prediction of the used electric quantity of a power supply area, schedule the power system according to the prediction result and avoid the occurrence of insufficient power supply.
(2) When the neural network model is trained, the genetic algorithm is adopted for training, so that the condition of local optimal solution is avoided.
(3) The method can accurately predict the power consumption of the power supply area in a future period of time, and provides a basis for the scheduling of the power system.
Drawings
Fig. 1 is a flowchart of a power system scheduling method based on carbon quota according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of an electric power system scheduling apparatus based on carbon quota according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a power system scheduling device based on a carbon quota, according to an embodiment of the present invention.
Wherein: 21-power generation amount acquisition module, 22-power consumption amount prediction module, 23-scheduling module, 31-memory, 32-processor and 33-bus.
With the above figures, there are shown specific embodiments of the present application, which will be described in more detail below. The drawings and written description are not intended to limit the scope of the inventive concepts in any manner, but rather to illustrate the concepts of the application by those skilled in the art with reference to specific embodiments.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
As shown in fig. 1, a method for scheduling a power system based on carbon quota includes:
s11, acquiring a carbon quota to be used of a power plant in a future period of time corresponding to a power supply area and power generation amount corresponding to the carbon quota to be used, wherein the carbon quota to be used and the corresponding power generation amount are stored in a cloud or a database in advance.
The future period of time may refer to a continuous time point of one, two, or three days in the future, where the continuous time point refers to a continuous whole number of points per day, for example, the carbon quota for the future day may include a carbon quota corresponding to a whole point in 1-24 hours in the future, the carbon quota corresponding to 1 point indicates that the carbon quota can be used between 24 points and 1 point, and the carbon quota corresponding to 2 points indicates that the carbon quota can be used between 1 point and 2 points.
The power generation amount corresponding to the carbon quota can be obtained by the following method: and predicting the power generation amount corresponding to the carbon quota according to the historical carbon quota and the generated power generation amount data, or training a power generation amount prediction model according to the historical carbon quota and the generated power generation amount data, and predicting the power generation amount corresponding to the carbon quota according to the prediction model. For example, a BP (Back Propagation, multilayer feed-forward network) neural network may be used as the power generation amount prediction module, and training of the BP neural network is conventional training, which is not described herein again.
S12, collecting power supply area data, and predicting the use electric quantity of the power supply area in a future period of time according to the power supply area data.
For example, power supply area data corresponding to M continuous time points in the future (the time points are the whole time point in a day) are collected, and the power consumption of the power supply area at the M continuous time points in the future is predicted according to the power supply area data corresponding to the M time points.
And S13, scheduling the power system according to the power generation amount corresponding to the power plant and the used power amount of the power supply area.
In one possible embodiment, collecting power supply area data and predicting the power consumption of the power supply area in a future period of time according to the power supply area data comprises:
and collecting a plurality of sample data, and preprocessing the sample data to obtain preprocessed sample data.
And constructing a neural network model, and training the neural network model by adopting the preprocessed sample data to obtain a power consumption prediction model of the power supply area.
The method comprises the steps of collecting power supply area data, using the power supply area data as input of a power supply area power consumption prediction model, predicting the power consumption of a power supply area in a future period of time, wherein the sample data and the power supply area data are the same type of data.
In one possible embodiment, the sample data includes seasonal data, weather data, temperature data, date type data and corresponding used electricity amount of the power supply area at continuous time points, the seasonal data includes spring, summer, autumn or winter, the weather data includes raining, sunny days, cloudy days or snow days, and the date type data includes working days or non-working days; the continuous time points represent time points corresponding to a plurality of consecutive days, and the time points represent integral values of each day.
Before the sample data is preprocessed, the text data in the sample data can be converted into the input data form of the neural network model, and an input vector is constructed according to the sequence of seasonal data, weather data, temperature data and date type data so as to perform subsequent training.
In a possible implementation manner, the preprocessing the sample data to obtain preprocessed sample data includes:
judging whether missing values exist in the seasonal data, the weather data, the temperature data, the date type data and the corresponding used electric quantity in the sample data, if so, completing the missing values, and performing normalization processing on the completed sample data to obtain the preprocessed sample data, otherwise, directly performing normalization processing on the sample data to obtain the preprocessed sample data.
In this embodiment, the missing value may be complemented by the following method:
acquiring time points and types corresponding to the missing values, wherein the types can comprise season data, weather data, temperature data, date type data and corresponding use electric quantity;
acquiring a front time point or a rear time point of a time point corresponding to the missing value, and completing the missing value according to the data of the same type as the missing value corresponding to the front time point and the data of the same type as the missing value corresponding to the rear time point, wherein the method specifically comprises the following steps:
T M indicating a missing value, which may be seasonal data, weather data, temperature data, date type data or corresponding electricity usage, T M+1 Representing data of the same type, T, corresponding to a point in time after the missing value M-1 And representing the same type of data corresponding to the time point before the missing value.
Or, the data at the time point corresponding to the missing value on the previous day is filled in the missing value part, or the data at the time point corresponding to the missing value on the next day is filled in the missing value part.
In a possible implementation manner, constructing a neural network model, and training the neural network model by using the preprocessed sample data to obtain a power supply area usage power consumption prediction model, including:
A. an RBF (radial basis function) neural network is constructed and used as a neural network model.
B. Randomly initializing the network center, the network width and the network weight of the neural network model, constructing a vector by using the network center, the network width and the network weight, taking the vector as a primary chromosome, and repeating the steps to obtain two parent chromosomes.
C. Adding the two parent chromosomes to a population, wherein the individuals in the population are used for characterizing the chromosomes.
For example, the first parent chromosome may be f 1 ＝[c 11 ,c 12 ,...,c 1j ,...,c 1J ,σ 11 ,σ 12 ,...,σ 1j ,...,σ 1J ,w 11 ,w 12 ,...,w 1j ,...,w 1J ]The second parent chromosome may be f 2 ＝[c 21 ,c 22 ,...,c 2j ,...,c 2J ,σ 21 ,σ 22 ,...,σ 2j ,...,σ 2J ,w 21 ,w 22 ,...,w 2j ,...,w 1J ]，c 1j 、σ 1j And w 1j Respectively to indicate the implicationA first network center, a first network width and a first network weight corresponding to the jth neuron in the layer, c 2j 、σ 2j And w 2j And respectively representing a second network center, a second network width and a second network weight corresponding to the jth neuron in the hidden layer.
D. And acquiring the fitness of each individual in the population by adopting the preprocessed sample data.
E. And G, judging whether the fitness of the individuals in the population is greater than a preset fitness threshold, if so, taking the individual with the maximum fitness as an optimal individual to obtain an optimal network center, an optimal network width and an optimal network weight, and entering the step G, otherwise, entering the step F.
F. And D, carrying out mutation and cross operation on individuals of the population to obtain offspring chromosomes, adding the offspring chromosomes into the population, and returning to the step D.
In this embodiment, the interleaving operation may include: and randomly exchanging the network centers, the network widths or the network weights in the two individuals. The mutation operation may include: and limiting the value range of the network center, the value range of the network width and the value range of the network weight, randomly selecting the network center, the network width and/or the network weight, randomly generating a new network center, the network width and/or the network weight in the value range, and replacing the selected network center, the network width and/or the network weight.
It should be noted that the above-mentioned crossover operation and mutation operation are only examples, and other methods in the prior art may be adopted to perform the mutation operation and crossover operation.
G. And carrying out secondary training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width and the optimal network weight, and taking the trained neural network model as a power utilization prediction model of a power supply area.
The network center, the network width and the network weight of the neural network model are initially trained through a genetic algorithm, the optimal individuals (the optimal network center, the optimal network width and the optimal network weight) obtained through the genetic algorithm are used as the initial network center, the initial network width and the initial network weight of the neural network model during secondary training, secondary training is carried out on the basis, and the situation of local optimal solution is avoided.
In a possible embodiment, obtaining the fitness of each individual in the population by using the preprocessed sample data includes:
and taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model.
Taking the corresponding use electric quantity in the preprocessed sample data as expected output data of the neural network model, and obtaining the error corresponding to the neural network model according to the output data and the expected output data as follows:
wherein E represents the corresponding error of the neural network model, D i Indicating the power consumption Y in the ith preprocessed sample data i The method comprises the steps of obtaining a sample data after the ith preprocessing, obtaining output data of a neural network model according to the sample data after the ith preprocessing, wherein I =1,2, \ 8230, and I represent the number of the sample data used when the fitness is obtained.
According to the error corresponding to the neural network model, obtaining the fitness of the individual as follows:
wherein F represents the fitness of the individual;
and traversing all the individuals to obtain the fitness of each individual in the population.
In a possible implementation manner, performing secondary training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width and the optimal network weight, and using the trained neural network model as a power consumption prediction model of a power supply area, includes:
taking the optimal network center, the optimal network width and the optimal network weight as an initial network center, an initial network width and an initial network weight of the neural network model in a secondary training process;
and taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model.
Taking the corresponding used electric quantity in the preprocessed sample data as expected output data of the neural network model, and acquiring a residual signal corresponding to the neural network model according to the output data and the expected output data as follows:
e n ＝d n -y n
wherein e is n Representing a residual signal corresponding to the neural network model; y is n Indicating the input nth sample data x n Output data of the neural network model; d is a radical of n Representing output data y n Corresponding expected output data, N =1,2, \8230, N, denotes the total number of input sample data.
Adjusting the network center, the network width and the network weight of the neural network model according to the residual signal corresponding to the neural network model, wherein the adjusting of the network center, the network width and the network weight of the neural network model comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,
representing the adjusted network weight, namely the network weight corresponding to the jth neuron in the hidden layer during the t +1 training;
representing the network weight before adjustment, namely the network weight corresponding to the jth neuron in the hidden layer during the tth training; eta w Representing a first learning rate;
representing a residual signal e corresponding to the neural network model during the t-th training n ；
Representing sample data x input by jth neuron pair of hidden layer in the neural network model during the t training n (ii) a response of (d); beta is a w Represents a first operation parameter value corresponding to the network weight, and beta w Has a value range of [0,1 ]]；
Representing the adjustment quantity corresponding to the network weight value during the t-th training;
representing the network weight corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal network weight; j =1,2, \8230j, J representing the total number of neurons in the neural network model;
representing the adjusted network center, namely the network center corresponding to the jth neuron in the hidden layer during the t +1 th training;
representing a network center before adjustment, namely the network center corresponding to the jth neuron in the hidden layer during the tth training; eta c Representing a second learning rate; beta is a c Represents a second operation parameter value corresponding to the network center, and beta c Has a value range of [0,1 ]]；
The adjustment amount corresponding to the network center during the t training is shown,
representing the network center corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal hub;
the adjusted network width is represented, namely the network width corresponding to the jth neuron in the hidden layer during the t +1 training;
representing the network width before adjustment, namely the network width corresponding to the jth neuron in the hidden layer during the tth training; eta σ Represents the third learning rate, β c Represents a third operation parameter value corresponding to the network width, and beta c Has a value range of [0,1 ]]；
The adjustment quantity corresponding to the network width is shown when the training is carried out for the t time;
representing the network width corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing the optimal network width.
In the present embodiment, it is preferred that,
and
both represent moment terms.
According to the residual signal e corresponding to the neural network model n Obtaining the sum of squares E of the total training error e Comprises the following steps:
determining the sum of squares E of the total training error e If the current network center, the network width and the network weight are respectively used as the final network center, the network width and the network weight of the neural network model to obtain the trained neural network model, the trained neural network model is used as a power supply area to use the power consumption prediction model, otherwise, N sample data are selected again, and the next network center is carried out on the basis of the N newly selected sample dataAnd adjusting the network width and the network weight until the training is finished.
In this embodiment, each training time, the first learning rate, the second learning rate, and the third learning rate may be adjusted to accelerate the convergence rate of the neural network model.
For example,
eta (t + 1) represents the learning rate after adjustment, eta (t) represents the learning rate before adjustment, E e (t) represents the sum of the squares of the total training errors in the t-th training, E e (t-1) represents the sum of the squares of the total training errors at the t-1 st training.
In a possible implementation manner, the scheduling of the power system according to the power generation amount corresponding to the power plant and the used power amount of the power supply area comprises the following steps:
judging whether the generated energy corresponding to the power plant is larger than the used electric quantity of the power supply area, if so, supplying power only by the power plant corresponding to the power supply area in a period of time in the future, otherwise, acquiring redundant electric quantity of other power plants in the power system, wherein the redundant electric quantity represents the extra electric quantity after the other power plants supply power for the corresponding power supply area;
judging whether the surplus electric quantity of other power plants in the distance threshold range is larger than or equal to the electric quantity difference of the current power supply area, if so, scheduling the surplus electric quantity of the other power plants closest to the current power supply area, otherwise, scheduling the surplus electric quantity of at least two other power plants closest to the current power supply area to complete the scheduling of the power system, wherein the electric quantity difference is the difference value obtained by subtracting the generated energy corresponding to the current power plant from the used electric quantity of the current power supply area, the surplus electric quantity of the at least two other power plants closest to the current power plant is larger than or equal to the electric quantity difference, and the distance between the current power plant and the other power plants is stored in a database or a cloud in advance.
When the surplus power of other power plants is scheduled for a part, the scheduled part is subtracted from the surplus power to ensure the scheduling accuracy.
It should be noted that the data used in the present invention may be pre-stored in the database or in the cloud, so as to facilitate the execution of the whole method.
Example 2
As shown in fig. 2, an electric power system scheduling apparatus based on carbon quota according to an embodiment of the present application includes an electric power generation amount obtaining module 21, an electric power usage amount predicting module 22, and a scheduling module 23;
the power generation amount obtaining module 21 is configured to obtain a carbon quota to be used by a power plant in a future period of time corresponding to a power supply area and a power generation amount corresponding to the carbon quota to be used, where the carbon quota to be used and the power generation amount corresponding to the carbon quota to be used are stored in a cloud or a database in advance;
the power consumption predicting module 22 is configured to collect data of a power supply area, and predict power consumption of the power supply area in a future period of time according to the data of the power supply area;
the scheduling module 23 is configured to schedule the power system according to the power generation amount corresponding to the power plant and the power consumption of the power supply area.
The power system scheduling device based on carbon quota according to this embodiment may implement the technical solution described in embodiment 1, and the implementation principle and the beneficial effect thereof are similar, and details are not repeated here.
Example 3
As shown in fig. 3, an embodiment of the present application provides a carbon quota-based power system scheduling apparatus, which includes a memory 31 and a processor 32, where the memory 31 and the processor 32 are connected to each other through a bus 33.
The memory 31 stores computer-executable instructions.
The processor 32 executes the memory-stored computer-executable instructions to cause the processor to perform a method for carbon quota-based scheduling of an electrical power system as described in embodiment 1.
Example 4
The present embodiment provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, the computer-readable storage medium is used for implementing the carbon quota-based power system scheduling method according to embodiment 1.
For example, the Memory may include, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Flash Memory (Flash Memory), a First In First Out (FIFO), a First In Last Out (FILO), and/or a First In Last Out (FILO); in particular, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so on. The processor may be implemented in at least one hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array), and meanwhile, the processor may also include a main processor and a coprocessor, where the main processor is a processor for Processing data in an awake state, and is also called a CPU (Central Processing Unit); a coprocessor is a low power processor for processing data in a standby state.
In some embodiments, the processor may be integrated with a GPU (Graphics Processing Unit) which is responsible for rendering and drawing contents required to be displayed on the display screen, for example, the processor may not be limited to a processor using a model STM32F105 series microprocessor, a Reduced Instruction Set Computer (RISC) microprocessor, an X86 or the like architecture processor or an integrated embedded neural Network Processor (NPU); the transceiver may be, but is not limited to, a wireless fidelity (WIFI) wireless transceiver, a bluetooth wireless transceiver, a General Packet Radio Service (GPRS) wireless transceiver, a ZigBee wireless transceiver (ieee802.15.4 standard-based low power local area network protocol), a 3G transceiver, a 4G transceiver, and/or a 5G transceiver, etc. In addition, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.
Example 5
Embodiments of the present application may also provide a computer program product, which includes a computer program, and when the computer program is executed by a processor, the method for scheduling a power system based on carbon quota as described in embodiment 1 is implemented.
It should be noted that any method utilizing the concepts of the present invention should be considered within the scope of the present application. Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains.
It will be understood that the present application is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.
Claims (10)
1. A power system scheduling method based on carbon quota is characterized by comprising the following steps:
acquiring a carbon quota to be used of a power plant in a future period of time and power generation amount corresponding to the carbon quota to be used of the power plant in a power supply area, wherein the carbon quota to be used and the corresponding power generation amount are stored in a cloud or a database in advance;
acquiring power supply area data, and predicting the use electric quantity of the power supply area in a future period of time according to the power supply area data;
and scheduling the power system according to the generated energy corresponding to the power plant and the used electric quantity of the power supply area.
2. The carbon quota-based power system scheduling method of claim 1, wherein collecting power supply area data and predicting the amount of electricity used by the power supply area in a future period of time according to the power supply area data comprises:
collecting a plurality of sample data, and preprocessing the sample data to obtain preprocessed sample data;
constructing a neural network model, and training the neural network model by adopting the preprocessed sample data to obtain a power consumption prediction model of a power supply area;
and acquiring power supply area data, and predicting the power consumption of the power supply area in a future period by taking the power supply area data as the input of a power consumption prediction model of the power supply area, wherein the sample data and the power supply area data are the same type of data.
3. The carbon quota based power system scheduling method of claim 2, wherein the sample data comprises seasonal data, weather data, temperature data, date type data, and corresponding amounts of used power of the power supply area at successive points in time, the seasonal data comprising spring, summer, fall, or winter, the weather data comprising rain, sunny, cloudy, or snow, and the date type data comprising working days or non-working days;
the continuous time points represent time points corresponding to a plurality of consecutive days, and the time points represent integral values of each day.
4. The carbon quota-based power system scheduling method of claim 3, wherein preprocessing the sample data to obtain preprocessed sample data comprises:
judging whether missing values exist in season data, weather data, temperature data, date type data and corresponding used electric quantity in the sample data, if so, completing the missing values, and performing normalization processing on the completed sample data, otherwise, directly performing normalization processing on the sample data to obtain the preprocessed sample data.
5. The carbon quota-based power system scheduling method according to claim 3, wherein a neural network model is constructed, and the neural network model is trained by using the preprocessed sample data to obtain a power supply area usage power consumption prediction model, and the method comprises the steps of:
A. constructing an RBF neural network, and taking the RBF neural network as a neural network model;
B. randomly initializing a network center, a network width and a network weight of the neural network model, constructing a vector by using the network center, the network width and the network weight, taking the vector as a primary chromosome, and repeating the step to obtain two parent chromosomes;
C. adding the two parent chromosomes into a population, wherein individuals in the population are used for representing the chromosomes;
D. acquiring the fitness of each individual in the population by adopting the preprocessed sample data;
E. judging whether the fitness of the individuals in the population is greater than a preset fitness threshold, if so, taking the individual with the maximum fitness as an optimal individual to obtain an optimal network center, an optimal network width and an optimal network weight, and entering a step G, otherwise, entering a step F;
F. carrying out variation and cross operation on individuals of the population to obtain offspring chromosomes, adding the offspring chromosomes into the population, and returning to the step D;
G. and carrying out secondary training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width and the optimal network weight, and taking the trained neural network model as a power utilization prediction model of a power supply area.
6. The carbon quota-based power system scheduling method of claim 5, wherein obtaining the fitness of each individual in the population using the preprocessed sample data comprises:
taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model;
taking the corresponding use electric quantity in the preprocessed sample data as expected output data of the neural network model, and obtaining the error corresponding to the neural network model according to the output data and the expected output data as follows:
wherein E represents the corresponding error of the neural network model, D i Indicating the power consumption Y in the ith preprocessed sample data i The method comprises the steps of obtaining a neural network model according to ith preprocessed sample data, wherein the output data of the neural network model obtained according to the ith preprocessed sample data is represented by I =1,2, \ 8230;
according to the error corresponding to the neural network model, obtaining the fitness of the individual as follows:
wherein F represents the fitness of the individual;
and traversing all the individuals to obtain the fitness of each individual in the population.
7. The method according to claim 5, wherein performing a second training on the neural network model according to the preprocessed sample data, the optimal network center, the optimal network width, and the optimal network weight, and using the trained neural network model as a power consumption prediction model for the power supply area comprises:
taking the optimal network center, the optimal network width and the optimal network weight as an initial network center, an initial network width and an initial network weight of the neural network model in a secondary training process;
taking seasonal data, weather data, temperature data and date type data in the preprocessed sample data as input data of the neural network model, and acquiring output data of the neural network model;
taking the corresponding used electric quantity in the preprocessed sample data as expected output data of the neural network model, and acquiring a residual signal corresponding to the neural network model according to the output data and the expected output data as follows:
e n ＝d n -y n
wherein e is n Representing a residual signal corresponding to the neural network model; y is n Indicating the input nth sample data x n Output data of the neural network model; d n Representing output data y n Corresponding desired output data, N =1,2, \ 8230, N, N representing the total number of input sample data;
adjusting the network center, the network width and the network weight of the neural network model according to the residual signal corresponding to the neural network model, wherein the adjusting of the network center, the network width and the network weight of the neural network model comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,
representing the adjusted network weight, namely the network weight corresponding to the jth neuron in the hidden layer during the t +1 training;
indicating the network weights before adjustment, i.e. implicit during the t-th trainingNetwork weight corresponding to jth neuron in layer; eta w Representing a first learning rate;
representing a residual signal e corresponding to the neural network model during the t-th training n ；
Representing sample data x input by jth neuron pair of hidden layer in the neural network model during the t training n (ii) a response of (d); beta is a beta w Represents a first operation parameter value corresponding to the network weight, and beta w Has a value range of [0,1 ]]；
Representing the adjustment quantity corresponding to the network weight when the training is performed for the t time;
representing the network weight corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal network weight; j =1,2, \8230j, J representing the total number of neurons in the neural network model;
representing the adjusted network center, namely the network center corresponding to the jth neuron in the hidden layer during the t +1 th training;
before the indication is adjustedThe network center of (a), namely the network center corresponding to the jth neuron in the hidden layer during the t training; eta c Representing a second learning rate; beta is a beta c Represents a second operation parameter value corresponding to the network center, and beta c Has a value range of [0,1 ]]；
The adjustment amount corresponding to the network center during the t training is shown,
representing the network center corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal hub;
the adjusted network width is represented, namely the network width corresponding to the jth neuron in the hidden layer during the t +1 th training;
representing the network width before adjustment, namely the network width corresponding to the jth neuron in the hidden layer during the tth training; eta σ Represents the third learning rate, β c Represents a third operation parameter value corresponding to the network width, and beta c Has a value range of [0,1 ]]；
The adjustment quantity corresponding to the network width is shown when the training is carried out for the t time;
representing the network width corresponding to the jth neuron in the hidden layer during the t-1 training; when the t =1, the signal intensity of the signal is increased,
representing an optimal network width;
according to the residual signal e corresponding to the neural network model n Obtaining the sum of squares E of the total training error e Comprises the following steps:
determining the sum of squares E of the total training error e And if the current network center, the network width and the network weight are respectively used as the final network center, the network width and the network weight of the neural network model to obtain the trained neural network model, the trained neural network model is used as a power supply area power consumption prediction model, otherwise, N sample data are reselected, and the next adjustment of the network center, the network width and the network weight is carried out on the basis of the reselected N sample data until the training is finished.
8. The method for scheduling the power system according to any one of claims 1 to 7, wherein the scheduling the power system according to the power generation amount corresponding to the power plant and the power usage amount of the power supply area comprises:
judging whether the generated energy corresponding to the power plant is larger than the used electric quantity of the power supply area, if so, supplying power only by the power plant corresponding to the power supply area in a period of time in the future, otherwise, acquiring redundant electric quantity of other power plants in the power system, wherein the redundant electric quantity represents the extra electric quantity after the other power plants supply power for the corresponding power supply area;
judging whether the surplus electric quantity of other power plants in the distance threshold range is larger than or equal to the electric quantity difference of the current power supply area, if so, scheduling the surplus electric quantity of the other power plants closest to the current power supply area, otherwise, scheduling the surplus electric quantity of at least two other power plants closest to the current power supply area to complete the scheduling of the power system, wherein the electric quantity difference is the difference value obtained by subtracting the generated energy corresponding to the current power plant from the used electric quantity of the current power supply area, the surplus electric quantity of the at least two other power plants closest to the current power plant is larger than or equal to the electric quantity difference, and the distance between the current power plant and the other power plants is stored in a database or a cloud in advance.
9. A power system scheduling device based on carbon quota is characterized by comprising a power generation capacity acquisition module, a used power prediction module and a scheduling module;
the power generation amount acquisition module is used for acquiring a carbon quota to be used of a power plant in a future period of time corresponding to a power supply area and power generation amount corresponding to the carbon quota to be used, and the carbon quota to be used and the corresponding power generation amount are stored in a cloud or a database in advance;
the power consumption prediction module is used for acquiring data of a power supply area and predicting the power consumption of the power supply area in a period of time in the future according to the data of the power supply area;
and the scheduling module is used for scheduling the power system according to the power generation amount corresponding to the power plant and the use electric quantity of the power supply area.
10. The power system scheduling device based on the carbon quota is characterized by comprising a memory and a processor, wherein the memory and the processor are connected with each other through a bus;
the memory stores computer execution instructions;
the processor executes computer-executable instructions stored by the memory to cause the processor to perform the method of carbon quota based scheduling of an electrical power system as claimed in any of claims 2 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211478786.9A CN115796023A (en) | 2022-11-23 | 2022-11-23 | Power system scheduling method, device and equipment based on carbon quota |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211478786.9A CN115796023A (en) | 2022-11-23 | 2022-11-23 | Power system scheduling method, device and equipment based on carbon quota |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115796023A true CN115796023A (en) | 2023-03-14 |
Family
ID=85440787
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211478786.9A Pending CN115796023A (en) | 2022-11-23 | 2022-11-23 | Power system scheduling method, device and equipment based on carbon quota |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796023A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116187788A (en) * | 2023-05-04 | 2023-05-30 | 江苏智能低碳科技发展有限公司 | Application platform of carbon management algorithm for factory |
| CN117852848A (en) * | 2024-03-08 | 2024-04-09 | 山东黄金电力有限公司 | Data information management system for configuring power system based on big data |
-
2022
- 2022-11-23 CN CN202211478786.9A patent/CN115796023A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116187788A (en) * | 2023-05-04 | 2023-05-30 | 江苏智能低碳科技发展有限公司 | Application platform of carbon management algorithm for factory |
| CN117852848A (en) * | 2024-03-08 | 2024-04-09 | 山东黄金电力有限公司 | Data information management system for configuring power system based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11581740B2 (en) | Method, system and storage medium for load dispatch optimization for residential microgrid | |
| CN115796023A (en) | Power system scheduling method, device and equipment based on carbon quota | |
| CN112614009B (en) | Power grid energy management method and system based on deep expectation Q-learning | |
| Ding et al. | Rolling optimization of wind farm and energy storage system in electricity markets | |
| Korpaas et al. | Operation and sizing of energy storage for wind power plants in a market system | |
| CN102183621B (en) | Aquaculture dissolved oxygen concentration online forecasting method and system | |
| CN110288210A (en) | Project level road infrastructure comprehensive maintenance decision-making technique and system | |
| Huang et al. | A control strategy based on deep reinforcement learning under the combined wind-solar storage system | |
| CN116911076B (en) | Toughness support simulation method and device for power distribution network by multiple micro-grids and electronic equipment | |
| CN106374536A (en) | Low-carbon power supply investment decision-making method of power capacity market under new energy access condition | |
| CN112215428B (en) | Photovoltaic power generation power prediction method and system based on error correction and fuzzy logic | |
| Bui et al. | Forecasting energy output of a solar power plant in curtailment condition based on LSTM using P/GHI coefficient and validation in training process, a case study in Vietnam | |
| Endemaño-Ventura et al. | Optimal energy bidding for renewable plants: A practical application to an actual wind farm in Spain | |
| CN105447596A (en) | Optimized power data short-term prediction method | |
| CN114444805A (en) | Control method for smooth output of multi-photovoltaic power station shared energy storage system | |
| Hosseini et al. | Data-driven multi-resolution probabilistic energy and reserve bidding of wind power | |
| CN110147908A (en) | A kind of wind power forecasting method based on three-dimensional optimal similarity and improvement cuckoo algorithm | |
| Du et al. | Optimal whole-life-cycle planning for battery energy storage system with normalized quantification of multi-services profitability | |
| CN109687452B (en) | Analysis method and device for power supply capacity of power distribution network and storage medium | |
| CN116596252A (en) | Multi-target charging scheduling method for electric automobile clusters | |
| CN112926811A (en) | Load aggregation commercial power purchasing method considering renewable energy consumption responsibility weight | |
| CN116362136A (en) | Self-dispatching optimization method and system for independent energy storage system | |
| CN115936756A (en) | Hydropower plant power generation control method considering power grid operation stability | |
| CN109740825A (en) | A kind of electric car charging/discharging thereof considered under traffic congestion factor | |
| CN112510678A (en) | Time sequence simulation-based power selling company distributed power capacity configuration 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 |