CN110601193B - Power load prediction system and method based on environment variable perception and hysteresis - Google Patents

Power load prediction system and method based on environment variable perception and hysteresis Download PDF

Info

Publication number
CN110601193B
CN110601193B CN201910951985.9A CN201910951985A CN110601193B CN 110601193 B CN110601193 B CN 110601193B CN 201910951985 A CN201910951985 A CN 201910951985A CN 110601193 B CN110601193 B CN 110601193B
Authority
CN
China
Prior art keywords
module
power
region
power load
load
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.)
Active
Application number
CN201910951985.9A
Other languages
Chinese (zh)
Other versions
CN110601193A (en
Inventor
陈佳喜
刘兴川
赵迎迎
张聪冲
魏瑞超
彭腾
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Smart City Research Institute Of China Electronics Technology Group Corp
Original Assignee
Smart City Research Institute Of China Electronics Technology Group Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Smart City Research Institute Of China Electronics Technology Group Corp filed Critical Smart City Research Institute Of China Electronics Technology Group Corp
Priority to CN201910951985.9A priority Critical patent/CN110601193B/en
Publication of CN110601193A publication Critical patent/CN110601193A/en
Application granted granted Critical
Publication of CN110601193B publication Critical patent/CN110601193B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

A power load prediction system and method based on environment variable perception and hysteresis comprises a sensor module, a storage module, a variable hysteresis module, a Res-LSTM prediction module, an error redundancy calculation module, an electric energy storage calculation module and a power scheduling calculation module; the sensor modules are arranged in plurality; the sensor module is in communication connection with the storage module; the storage module is in communication connection with the variable hysteresis module; the variable hysteresis module is in communication connection with the Res-LSTM prediction module, the Res-LSTM prediction module is in communication connection with the error redundancy calculation module, and the error redundancy calculation module is in communication connection with the power scheduling calculation module and the electric energy storage calculation module. After the power loads in each area are pre-distributed, the actual power loads are compared in the future, so that reference is made for power dispatching and electric energy storage, and electric energy waste is reduced.

Description

Power load prediction system and method based on environment variable perception and hysteresis
Technical Field
The invention relates to the technical field of power internet of things, in particular to a power load prediction system and method based on environment variable perception and hysteresis.
Background
The power load is a numerical value of power consumption of various electric devices installed in users such as state organs, enterprises, residents, and the like. The power load prediction is a process of predicting the temporal distribution and the spatial distribution of the future power consumption or the power consumption with the power load as a research object. The power load prediction is a decision basis of power dispatching in a power grid, is an important component of power system planning, is the basis of economic operation of the power system, and is important to the planning and stable operation of the whole power grid.
At present, the practical aspect of the power load in the power grid is mainly monitored, and the prediction of the power load is mainly based on the historical data of the power load to roughly estimate the short-term power demand in the future.
The prediction of the power load in the prior art is mainly based on the historical data of the power load, and the defects or shortcomings are mainly as follows:
1. the prediction deviation is large: the actual power load is greatly influenced by various complex factors such as weather (temperature, humidity, precipitation and the like), holidays, special conditions, large-scale industrial user emergencies, economic operation conditions, national policy and regulation and the like, so that the traditional prediction method purely based on historical data of the power load has large deviation.
2. The prediction area is small: the existing power load prediction model does not usually consider environmental factors, so that the power load can be predicted only in a small-range area on the space, and once a plurality of areas with large differences of the environmental factors are involved, the existing prediction method based on historical load data is not reliable any more.
In order to solve the above problems, the present application provides a power load prediction system and method based on environment variable sensing and hysteresis.
Disclosure of Invention
Objects of the invention
In order to solve the technical problems of large prediction deviation and small prediction area of the power load in the prior art in the background art, the invention provides a power load prediction system and method based on environment variable perception and hysteresis.
(II) technical scheme
In order to solve the problems, the invention provides a power load prediction system based on environment variable sensing and hysteresis, which comprises a sensor module, a storage module, a variable hysteresis module, a Res-LSTM prediction module, an error redundancy calculation module, an electric energy storage calculation module and a power scheduling calculation module;
the sensor modules are arranged in plurality; the sensor module is in communication connection with the storage module; the storage module is in communication connection with the variable hysteresis module; the variable hysteresis module is in communication connection with the Res-LSTM prediction module, the Res-LSTM prediction module is in communication connection with the error redundancy calculation module, and the error redundancy calculation module is in communication connection with the power scheduling calculation module and the electric energy storage calculation module.
Preferably, the sensor modules comprise rainfall sensors, temperature sensors and humidity sensors, and the sensor modules are distributed in the areas 1, 2, … and N; the rainfall sensor is used for measuring the rainfall of the region; simultaneously converting rainfall information into digital signals and transmitting the digital signals; the temperature sensor monitors the current temperature of the area, converts the current temperature into an available output signal and transmits the available output signal; the humidity sensor can measure the current humidity of the area, convert the current humidity into a usable output signal and transmit the usable output signal.
Preferably, at least one storage module is arranged in each area; the system is used for storing the historical information and the current information of rainfall, temperature and humidity uploaded by the sensor module.
Preferably, the inputs of the variable hysteresis module are the sensing data and the power load history data of each regional sensor module, and the outputs are the data set used for the Res-LSTM prediction module training or prediction.
Preferably, the Res-LSTM prediction module is configured to train historical data about environmental variables and power loads of each area, which are formed by the variable hysteresis module, and determine parameters of the model, so that future power loads can be predicted according to the subsequently input environmental variables and historical power loads.
Preferably, the power scheduling calculation module calculates the actual power load value of each future area.
Preferably, the electric energy storage calculation module calculates the electric energy storage power according to the actual electric power load of each region and the pre-distributed power of each region.
Preferably, the error redundancy calculation module calculates a pre-allocated power for each region according to the predicted power load of each region and the variance of the historical power load of the region.
A method for predicting power load based on environment variable perception and hysteresis comprises the following specific steps:
s10, the sensor modules arranged in each area periodically monitor information such as rainfall, temperature, humidity and the like, and after the information is collected, the information is uploaded and stored to the storage modules of each area;
s20, collecting power load historical data of each region, and matching the historical data with the sensor data in S10 by using an environment variable hysteresis module through setting a reasonable time interval to form an environment information-power load data set;
and S30, dividing the environmental information-power load data set generated in S20 into a training set, a verification set and a test set according to a certain proportion, training the Res-LSTM model by using the training set, determining the hyper-parameters in the Res-LSTM by using the verification set, and testing the accuracy of the model by using the test set.
S40, inputting the current environmental monitoring data such as rainfall, temperature and humidity in S10 and the historical data of the power load in S20 into the Res-LSTM model trained in S30 to obtain the predicted power load value of each area;
s50, combining the power load values obtained by prediction of each region in S40 with the variance of the power load historical data of each region, and obtaining the pre-distributed power of each region through an error redundancy calculation module;
and S60, inputting the pre-distributed power of each region in the S50 and the actual power load calculated by the power load scheduling module of each region into the electric energy storage calculation module to obtain an electric energy storage calculation value, and providing an operation basis for the actual power scheduling process.
Preferably, the training, verifying and testing process of the Res-LSTM model in step S30 is as follows:
s31, initializing parameters in Res-LSTM model, including matrix U in each neural network unitf，Uu，Uc，UoAnd matrix Gf，Gu，GoAnd a matrix Wf，Wu，Wc，Wo；
S32, inputting a training set of the data set formulated in S20, and repeatedly iterating parameters in the Res-LSTM model by using a mainstream BPTT algorithm until the algorithm converges or the error is lower than a preset threshold value;
s33, inputting a verification set of the data set formulated in S20, verifying the hyperparameters such as the number of neural network units by using methods such as grid search or random search, and selecting the hyperparameter combination with the minimum generalization error;
and S34, testing the trained Res-LSTM, and if the error is higher than a preset threshold value, re-training and verifying the model until the model precision meets the requirement.
The technical scheme of the invention has the following beneficial technical effects: 1. the prediction precision is high: according to the method, environmental factors including temperature, humidity and precipitation are considered, historical data of the regional power load are combined, namely universality of historical rules and specificity of the current condition of the regional power load are considered, and therefore the prediction accuracy is higher compared with the existing power load prediction method.
2. The number of prediction regions is as follows: the invention inputs the environmental data of temperature, humidity, precipitation and the like of a plurality of areas and the historical data of the power load of each area into a Res-LSTM model for prediction, considers the relevance of the power load of each area to a certain extent, and simultaneously realizes the prediction of the power load of the plurality of areas.
3. Pre-allocating redundancy: after the power load of each region is predicted, the predicted power load is not simply referred to a power system for power scheduling, but the variance of the historical power load prediction of the region is referred to, so that the power of regions with large load variation is pre-distributed with good elasticity.
4. Scheduling can refer to: after the power loads in each area are pre-distributed, the actual power loads are compared in the future, so that reference is made for power dispatching and electric energy storage, and electric energy waste is reduced.
Drawings
Fig. 1 is a structural block diagram of a power internet of things load prediction system of a power load prediction system and a power internet of things load prediction method based on environmental variable sensing and hysteresis provided by the invention.
Fig. 2 is a block diagram of a variable hysteresis module in the system and method for predicting a power load based on environmental variable sensing and hysteresis according to the present invention.
FIG. 3 is a Res-LSTM prediction module in the system and method for predicting electrical load based on environment variable sensing and hysteresis according to the present invention.
Fig. 4 is a block diagram of an error redundancy calculation module in the power load prediction system and method based on environment variable sensing and hysteresis according to the present invention.
Fig. 5 is a block diagram of an electrical energy storage calculation module in the system and method for predicting an electrical load based on environment variable sensing and hysteresis according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
As shown in fig. 1 to 5, the power load prediction system based on environmental variable sensing and hysteresis provided by the present invention includes a sensor module, a storage module, a variable hysteresis module, a Res-LSTM prediction module, an error redundancy calculation module, an electric energy storage calculation module, and an electric power scheduling calculation module;
the sensor modules are arranged in plurality; the sensor module is in communication connection with the storage module; the storage module is in communication connection with the variable hysteresis module; the variable hysteresis module is in communication connection with the Res-LSTM prediction module, the Res-LSTM prediction module is in communication connection with the error redundancy calculation module, and the error redundancy calculation module is in communication connection with the power scheduling calculation module and the electric energy storage calculation module.
In an alternative embodiment, the sensor modules comprise rain sensors, temperature sensors and humidity sensors, and the sensor modules are distributed in zone 1, zone 2, …, zone N; the rainfall sensor is used for measuring the rainfall of the region; simultaneously converting rainfall information into digital signals and transmitting the digital signals; the temperature sensor monitors the current temperature of the area, converts the current temperature into an available output signal and transmits the available output signal; the humidity sensor can measure the current humidity of the area, convert the current humidity into a usable output signal and transmit the usable output signal.
In an optional embodiment, at least one storage module is arranged in each area; the system is used for storing the historical information and the current information of rainfall, temperature and humidity uploaded by the sensor module.
In an alternative embodiment, the inputs to the variable hysteresis module are the sensory data and the power load history data of each regional sensor module, and the outputs are the data sets used for the Res-LSTM prediction module training or prediction. As shown in fig. 2, the main function of the variable hysteresis module is to shift the environmental history data and the historical power load of each region according to the required prediction span, so as to form a data set for training the load prediction model. The corresponding relation between the original environment variable data and the power load data is as follows:
Figure BDA0002226028500000061
wherein the content of the first and second substances,
Figure BDA0002226028500000062
indicating the amount of rainfall at a certain area time t,
Figure BDA0002226028500000063
which indicates the temperature at a certain time t of the area,
Figure BDA0002226028500000064
indicating the humidity, y, of a certain area over time t(t)Representing load data for a certain area time t.
Assuming that the prediction lag is Δ, that is, the future power load after the Δ time period needs to be predicted according to the current environmental data and the historical load data, after being processed by the variable lag module, the corresponding relationship between the environmental variable data and the power load data is as follows:
Figure BDA0002226028500000065
in an optional embodiment, the Res-LSTM prediction module is configured to train historical data about environmental variables and power loads of each area, which is formed by the variable hysteresis module, and determine parameters of the model so that future power loads can be predicted according to subsequently input environmental variables and historical power loads; as shown in fig. 3, compared with the conventional LSTM model, the network structure of the Res-LSTM model increases Res residual operation, which is mainly to make the selection of the prediction interval of the power load more flexible and avoid the problem of gradient disappearance caused by excessive number of layers of the neural network during model training, and a plurality of neural network units are set in the Res-LSTM model according to the time interval between the environment variable and the historical power load data, each neural network unit takes the environment variable as input, the predicted power load as output, and is marked as a(1),A(2),...,A(t)With neural network element A(t)For example, its environment variable input x(t)And the predicted value y (of the power load)t) The input and output relationship of (1) is as follows:
Figure BDA0002226028500000071
the classical LSTM model is implemented in neural network element A(t-1)Coupled output of
Figure BDA0002226028500000072
And
Figure BDA0002226028500000073
directly as a neural network element A(t)Is connected to the input c(t-1)And a(t-1)Wherein the Res residual operation is to connect neural network element A(t-2)Coupled output c of(t-2)And a(t-2)And neural network element A(t-1)Coupled output of
Figure BDA0002226028500000074
And
Figure BDA0002226028500000075
together, also as neural network element A(t)Is connected to the input c(t-1)And a(t-1)The specific mode of the above operation is as follows:
Figure BDA0002226028500000076
and, there are:
Figure BDA0002226028500000077
similarly, for neural network element A(t-1)Coupled output c of(t-1)And a(t-1)And neural network element A(t)Coupled output of
Figure BDA0002226028500000078
And
Figure BDA0002226028500000079
together, also as neural network element A(t+1)Is connected to the input c(t)And a(t)The specific mode of the operation is as follows:
Figure BDA00022260285000000710
and, there are:
Figure BDA00022260285000000711
the parts of the Res-LSTM prediction module that are the same as the conventional LSTM model will not be described in detail here.
In an alternative embodiment, the error redundancy calculation module calculates a prediction for each region based on the predicted power load for each region in combination with the variance of the historical power loads for that regionThe power is distributed. As shown in FIG. 4, assume that the predicted load for region n at time t is
Figure BDA0002226028500000081
The variance of the regional load historical data is varRegion nAnd the redundancy ratio is theta, the pre-distribution power of the region n at the time t
Figure BDA0002226028500000082
Comprises the following steps:
Figure BDA0002226028500000083
in an optional embodiment, the power scheduling calculation module calculates the actual power load value of each future region, and the actual load of the region n at the time t is recorded as
Figure BDA0002226028500000084
In an optional embodiment, the electric energy storage calculation module calculates the electric energy storage power according to the actual electric load of each region and the pre-distributed power of each region. As shown in fig. 5, the actual power load of each region at time t is calculated by the power scheduling calculation module
Figure BDA0002226028500000085
Calculating the sum Load of the actual power loads of each region at the time t(t)It can be expressed as:
Figure BDA0002226028500000086
meanwhile, the pre-distributed power of each area at the time t is obtained according to the error redundancy calculation module
Figure BDA0002226028500000087
Obtaining the sum P of the pre-distributed power of all the areas at the time t(t)It can be expressed as:
Figure BDA0002226028500000088
the electrical energy storage calculation module may then obtain a calculation of the electrical energy storage value, which may be expressed as:
Ch(t)＝P(t)-Load(t) (11)
wherein, if Ch(t)> 0 indicates that the energy storage device should be in a charged state, whereas Ch(t)< 0 indicates that the energy storage device should be in a discharged state.
A method for predicting power load based on environment variable perception and hysteresis comprises the following specific steps:
s10, the sensor modules arranged in each area periodically monitor information such as rainfall, temperature, humidity and the like, and after the information is collected, the information is uploaded and stored to the storage modules of each area;
s20, collecting power load historical data of each region, and matching the historical data with the sensor data in S10 by using an environment variable hysteresis module through setting a reasonable time interval to form an environment information-power load data set;
and S30, dividing the environmental information-power load data set generated in S20 into a training set, a verification set and a test set according to a certain proportion, training the Res-LSTM model by using the training set, determining the hyper-parameters in the Res-LSTM by using the verification set, and testing the accuracy of the model by using the test set.
S40, inputting the current environmental monitoring data such as rainfall, temperature and humidity in S10 and the historical data of the power load in S20 into the Res-LSTM model trained in S30 to obtain the predicted power load value of each area;
s50, combining the power load values obtained by prediction of each region in S40 with the variance of the power load historical data of each region, and obtaining the pre-distributed power of each region through an error redundancy calculation module;
and S60, inputting the pre-distributed power of each region in the S50 and the actual power load calculated by the power load scheduling module of each region into the electric energy storage calculation module to obtain an electric energy storage calculation value, and providing an operation basis for the actual power scheduling process.
In an alternative embodiment, the training, validation and testing process for the Res-LSTM model in step S30 is as follows:
s31, initializing parameters in Res-LSTM model, including matrix U in each neural network unitf，Uu，Uc，UoAnd matrix Gf，Gu，GoAnd a matrix Wf，Wu，Wc，Wo；
S32, inputting a training set of the data set formulated in S20, and repeatedly iterating parameters in the Res-LSTM model by using a mainstream BPTT algorithm until the algorithm converges or the error is lower than a preset threshold value;
s33, inputting a verification set of the data set formulated in S20, verifying the hyperparameters such as the number of neural network units by using methods such as grid search or random search, and selecting the hyperparameter combination with the minimum generalization error;
and S34, testing the trained Res-LSTM, and if the error is higher than a preset threshold value, re-training and verifying the model until the model precision meets the requirement.
According to the method, environmental factors including temperature, humidity and precipitation are considered, historical data of the regional power load are combined, namely universality of historical rules and specificity of the current condition of the regional power load are considered, and therefore the prediction accuracy is higher compared with the existing power load prediction method.
The invention inputs the environmental data of temperature, humidity, precipitation and the like of a plurality of areas and the historical data of the power load of each area into a Res-LSTM model for prediction, considers the relevance of the power load of each area to a certain extent, and simultaneously realizes the prediction of the power load of the plurality of areas.
After the power load of each region is predicted, the predicted power load is not simply referred to a power system for power scheduling, but the variance of the historical power load prediction of the region is referred to, so that the power of regions with large load variation is pre-distributed with good elasticity.
After the power loads in each area are pre-distributed, the actual power loads are compared in the future, so that reference is made for power dispatching and electric energy storage, and electric energy waste is reduced.
The existing power load monitoring or predicting technology has two defects of large prediction deviation and small prediction area, so the method has certain irreplaceability in the following two practical application scenes:
1. power load fluctuation: in practical application, if the power load in some regions fluctuates greatly between seasons or years due to climate factors or economic operation factors, the existing load prediction method cannot consider the particularity of the power load at the current stage relative to the historical synchronous power load, and the invention can play an irreplaceable role to a certain extent in the scene.
2. Power scheduling changes: in practical application, if a situation that a difference between an actual load and pre-distributed power is large occurs in a part of regions due to some unknown factors, the relevance of power scheduling of each region cannot be considered in the existing method because the power load of the region is simply considered and researched, and the method has certain irreplaceability in this scene.
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.

Claims (7)

1. A power load prediction system based on environment variable perception and hysteresis is characterized by comprising a sensor module, a storage module, a variable hysteresis module, a Res-LSTM prediction module, an error redundancy calculation module, an electric energy storage calculation module and an electric power scheduling calculation module;
the sensor modules are arranged in plurality; the sensor module is in communication connection with the storage module; the storage module is in communication connection with the variable hysteresis module; the variable lag module is in communication connection with the Res-LSTM prediction module, and the Res-LSTM prediction module is in communication connection with the error redundancy calculation module;
the input of the variable hysteresis module is sensing data and power load historical data of each regional sensor module, and the output is a data set used for Res-LSTM prediction module training or prediction;
the variable hysteresis module is used for shifting the environmental historical data and the historical power load of each region according to the required prediction span so as to form a data set to train a load prediction model behind the data set;
the corresponding relation between the original environment variable data and the power load data is as follows:
Figure FDA0002801396380000011
wherein the content of the first and second substances,
Figure FDA0002801396380000012
indicating the amount of rainfall at a certain area time t,
Figure FDA0002801396380000013
which indicates the temperature at a certain time t of the area,
Figure FDA0002801396380000014
indicating the humidity, y, of a certain area over time t(t)Load data indicating a time t of a certain area;
assuming that the prediction lag is Δ, that is, the future power load after the Δ time period is predicted according to the current environmental data and the historical load data, after being processed by the variable lag module, the corresponding relationship between the environmental variable data and the power load data is as follows:
Figure FDA0002801396380000015
the Res-LSTM prediction module is used for training historical data which are formed by the variable hysteresis module and are related to environment variables and power loads of each region to determine parameters of the model and predicting future power loads according to the environment variables and the historical power loads which are input subsequently;
adding Res residual operation to the network structure of the Res-LSTM model;
setting a plurality of neural network units in the Res-LSTM model according to the time interval between the environment variable and the historical power load data, wherein each neural network unit takes the environment variable as input and the predicted power load as output and is marked as A(1),A(2),...,A(t)With neural network element A(t)For example, its environment variable input x(t)And the predicted value y of the power load(t)The input and output relationship of (1) is as follows:
Figure FDA0002801396380000021
the error redundancy calculation module is in communication connection with the power scheduling calculation module and the electric energy storage calculation module;
the error redundancy calculation module is used for calculating a pre-distribution power for each region according to the predicted power load of each region and the variance of the historical power load of the region;
wherein, the predicted load of the region n at time t is assumed to be
Figure FDA0002801396380000022
The variance of the regional load historical data is varRegion nAnd the redundancy ratio is theta, the pre-distribution power of the region n at the time t
Figure FDA0002801396380000023
Comprises the following steps:
Figure FDA0002801396380000024
2. the system according to claim 1, wherein the sensor modules comprise rain sensors, temperature sensors and humidity sensors, and are distributed in zone 1, zone 2, …, zone N; the rainfall sensor is used for measuring the rainfall of the region; simultaneously converting rainfall information into digital signals and transmitting the digital signals; the temperature sensor monitors the current temperature of the area, converts the current temperature into an available output signal and transmits the available output signal; the humidity sensor can measure the current humidity of the area, convert the current humidity into a usable output signal and transmit the usable output signal.
3. The system according to claim 2, wherein at least one memory module is disposed in each region; the system is used for storing the historical information and the current information of rainfall, temperature and humidity uploaded by the sensor module.
4. The system according to claim 1, wherein the power scheduling calculation module calculates actual power load values of each future region;
wherein the actual load of the region n at time t is recorded as
Figure FDA0002801396380000031
5. The system according to claim 1, wherein the electrical energy storage calculation module calculates the electrical energy storage power according to the actual electrical load of each region and the pre-allocated power of each region;
which comprises the following steps:
firstly, the actual power load of each region at the time t is calculated and obtained according to a power scheduling calculation module
Figure FDA0002801396380000032
Calculating the sum Load of the actual power loads of each region at the time t(t)Expressed as:
Figure FDA0002801396380000033
meanwhile, the pre-distributed power of each area at the time t is obtained according to the error redundancy calculation module
Figure FDA0002801396380000034
Obtaining the sum P of the pre-distributed power of all the areas at the time t(t)Expressed as:
Figure FDA0002801396380000035
then, the electric energy storage calculation module obtains a calculation result of the electric energy storage value, which is expressed as:
Ch(t)＝P(t)-Load(t)；
if Ch(t)> 0 indicates that the energy storage device should be in a charged state;
if Ch(t)< 0 indicates that the energy storage device should be in a discharged state.
6. The system according to any one of claims 1-5, further comprising a method for predicting an electrical load based on the perception and the hysteresis of the environmental variables, comprising the steps of:
s10, the sensor modules arranged in each area periodically monitor information such as rainfall, temperature, humidity and the like, and after the information is collected, the information is uploaded and stored to the storage modules of each area;
s20, collecting power load historical data of each region, and matching the historical data with the sensor data in S10 by using an environment variable hysteresis module through setting a reasonable time interval to form an environment information-power load data set;
s30, dividing the environmental information-power load data set generated in S20 into a training set, a verification set and a test set according to a certain proportion, training a Res-LSTM model by using the training set, determining a hyper-parameter in the Res-LSTM by using the verification set, and testing the accuracy of the model by using the test set;
s40, inputting the current environmental monitoring data such as rainfall, temperature and humidity in S10 and the historical data of the power load in S20 into the Res-LSTM model trained in S30 to obtain the predicted power load value of each area;
s50, combining the power load values obtained by prediction of each region in S40 with the variance of the power load historical data of each region, and obtaining the pre-distributed power of each region through an error redundancy calculation module;
and S60, inputting the pre-distributed power of each region in the S50 and the actual power load calculated by the power load scheduling module of each region into the electric energy storage calculation module to obtain an electric energy storage calculation value, and providing an operation basis for the actual power scheduling process.
7. The method of claim 6, wherein the method comprises the steps of: the training, verifying and testing process of the Res-LSTM model in step S30 is as follows:
s31, initializing parameters in Res-LSTM model, including matrix U in each neural network unitf，Uu，Uc，UoAnd matrix Gf，Gu，GoAnd a matrix Wf，Wu，Wc，Wo；
S32, inputting a training set of the data set formulated in S20, and repeatedly iterating parameters in the Res-LSTM model by using a mainstream BPTT algorithm until the algorithm converges or the error is lower than a preset threshold value;
s33, inputting a verification set of the data set formulated in S20, verifying the hyperparameters such as the number of neural network units by using methods such as grid search or random search, and selecting the hyperparameter combination with the minimum generalization error;
and S34, testing the trained Res-LSTM, and if the error is higher than a preset threshold value, re-training and verifying the model until the model precision meets the requirement.
CN201910951985.9A 2019-10-09 2019-10-09 Power load prediction system and method based on environment variable perception and hysteresis Active CN110601193B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910951985.9A CN110601193B (en) 2019-10-09 2019-10-09 Power load prediction system and method based on environment variable perception and hysteresis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910951985.9A CN110601193B (en) 2019-10-09 2019-10-09 Power load prediction system and method based on environment variable perception and hysteresis

Publications (2)

Publication Number Publication Date
CN110601193A CN110601193A (en) 2019-12-20
CN110601193B true CN110601193B (en) 2021-02-05

Family

ID=68865855

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910951985.9A Active CN110601193B (en) 2019-10-09 2019-10-09 Power load prediction system and method based on environment variable perception and hysteresis

Country Status (1)

Country Link
CN (1) CN110601193B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113033898A (en) * 2021-03-26 2021-06-25 国核电力规划设计研究院有限公司 Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network
CN113224852A (en) * 2021-05-28 2021-08-06 广东电网有限责任公司广州供电局 Power communication management method and system based on AI auxiliary decision

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102968111B (en) * 2012-12-14 2015-09-02 廊坊新奥节能服务有限公司 Control the method and system of distributed energy resource system
CN105226659B (en) * 2015-10-27 2018-01-19 国家电网公司 A kind of multizone supports the micro-grid system of independent operating each other
CN105243454A (en) * 2015-11-06 2016-01-13 广州威沃电子有限公司 Big data-based electrical load prediction system
CN107562784A (en) * 2017-07-25 2018-01-09 同济大学 Short text classification method based on ResLCNN models
CN107704966A (en) * 2017-10-17 2018-02-16 华南理工大学 A kind of Energy Load forecasting system and method based on weather big data
CN109800898A (en) * 2017-11-17 2019-05-24 中国电力科学研究院有限公司 A kind of intelligence short-term load forecasting method and system

Also Published As

Publication number Publication date
CN110601193A (en) 2019-12-20

Similar Documents

Publication Publication Date Title
CN107591844B (en) Active power distribution network robust reconstruction method considering node injection power uncertainty
CN105740975B (en) A kind of equipment deficiency assessment and prediction technique based on data correlation relation
CN110601193B (en) Power load prediction system and method based on environment variable perception and hysteresis
CN104820876A (en) Short-term load forecasting method and system
Hu et al. A stochastic model based energy management system for off-grid solar houses
CN117318111B (en) Weather prediction-based dynamic adjustment method and system for light energy storage source
JP5923454B2 (en) Power system management apparatus and method
Chamorro et al. Information length quantification and forecasting of power systems kinetic energy
CN115514100A (en) Hybrid energy storage system based on multi-element energy storage and control
Laouafi et al. An evaluation of conventional and computational intelligence methods for medium and long-term load forecasting in Algeria
CN116544934B (en) Power scheduling method and system based on power load prediction
CN105958474B (en) Dynamic capacity increasing method and system for power transmission line for power grid regulation and control system
Ávila et al. Fuzzy demand forecasting in a predictive control strategy for a renewable-energy based microgrid
CN115907228A (en) Short-term power load prediction analysis method based on PSO-LSSVM
TW201727559A (en) Management method and system of renewable energy power plant checking whether the power generation of a renewable energy power plant is normal according to the estimated power generation amount
Gofman et al. Increasing the accuracy of the short-term and operational prediction of the load of a power system using an artificial neural network
US20220284156A1 (en) Real-time update of power system models for dynamic security assessment
Songpu et al. Domestic demand predictions considering influence of external environmental parameters
Maalej et al. Sensor data augmentation strategy for load forecasting in smart grid context
Songpu et al. External parameters contribution in domestic load forecasting using neural network
CN112819307A (en) Demand response method and system based on load supervision in smart power grid
Dadi et al. Estimation of transmission reliability margin using statistical methods and intelligent techniques
Merce et al. Load and demand side flexibility forecasting
Pusat Effect of time horizon on wind speed prediction with ANN
CN117575291B (en) Federal learning data collaborative management method based on edge parameter entropy

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
GR01 Patent grant
GR01 Patent grant