CN117039852A - Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station - Google Patents
Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station Download PDFInfo
- Publication number
- CN117039852A CN117039852A CN202310896582.5A CN202310896582A CN117039852A CN 117039852 A CN117039852 A CN 117039852A CN 202310896582 A CN202310896582 A CN 202310896582A CN 117039852 A CN117039852 A CN 117039852A
- Authority
- CN
- China
- Prior art keywords
- wind
- data
- preset
- training
- power
- 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
- 238000000034 method Methods 0.000 title claims abstract description 43
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 36
- 230000000295 complement effect Effects 0.000 title claims abstract description 28
- 238000003860 storage Methods 0.000 title claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 170
- 238000013499 data model Methods 0.000 claims abstract description 54
- 238000005259 measurement Methods 0.000 claims abstract description 28
- 238000013528 artificial neural network Methods 0.000 claims abstract description 24
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000012544 monitoring process Methods 0.000 claims abstract description 14
- 230000006870 function Effects 0.000 claims description 17
- 230000008859 change Effects 0.000 claims description 11
- 238000012546 transfer Methods 0.000 claims description 4
- 238000010248 power generation Methods 0.000 description 34
- 238000004891 communication Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006872 improvement Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 4
- 238000005859 coupling reaction Methods 0.000 description 4
- 238000005086 pumping Methods 0.000 description 4
- 230000001105 regulatory effect Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000005622 photoelectricity Effects 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/381—Dispersed generators
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
-
- 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
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/22—The renewable source being solar energy
- H02J2300/24—The renewable source being solar energy of photovoltaic origin
-
- 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
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/28—The renewable source being wind energy
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a power prediction method, a device, equipment and a storage medium of a wind-solar-water complementary power station, wherein the method comprises the steps of obtaining historical wind-solar data of the wind power station and a photoelectric station in a first preset time period through a weather station; defining a training model of wind and light data through a first preset algorithm; training the training model by using a neural network based on a preset data set to obtain a wind-light data model based on a first preset duration; acquiring a real-time stamp of the energy base region, and acquiring model data of a second preset duration in the wind-light data model by taking the real-time stamp as a measurement and report starting point as measurement and report reference data; acquiring wind and light trend of the measurement and report reference data; and outputting the calculated power result to a computer monitoring system of the hydropower station. According to the wind-light trend prediction method based on the neural network training, wind-light trend of a certain duration in the future based on the real-time stamp can be predicted according to the data model completed by the neural network training, and the result is output to adjust the operation of the unit, so that the safety and stability of a power grid and the unit are ensured.
Description
Technical Field
The application relates to the fields of new energy technology and artificial intelligence, in particular to a power prediction method, device, equipment and storage medium of a wind-solar-water complementary power station.
Background
In order to meet the construction requirements of a novel power system, the construction of the wind, light and water storage integrated clean energy base has a positive practical effect. The base takes hydropower (including pumping storage) in the river basin as an adjusting power supply, and the hydropower and wind power are coupled and output by solar power generation to be sent out in a complementary mode. In the new energy power generation of three main flows of wind, light and water, the real-time power of wind power and photoelectricity is greatly influenced by weather, weather and day and night, so that the power generation curves of wind power and photoelectricity have severe fluctuation and steep rise and fall, and have obvious randomness and fluctuation; the hydroelectric generation has stronger stability, so that the hydroelectric generating set is required to adjust the active power output according to the load characteristic of the power grid and the real-time power of new energy in order to ensure the balance and stability of the active power of the power grid, and the stability of the power grid is ensured while the load demand of the power system and the stable operation of the hydroelectric generating set are met.
The existing wind, light and water complementary control model is not formed, and in the prior art, for example, a pumped storage wind power coordinated control model adopts a load curve to replace each active power output, and the active power dynamic performance of a water motor unit is not considered; the cooperative game method has huge calculated amount and larger hysteresis when large-scale power generation members participate in the wind-light-water complementary power generation system, so that fluctuation of a wind power generation end can be fed back to a power grid to cause unstable power grid, or a hydroelectric generating set is caused to frequently pass through a vibration area to run, and the hydroelectric generating set is forced to discard light and wind in order to avoid the vibration area to run.
Disclosure of Invention
The application mainly aims to provide a power prediction method, device, equipment and storage medium for a wind-solar-water complementary power station, which are used for accurately predicting power of a photovoltaic power generation and wind power generation power and change time reaching a minute level, providing basic data for unit coupling operation control, automatic Generation Control (AGC) system control and the like of a wind-solar hydropower station (including pumping storage) in a clean energy base, and solving the problems of power grid fluctuation and frequent unit vibration caused by wind, light and water complementary regulation in the prior art.
In order to achieve the above purpose, the present application provides the following technical solutions:
a power prediction method for a wind-solar-water complementary power station, the wind-solar-water complementary power station comprising a meteorological station and a power station built in an energy base area, the power station comprising a wind power station, a photovoltaic power station and a hydropower station which are operated in a grid-connected mode, the hydropower station comprising a conventional hydropower station and a pumped storage power station, the power prediction method comprising:
acquiring historical wind and light data of the wind power station and the photoelectric station within a first preset duration through the meteorological station;
defining a training model of the wind-solar data through a first preset algorithm;
training the training model by using a neural network based on a preset data set to obtain a wind-light data model based on the first preset duration;
acquiring a real-time stamp of the energy base region, and acquiring model data of a second preset duration in the wind-solar data model by taking the real-time stamp as a measurement and report starting point as measurement and report reference data;
acquiring the trend of the wind and light trend of the forecast reference data;
and calculating predicted power according to the trend of the wind and light, and outputting a result to a computer monitoring system of the hydropower station.
As a further improvement of the present application, defining the training model of the wind-solar data by a first preset algorithm includes:
all characteristics of the wind and light data are acquired and respectively defined as an input node;
each input node is respectively endowed with a preset weight;
defining the training model according to equation (1):
wherein y is the training model; x is x n For the n-th input node,the preset weight values from the mth input node of the input layer to the nth input node of the hidden layer are obtained; />A threshold value for the n-th input node of the hidden layer; tan sig (x) is the transfer function, and +.>
As a further improvement of the present application, performing neural network training on the training model based on a preset data set to obtain a wind-solar data model based on the first preset duration, including:
training the training model by using a neural network based on a preset data set;
iteratively updating all preset weights of the training model by a back propagation algorithm with preset times so that the value of a loss function of the training model is smaller than or equal to a second preset threshold value;
and acquiring a training model after training and defining the training model as the wind-light data model.
As a further improvement of the present application, iteratively updating all preset weights of the training model by a back propagation algorithm at preset times so that a value of a loss function of the training model is less than or equal to a second preset threshold value, including:
acquiring an error signal after iterative updating of the training model;
back-propagating the error signal to an output layer of the training model to obtain an error value (2) of the output layer;
wherein E is the error value; target is the expected output of the training model; output is the actual output of the training model;
sequentially updating preset weights of the output layer, the hidden layer and the input layer according to the error value;
and performing the next iteration on the training model based on the updated preset weight until the value of the loss function of the training model is smaller than or equal to the second preset threshold.
As a further improvement of the present application, obtaining the wind-light trend of the forecast reference data includes:
establishing a rectangular coordinate system by taking natural time as an x axis and taking the numerical value of the model data as a y axis;
outputting the measurement and report reference data in the rectangular coordinate system to form a data curve;
outputting the real-time stamp in the rectangular coordinate system and acquiring a curve slope of the second preset duration of the data curve based on the real-time stamp;
and obtaining a change curve of the slope of the curve in the second preset time period, wherein the change curve is the trend of the wind and light trend.
In order to achieve the above purpose, the present application further provides the following technical solutions:
the power prediction device of the wind-light-water complementary power station is applied to the power prediction method of the wind-light-water complementary power station, and comprises the following components:
the historical wind-light data acquisition module is used for acquiring historical wind-light data of the wind power station and the photoelectric station in a first preset duration through the meteorological station;
the training model definition module is used for defining a training model of the wind-solar data through a first preset algorithm;
the wind-solar data model training module is used for training the training model by using a neural network based on a preset data set to obtain a wind-solar data model based on the first preset duration;
the measurement and report reference data acquisition module is used for acquiring a real-time stamp of the energy base area, and acquiring model data with a second preset duration in the wind-solar data model by taking the real-time stamp as a measurement and report starting point as measurement and report reference data;
the wind-light trend acquisition module is used for acquiring the wind-light trend of the observation and report reference data;
and the predicted power output module is used for calculating predicted power according to the trend of the wind and light trend, and outputting the result to a computer monitoring system of the hydropower station.
In order to achieve the above purpose, the present application further provides the following technical solutions:
an electronic device comprising a processor, a memory coupled to the processor, the memory storing program instructions executable by the processor; and the processor executes the program instructions stored in the memory to realize the power prediction method of the wind-solar-water complementary power station.
In order to achieve the above purpose, the present application further provides the following technical solutions:
a storage medium having stored therein program instructions which, when executed by a processor, implement a power prediction method capable of implementing a wind-solar-water complementary power station as described above.
According to the application, the historical wind-light data generated by the three parties are trained by a neural network, and model data with a certain time length in the future is intercepted in real time in a wind-light data model after the training is completed to be used as measurement and report reference data; calculating predicted power according to wind-light trend of a certain time length in the future, which is obtained from the forecast reference data, and outputting the result to a hydropower station computer monitoring system; the method and the system can accurately predict the wind power and photovoltaic power changes in the clean energy base in minute level, so that the hydroelectric generating set can be quickly and early regulated to corresponding standby operation working conditions or a plurality of machines can distribute the existing load, the hydroelectric generating set can timely regulate the load when the wind-solar fluctuation power supply changes, the fluctuation of the power grid is reduced, the power generation quality is improved, meanwhile, the pre-regulation capacity can reduce the frequency of the hydroelectric generating set frequently passing through a vibration area, the forced wind and light discarding of the generating set for avoiding the operation vibration area is avoided, and the safe and stable operation of the generating set and the power grid is ensured while the power requirement of a power station is met.
Drawings
FIG. 1 is a schematic diagram of the steps in a flow chart of an embodiment of a power prediction method for a wind-solar-water complementary power station of the present application;
FIG. 2 is a schematic diagram of functional modules of an embodiment of a power prediction apparatus for a wind-solar-water hybrid power station according to the present application;
FIG. 3 is a schematic diagram of an embodiment of an electronic device of the present application;
FIG. 4 is a schematic diagram illustrating the structure of an embodiment of a storage medium according to the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The terms "first," "second," "third," and the like in this disclosure are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "first," "second," and "third" may explicitly or implicitly include at least one such feature. In the description of the present application, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise. All directional indications (such as up, down, left, right, front, back … …) in embodiments of the present application are merely used to explain the relative positional relationship, movement, etc. between the components in a particular gesture (as shown in the drawings), and if the particular gesture changes, the directional indication changes accordingly. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
As shown in fig. 1, the present embodiment provides a power prediction method for a wind-solar-water complementary power station, where the wind-solar-water complementary power station includes a meteorological station and a power station built in an energy base area, the power station includes a wind power station, a photovoltaic power station, and a hydropower station that are operated in a grid-connected mode, and the hydropower station includes a conventional hydropower station and a pumped storage power station, and the power prediction method in this embodiment includes the following steps:
step S1, historical wind-light data of a wind power station and a photoelectric station in a first preset duration are obtained through a weather station.
And S2, defining a training model of wind and light data through a first preset algorithm.
And step S3, training the training model through a neural network based on a preset data set to obtain a wind-light data model based on a first preset duration.
And S4, acquiring a real-time stamp of the energy base region, and acquiring model data of a second preset duration in the wind-light data model by taking the real-time stamp as a measurement and report starting point as measurement and report reference data.
And S5, acquiring the trend of the wind and light trend of the forecast reference data.
And S6, calculating predicted power according to the trend of the wind and light trend, and outputting the result to a computer monitoring system of the hydropower station.
Preferably, a weather station and a wind-light forecasting workstation can be built in the energy base area, wind-light data (wind power, wind direction, cloud layer, cloud speed and the like) are measured through the weather station, cloud images are collected and transmitted to the workstation, the wind-light forecasting workstation adopts AI learning and correction according to the measured data and external input information, a self-learning environment database is built, the generated energy and change time of photovoltaic power generation and wind power generation are finely predicted, and basic data are provided for unit operation control of hydropower stations and pumping and accumulating power stations, an automatic power generation control (AGC) system and the like.
Preferably, the weather station is provided with a wind direction sensor, a wind speed sensor, an air pressure sensor, an air temperature sensor, a humidity sensor, a rainfall sensor, a cloud height instrument, an atmosphere transmission instrument, a background light brightness sensor, an omnibearing camera and other instruments for collecting real-time weather data in an energy base area; multiple weather stations may also be provided depending on the size of the clean energy base.
Preferably, the weather station is provided with a PLC controller of a storage system, so that the timing acquisition and storage of the original data can be completed, the automatic alarm can be timely given when the equipment fails, and meanwhile, a communication signal can be sent to the wind-light measuring and reporting workstation in real time. When communication is interrupted and the fault occurs, the controller can circularly store data for more than 7 days and transmit the stored data to the test report platform after the communication is recovered.
Preferably, when the weather station is far away from the electricity utilization area or an available external power supply cannot be adopted, the weather station can use the electricity generated by the light power station and store energy by a power generation and distribution battery, but wind power generation is not adopted so as not to influence wind power measurement.
Preferably, the signal transmission between the weather station and the wind-light measuring and forecasting workstation can adopt optical fiber, wireless network connection or satellite signal transmission, and can be determined according to implementation conditions.
Preferably, the wind-solar measuring and reporting workstation comprises an AI server, platform software and the like, is arranged in a central control room of a certain power supply point in a base or a base centralized control center and is used for receiving and storing data of a meteorological station and AI calculation, predicting the expected time of large-range fluctuation of the power generation power and the output power of solar energy or wind power in set time through an AI algorithm, transmitting the data to an AGC (automatic gain control) or a power station computer for monitoring, and combining other control parameters, so that a hydroelectric generating set and a pumping and storing set in the hydroelectric generating station enter a preset operation mode in advance.
Preferably, the steps are performed simultaneously, and real-time wind and light data can be acquired and added into a training data set, so that the data model is continuously trained through the real-time data.
Preferably, in practical application, the BP neural network establishes a mapping relation by learning and storing a large number of input-output modes, the learning rule is a gradient descent method, the principle is that the weight and bias of the network are adjusted through establishing a forward propagation process and through back propagation according to the error between the predicted output and the actual value, so that the sum of squares of a loss function reaches the minimum usable condition, and when the engineering requirement is met, a training model is output for monitoring and using of water motor group equipment and the like in a clean energy base, and during practical application, wind power generation and photovoltaic power generation can be divided into two parts for training respectively.
Further, the step S2 specifically includes the following steps:
step S21, all characteristics of wind and light data are acquired and respectively defined as an input node.
Step S22, each input node is respectively given a preset weight.
Step S23, defining a training model according to the formula (1):
wherein y is a training model; x is x n For the n-th input node,the preset weight values from the mth input node of the input layer to the nth input node of the hidden layer are obtained; />A threshold value for the n-th input node of the hidden layer; tan sig (x) is the transfer function, and +.>
Preferably, the neural network used in the embodiment only includes an hidden layer, where the input nodes include data such as wind power, wind direction, cloud layer, cloud speed, temperature, humidity, illumination duration, and the like, and each of the foregoing data is sequentially used as an input node.
Preferably, training a model to train a neural network typically requires providing a large amount of data, i.e., a data set; data sets are generally divided into three classes, namely training set (training set), validation set (validation set) and test set (test set).
One epoch is a process equal to one training time using all samples in the training set, and the training time refers to one forward propagation (forward pass) and one backward propagation (back pass); when the number of samples of one epoch (i.e., the training set) is too large, excessive time may be consumed for performing one training, and it is not necessary to use all data of the training set for each training, the whole training set needs to be divided into a plurality of small blocks, i.e., a plurality of latches for training; one epoch is made up of one or more latches, which are part of a training set, with only a portion of the data being used for each training process, i.e., one latch, and one latch being trained as an iteration.
Preferably, the neural network training specifically includes a Perceptron (Perceptron), the Perceptron is composed of two layers of neurons, an input layer receives an external input signal and transmits the external input signal to an output layer, the output layer is an M-P neuron, and if the formula (1) is a step function, then:
y j ＝f(∑ i w i ·x i -θ i ) ①。
preferably, given a training data set, the weights w i (i=1, 2,., n) and training threshold θ i Can be obtained by learning, theta i It can be understood that a weight w corresponding to a fixed value of-1, 0 is fixedly input i+1 。
Preferably, the first preset number of times may be set to 200 times.
Preferably, the learning rate of 1 st to 100 th epochs may be set to 0.01, the learning rate of 101 st to 150 th epochs may be set to 0.001, and the learning rate of 151 st to 200 th epochs may be set to 0.0001.
Illustrating:
training learning of AI for photovoltaic power generation prediction: the method comprises the steps of identifying the state and the property of remote cloud according to a camera picture, calculating the state and the time when the cloud reaches the upper air of a photovoltaic area according to measured data such as wind power, wind direction, cloud layer, cloud speed, temperature, humidity and the like, calculating photovoltaic power generation amount according to the irradiation angle, the sunshine time, the light intensity and the like of the sun within one year, comparing the calculated result with actual photovoltaic power generation amount, and then carrying out learning correction, wherein external influence factors comprise factors such as actual attenuation of a photovoltaic panel, cleaning state of the photovoltaic panel, equipment utilization rate, chemical energy storage capacity and the like.
Preferably, the AI training period of a complete photovoltaic power generation prediction is one year, the overall construction period of the clean energy base is long, the construction period of the artificial intelligent wind-light measuring and reporting platform is short, in order to enable the project to start training as early as possible, the artificial intelligent wind-light measuring and reporting platform can be built before the clean energy base is not built, illumination measuring points are arranged in a photovoltaic panel area, radiation sensors, scattering sensors and the like are arranged, photovoltaic power generation power in each daily period is simulated for identification and prediction training of cloud graphics, automatic learning and correction are carried out after photovoltaic power generation is put into production by using the practical output of photovoltaic, so that learning efficiency is improved, a model is built, prediction accuracy is improved, and simulation is realized on a server through subroutines of platform software.
Further, training learning of AI for wind power generation: and calculating the generated energy and time of the wind power generation area according to wind power, wind direction and the like, and comparing the calculated result with the actual generated energy to perform learning correction. The AI training learning process of wind power generation can also start training as early as possible before the wind farm is built, measurement points can be set in a fan arrangement area, wind power generation power of each period of each day is simulated, actual output of a fan is utilized to automatically learn and correct after the fan is generated and put into operation, so that learning efficiency is improved, a model is built, prediction accuracy is improved, and simulation is realized on a server by a subroutine of platform software.
Further, the step S3 specifically includes the following steps:
step S31, training the neural network on the training model based on the preset data set.
Step S32, iteratively updating all preset weights of the training model by a back propagation algorithm for preset times so that the value of the loss function of the training model is smaller than or equal to a second preset threshold value.
And step S33, acquiring a training model after training and defining the training model as a wind-light data model.
Further, the step S32 specifically includes the following steps:
step S321, obtaining an error signal after iterative updating of the training model.
In step S322, the error signal is back propagated to the output layer of the training model, so as to obtain an error value (2) of the output layer.
Wherein E is an error value; target is the expected output of the training model; output is the actual output of the training model.
Step S323, the preset weights of the output layer, the hidden layer and the input layer are updated in sequence according to the error value.
Step S324, performing the next iteration on the training model based on the updated preset weight until the value of the loss function of the training model is smaller than or equal to a second preset threshold.
Preferably, in this embodiment, when the error value E is less than or equal to 5% (when the predicted power value range is greater than or equal to 50% of the wind-solar installation), or when the error value E is less than or equal to 8% (when the predicted power value range is less than 50% of the wind-solar installation), the training model of this embodiment is output, and the model calculation after training is completed is used for monitoring and controlling the water motor group equipment in the clean energy base.
Preferably, the embodiment can be further carried in platform comprehensive software, and the platform comprehensive software also has the functions of historical meteorological data statistics query, report generation, extreme climate early warning, measuring element fault warning, photovoltaic power generation and wind power generation abnormality warning, photovoltaic power generation and wind power generation capacity prediction for the future month, season and year. And providing related data for power generation mission planning, operation and maintenance management and the like of the energy base; the platform has enough communication interfaces, can communicate with a computer monitoring system or AGC of a power station in the base in real time according to a specified communication protocol, and finally completes intelligent control operation of the unit.
Further, the step S5 specifically includes the following steps:
in step S51, a rectangular coordinate system is established with the natural time as the x-axis and the numerical value of the model data as the y-axis.
Step S52, outputting the forecast reference data in the rectangular coordinate system to form a data curve.
And step S53, outputting the real-time stamp in the rectangular coordinate system and acquiring the curve slope of the second preset duration of the data curve based on the real-time stamp.
And S54, obtaining a change curve of the slope of the curve in a second preset time length, wherein the change curve is the trend of the wind and light.
According to the embodiment, the neural network training is carried out on historical wind-light data generated by three parties, and model data with a certain time length in the future is intercepted in a wind-light data model after the training is completed in real time to serve as measurement and report reference data; and calculating predicted power according to the wind-light trend of a certain time length in the future obtained from the forecast reference data, and outputting the result to a hydropower station computer monitoring system. . The embodiment utilizes the controllable characteristic of hydroelectric generation, and can predict the trend of wind and light trend for a certain time length in the future based on a real-time stamp according to the data model after the training of the neural network while meeting the total output power of the power station; the training model of the embodiment carries out minute-level accurate prediction on wind power and photovoltaic power changes in the clean energy base, so that the hydroelectric generating set can be quickly and early regulated to corresponding standby operation working conditions or a plurality of machines are distributed with existing loads, the hydroelectric generating set can timely regulate loads when wind and light fluctuation power supplies change, power grid fluctuation is reduced, power generation quality is improved, meanwhile, the pre-regulation capacity can reduce the frequency of the hydroelectric generating set frequently passing through a vibration area, the situation that the generating set is forced to abandon wind and light for avoiding the operation vibration area is avoided, and the safe and stable operation of the generating set and the power grid is ensured while the power requirement of a power station is met.
As shown in fig. 2, this embodiment provides an embodiment of a power prediction apparatus of a wind-solar-water complementary power station, and referring to fig. 2, in this embodiment, the power prediction apparatus of a wind-solar-water complementary power station is applied to a power prediction method of a wind-solar-water complementary power station in the foregoing embodiment, where the power prediction apparatus includes a historical wind-solar data acquisition module 1, a training model definition module 2, a wind-solar data model training module 3, a measurement and report reference data acquisition module 4, a wind-solar trend acquisition module 5, and a predicted power output module 6 that are electrically connected in sequence.
The historical wind-light data acquisition module 1 is used for acquiring historical wind-light data of a wind power station and a photoelectric station in a first preset duration through a meteorological station; the training model definition module 2 is used for defining a training model of wind and light data through a first preset algorithm; the wind-solar data model training module 3 is used for training the training model by using a neural network based on a preset data set to obtain a wind-solar data model based on a first preset duration; the measurement and report reference data acquisition module 4 is used for acquiring a real-time timestamp of the energy base area, and acquiring model data of a second preset duration in the wind-light data model by taking the real-time timestamp as a measurement and report starting point as measurement and report reference data; the wind-light trend acquisition module 5 is used for acquiring wind-light trend of the measurement and report reference data; the predicted power output module 6 is used for calculating predicted power according to the trend of the wind and light trend, and outputting the result to the computer monitoring system of the hydropower station. .
Further, the training model definition module comprises a first training model definition sub-module, a second training model definition sub-module and a third training model definition sub-module which are electrically connected in sequence; the first training model definition submodule is electrically connected with the historical wind and light data acquisition module, and the third training model definition submodule is electrically connected with the wind and light data model training module.
The first training model definition submodule is used for acquiring all characteristics of wind and light data and respectively defines the characteristics as an input node.
The second training model definition submodule is used for respectively giving a preset weight value to each input node.
The third training model definition sub-module is for defining a training model according to equation (1):
wherein y is a training model; x is x n For the n-th input node,the preset weight values from the mth input node of the input layer to the nth input node of the hidden layer are obtained; />A threshold value for the n-th input node of the hidden layer; tan sig (x) is a transfer function, and
further, the wind-light data model training module comprises a first wind-light data model training sub-module, a second wind-light data model training sub-module and a third wind-light data model training sub-module which are electrically connected in sequence; the first wind-light data model training sub-module is electrically connected with the third training model defining sub-module, and the third wind-light data model training sub-module is electrically connected with the forecast reference data acquisition module.
The first wind-solar data model training submodule is used for training the training model by using a neural network based on a preset data set; the second wind-solar data model training submodule is used for iteratively updating all preset weights of the training model with preset times through a back propagation algorithm so that the value of a loss function of the training model is smaller than or equal to a second preset threshold value; the third wind-light data model training submodule is used for acquiring a training model after training is completed and defining the training model as a wind-light data model.
Further, the second wind-light data model training submodule further comprises a first wind-light data model training unit, a second wind-light data model training unit, a third wind-light data model training unit and a fourth wind-light data model training unit which are electrically connected in sequence; the first wind-light data model training unit is electrically connected with the first wind-light data model training submodule, and the fourth wind-light data model training unit is electrically connected with the third wind-light data model training submodule.
The first wind-solar data model training unit is used for acquiring error signals after iterative updating of the training model.
The second wind-solar data model training unit is used for reversely transmitting the error signal to an output layer of the training model to obtain an error value (2) of the output layer.
Wherein E is an error value; target is the expected output of the training model; output is the actual output of the training model.
The third wind-solar data model training unit is used for updating preset weights of the output layer, the hidden layer and the input layer in sequence according to the error value.
The fourth wind-solar data model training unit is used for carrying out next iteration on the training model based on the updated preset weight until the value of the loss function of the training model is smaller than or equal to a second preset threshold value.
The wind-light trend acquiring module comprises a first wind-light trend acquiring sub-module, a second wind-light trend acquiring sub-module, a third wind-light trend acquiring sub-module and a fourth wind-light trend acquiring sub-module which are electrically connected in sequence; the first wind and light trend acquisition sub-module is electrically connected with the observation and prediction reference data acquisition module, and the fourth wind and light trend acquisition sub-module is electrically connected with the predicted power output module.
The first wind-solar trend acquisition submodule is used for establishing a rectangular coordinate system by taking natural time as an x axis and taking a numerical value of model data as a y axis; the second wind-solar trend acquisition submodule is used for outputting measurement and report reference data in a rectangular coordinate system to form a data curve; the third wind-solar trend obtaining submodule is used for outputting a real-time timestamp in the rectangular coordinate system and obtaining a curve slope of a second preset duration of the data curve based on the real-time timestamp; the fourth wind-light trend obtaining submodule is used for obtaining a change curve of the slope of the curve in a second preset duration, and the change curve is the wind-light trend.
It should be noted that, the present embodiment is an apparatus embodiment based on the foregoing method embodiment, and the expansion and limitation of the present embodiment may be referred to the foregoing method embodiment, which is not described in detail.
According to the embodiment, the neural network training is carried out on historical wind-light data generated by three parties, and model data with a certain time length in the future is intercepted in a wind-light data model after the training is completed in real time to serve as measurement and report reference data; and calculating predicted power according to the wind-light trend of a certain time length in the future obtained from the forecast reference data, and outputting the result to a hydropower station computer monitoring system. The application utilizes the controllable characteristic of hydroelectric generation, and can predict the trend of wind and light trend in a certain time based on the real-time stamp according to the data model trained by the neural network while meeting the total output power of the power station. The training model disclosed by the application can be used for accurately predicting the wind power and photovoltaic power changes in a clean energy base in a minute level, so that the hydroelectric generating set can be quickly and early regulated to corresponding standby operation working conditions or a plurality of machines can distribute the existing load, the hydroelectric generating set can be timely regulated when a wind-solar fluctuation power supply changes, the fluctuation of a power grid is reduced, the power generation quality is improved, meanwhile, the pre-regulation capacity can be used for reducing the frequency of the hydroelectric generating set frequently passing through a vibration area, the forced wind discarding and the light discarding of the generating set for avoiding the operation vibration area are avoided, and the safe and stable operation of the generating set and the power grid is ensured while the power requirement of a power station is met.
Fig. 3 illustrates an embodiment of the electronic device of the application, see fig. 3, the electronic device 7 comprising a processor 71 and a memory 72 coupled to the processor 71.
The memory 72 stores program instructions for implementing the method for predicting power of a wind-solar-water hybrid power plant according to any of the embodiments described above.
The processor 71 is configured to execute program instructions stored in the memory 72 to perform power prediction of the wind, solar and water hybrid power station.
The processor 71 may also be referred to as a CPU (Central Processing Unit ). The processor 71 may be an integrated circuit chip with signal processing capabilities. Processor 71 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Further, fig. 4 is a schematic structural diagram of a storage medium according to an embodiment of the present application, referring to fig. 4, where the storage medium 8 according to an embodiment of the present application stores a program instruction 81 capable of implementing all the methods described above, where the program instruction 81 may be stored in the storage medium in the form of a software product, and includes several instructions for making a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) execute all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, an optical disk, or other various media capable of storing program codes, or a terminal device such as a computer, a server, a mobile phone, a tablet, or the like.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of elements is merely a logical functional division, and there may be additional divisions of actual implementation, e.g., multiple elements or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other forms.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units. The foregoing is only the embodiments of the present application, and the patent scope of the application is not limited thereto, but is also covered by the patent protection scope of the application, as long as the equivalent structure or equivalent flow changes made by the description and the drawings of the application or the direct or indirect application in other related technical fields are adopted.
The embodiments of the application have been described in detail above, but they are merely examples, and the application is not limited to the above-described embodiments. It will be apparent to those skilled in the art that any equivalent modifications or substitutions to this application are within the scope of the application, and therefore, all equivalent changes and modifications, improvements, etc. that do not depart from the spirit and scope of the principles of the application are intended to be covered by this application.
Claims (8)
1. A power prediction method for a wind-solar-water complementary power station, the wind-solar-water complementary power station comprising a meteorological station and a power station built in an energy base area, the power station comprising a wind power station, a photovoltaic power station and a hydropower station which are operated in a grid-connected mode, the hydropower station comprising a conventional hydropower station and a pumped storage power station, the power prediction method comprising:
acquiring historical wind and light data of the wind power station and the photoelectric station within a first preset duration through the meteorological station;
defining a training model of the wind-solar data through a first preset algorithm;
training the training model by using a neural network based on a preset data set to obtain a wind-light data model based on the first preset duration;
acquiring a real-time stamp of the energy base region, and acquiring model data with a second preset duration in the wind-solar data model by taking the real-time stamp as a measurement starting point as measurement reference data, wherein the preset duration can reach a minute level;
acquiring the trend of the wind and light trend of the forecast reference data;
and calculating predicted power according to the trend of the wind and light, and outputting a result to a computer monitoring system of the hydropower station.
2. The power prediction method according to claim 1, wherein defining the training model of the wind-solar data by a first preset algorithm comprises:
all characteristics of the wind and light data are acquired and respectively defined as an input node;
each input node is respectively endowed with a preset weight;
defining the training model according to equation (1):
wherein y is the training model; x is x n For the n-th input node,the preset weight values from the mth input node of the input layer to the nth input node of the hidden layer are obtained; />A threshold value for the n-th input node of the hidden layer; tan sig (x) is the transfer function, and +.>
3. The power prediction method according to claim 1, wherein the training model is trained by a neural network based on a preset data set to obtain a wind-solar data model based on the first preset duration, and the method comprises:
training the training model by using a neural network based on a preset data set;
iteratively updating all preset weights of the training model by a back propagation algorithm with preset times so that the value of a loss function of the training model is smaller than or equal to a second preset threshold value;
and acquiring a training model after training and defining the training model as the wind-light data model.
4. The power prediction method according to claim 1, wherein iteratively updating all preset weights of the training model by a back propagation algorithm for a preset number of times so that a value of a loss function of the training model is equal to or less than a second preset threshold value comprises:
acquiring an error signal after iterative updating of the training model;
back-propagating the error signal to an output layer of the training model to obtain an error value (2) of the output layer;
wherein E is the error value; target is the expected output of the training model; output is the actual output of the training model;
sequentially updating preset weights of the output layer, the hidden layer and the input layer according to the error value;
and performing the next iteration on the training model based on the updated preset weight until the value of the loss function of the training model is smaller than or equal to the second preset threshold.
5. The power prediction method according to claim 1, wherein obtaining the wind-solar trend of the prediction reference data comprises:
establishing a rectangular coordinate system by taking natural time as an x axis and taking the numerical value of the model data as a y axis;
outputting the measurement and report reference data in the rectangular coordinate system to form a data curve;
outputting the real-time stamp in the rectangular coordinate system and acquiring a curve slope of the second preset duration of the data curve based on the real-time stamp;
and obtaining a change curve of the slope of the curve in the second preset time period, wherein the change curve is the trend of the wind and light trend.
6. A power prediction apparatus of a wind-solar-water complementary power station, the power prediction apparatus of the wind-solar-water complementary power station being applied to the power prediction method of the wind-solar-water complementary power station according to any one of claims 1 to 5, characterized in that the power prediction apparatus of the wind-solar-water complementary power station comprises:
the historical wind-light data acquisition module is used for acquiring historical wind-light data of the wind power station and the photoelectric station in a first preset duration through the meteorological station;
the training model definition module is used for defining a training model of the wind-solar data through a first preset algorithm;
the wind-solar data model training module is used for training the training model by using a neural network based on a preset data set to obtain a wind-solar data model based on the first preset duration;
the measurement and report reference data acquisition module is used for acquiring a real-time stamp of the energy base area, and acquiring model data with a second preset duration in the wind-solar data model by taking the real-time stamp as a measurement and report starting point as measurement and report reference data;
the wind-light trend acquisition module is used for acquiring the wind-light trend of the observation and report reference data;
and the predicted power output module is used for calculating predicted power according to the trend of the wind and light trend, and outputting the result to a computer monitoring system of the hydropower station.
7. An electronic device comprising a processor, and a memory coupled to the processor, the memory storing program instructions executable by the processor; the processor, when executing the program instructions stored in the memory, implements a method for predicting power of a wind-solar-water complementary power station according to any one of claims 1 to 5.
8. A storage medium having stored therein program instructions which, when executed by a processor, implement a power prediction method capable of implementing a wind-solar-water hybrid power station according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896582.5A CN117039852A (en) | 2023-07-20 | 2023-07-20 | Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310896582.5A CN117039852A (en) | 2023-07-20 | 2023-07-20 | Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117039852A true CN117039852A (en) | 2023-11-10 |
Family
ID=88627147
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310896582.5A Pending CN117039852A (en) | 2023-07-20 | 2023-07-20 | Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117039852A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239738A (en) * | 2023-11-10 | 2023-12-15 | 华南理工大学 | Wind power prediction method and system based on combined prediction model |
-
2023
- 2023-07-20 CN CN202310896582.5A patent/CN117039852A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117239738A (en) * | 2023-11-10 | 2023-12-15 | 华南理工大学 | Wind power prediction method and system based on combined prediction model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021238505A1 (en) | Federated learning-based regional photovoltaic power probability prediction method, and cooperative regulation and control system | |
| CN113128793A (en) | Photovoltaic power combination prediction method and system based on multi-source data fusion | |
| CN103390116B (en) | Use the photovoltaic power station power generation power forecasting method of stepping mode | |
| CN106446494B (en) | Honourable power forecasting method based on wavelet packet-neural network | |
| CN109284870A (en) | Short-term method for forecasting photovoltaic power generation quantity based on shot and long term Memory Neural Networks | |
| CN106950902A (en) | A kind of electric integrated automatic monitoring system | |
| Guo et al. | An ensemble solar power output forecasting model through statistical learning of historical weather dataset | |
| CN106485093B (en) | Based on the solar irradiance time series synthetic method for improving Markov chain | |
| CN112510703B (en) | Multi-energy access power grid optimal scheduling method considering coal consumption curve correction | |
| CN112329977A (en) | Wind power prediction system for extreme scene | |
| CN111612244B (en) | QRA-LSTM-based method for predicting nonparametric probability of photovoltaic power before day | |
| CN102545707A (en) | Power generation power forecasting method and system by taking power generation units as basic prediction units | |
| Li et al. | Day-ahead hourly photovoltaic generation forecasting using extreme learning machine | |
| CN110263915A (en) | A kind of wind power forecasting method based on deepness belief network | |
| CN117039852A (en) | Power prediction method, device, equipment and storage medium of wind-solar-water complementary power station | |
| Shah et al. | A novel prediction error-based power forecasting scheme for real pv system using pvusa model: A grey box-based neural network approach | |
| Chen et al. | Research on wind power prediction method based on convolutional neural network and genetic algorithm | |
| Tao et al. | On comparing six optimization algorithms for network-based wind speed forecasting | |
| CN105844350A (en) | Short period wind power prediction system based on covariance preferable combination model | |
| Sahoo et al. | Energy forecasting for grid connected MW range solar PV system | |
| Khoury et al. | A predictive convolutional neural network model for source-load forecasting in smart grids | |
| TW201740296A (en) | Method and system for predicting power generation capacity of renewable energy using a neural network to accurately calculate the power generation capacity of renewable energy | |
| Jun-Ma et al. | Photovoltaic power generation prediction based on MEA-BP neural network | |
| CN112491049B (en) | Multi-energy access power grid optimized scheduling method considering on-line coal consumption curve | |
| CN114861993A (en) | Regional photovoltaic power generation prediction method based on federal learning and deep neural network |
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 |