WO2024046137A1 - 一种多能源联合发电系统的功率预测模型构建及预测方法 - Google Patents

一种多能源联合发电系统的功率预测模型构建及预测方法 Download PDF

Info

Publication number
WO2024046137A1
WO2024046137A1 PCT/CN2023/113555 CN2023113555W WO2024046137A1 WO 2024046137 A1 WO2024046137 A1 WO 2024046137A1 CN 2023113555 W CN2023113555 W CN 2023113555W WO 2024046137 A1 WO2024046137 A1 WO 2024046137A1
Authority
WO
WIPO (PCT)
Prior art keywords
power generation
data
historical
power
generation system
Prior art date
Application number
PCT/CN2023/113555
Other languages
English (en)
French (fr)
Inventor
张玮
李梦杰
张璐
黄康迪
刘瑞阔
刘志武
Original Assignee
中国长江三峡集团有限公司
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 中国长江三峡集团有限公司 filed Critical 中国长江三峡集团有限公司
Publication of WO2024046137A1 publication Critical patent/WO2024046137A1/zh

Links

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
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/381—Dispersed generators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • 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
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H02J3/466—Scheduling the operation of the generators, e.g. connecting or disconnecting generators to meet a given demand
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
    • 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
    • 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
    • H02J2300/40—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation wherein a plurality of decentralised, dispersed or local energy generation technologies are operated simultaneously
    • 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

Definitions

  • the invention relates to the field of electric power energy technology, and in particular to a power prediction model construction and prediction method for a multi-energy joint power generation system.
  • Vigorously developing power generation technologies from renewable energy sources such as water energy, wind energy, and solar energy can effectively solve the environmental pollution problems caused by conventional fossil energy power generation. Judging from the current resource status and technological development level of renewable energy, it is an effective way to aggregate hydro energy, wind energy, solar energy and other renewable energy sources to form a multi-energy complementary power generation system.
  • the randomness and volatility of photovoltaic power generation and wind power generation are inherent flaws of these two renewable energy sources.
  • Large-scale photovoltaic and wind power grid integration will inevitably pose a threat to the safety and stable operation of the power generation system. Therefore, improving the accuracy of combined power generation prediction of multiple renewable energy sources is of great significance to the safe and stable operation of the power generation system, improving power quality, and improving the effective utilization of renewable energy sources.
  • most of the existing renewable energy power generation forecasts rely on artificial intelligence algorithms such as deep learning to analyze a single energy source, and have not yet carried out joint power forecasts based on multiple renewable energy sources.
  • embodiments of the present invention provide a power prediction model construction and prediction method related to a multi-energy joint power generation system to achieve joint prediction of multiple renewable energy sources.
  • the first aspect of the embodiment of the present invention provides a method for constructing a power prediction model of a multi-energy joint power generation system.
  • the method for constructing a power prediction model of a multi-energy joint power generation system includes: obtaining the history of each power generation mode in the multi-energy joint power generation system. Power generation data and meteorological factors corresponding to historical power generation data, the meteorological factors representing meteorological factors affecting power generation; input historical power generation data of each power generation method and meteorological factors corresponding to historical power generation data
  • the preset network model is such that the preset network model calculates the power generation correlation between any two power generation modes among the multiple power generation modes according to the preset correlation calculation method; The correlation between power generation and the Nash efficiency coefficient corresponding to each power generation mode is used to construct a loss function.
  • the historical power generation data of each power generation mode and the meteorological factors corresponding to the historical power generation data are used as training data to predict the prediction.
  • the network model is trained until the preset training conditions are met and the corresponding power prediction model of the multi-energy combined power generation system is obtained.
  • the method before inputting the historical power generation data of each power generation mode and the meteorological factors corresponding to the historical power generation data into the preset network model, the method further includes: analyzing the acquired historical power generation data. Carry out the identification operation of abnormal data and/or missing data; identify abnormal data of historical power generation and missing historical power generation. Data is handled with exceptions.
  • performing a missing data identification operation on the acquired historical power generation data includes: when the time interval corresponding to any two adjacent acquired historical power generation data is greater than a preset time length, determining that the Two adjacent historical power generation data are missing data of historical power generation.
  • perform exception processing on the identified historical power generation abnormal data and historical power generation missing data including: deleting the historical power generation missing data in the acquired historical power generation data and using a preset supervised learning method to The historical power generation abnormal data is processed.
  • the multi-energy combined power generation system includes a hydropower power generation system.
  • obtaining historical power generation data of each power generation mode in the multi-energy combined power generation system includes: obtaining historical data of hydropower stations and calculating corresponding historical hydropower power generation data based on the historical data of hydropower stations.
  • a second aspect of the embodiment of the present invention provides a power prediction method for a multi-energy joint power generation system.
  • the power prediction method for the multi-energy joint power generation system includes: obtaining power generation data of each power generation mode in the multi-energy joint power generation system to be predicted. Corresponding meteorological factors; input the meteorological factors corresponding to the power generation data of each power generation method into the power prediction model construction of the multi-energy combined power generation system as described in the first aspect of the embodiment of the present invention and any one of the first aspects.
  • the method constructs a power prediction model of the obtained multi-energy combined power generation system, and obtains the power of the multi-energy combined power generation system to be predicted.
  • a third aspect of the embodiment of the present invention provides a device for constructing a power prediction model of a multi-energy joint power generation system.
  • the device for constructing a power prediction model of a multi-energy joint power generation system includes: a first acquisition module for obtaining the power of a multi-energy joint power generation system.
  • the meteorological factors represent the meteorological factors that affect the power generation power; the first input module is used to convert the historical power generation data of each power generation method.
  • the power generation data and the meteorological factors corresponding to the historical power generation data are input into the preset network model, so that the preset network model calculates the power generation between any two power generation modes among the multiple power generation modes according to the preset correlation calculation method.
  • Data and meteorological factors corresponding to historical power generation data are used as training data, and the preset network model is trained until the preset training conditions are met and the corresponding power prediction model of the multi-energy combined power generation system is obtained.
  • the fourth aspect of the embodiment of the present invention provides a power prediction device for a multi-energy joint power generation system.
  • the power prediction device for the multi-energy joint power generation system includes: a second acquisition module for obtaining each of the multi-energy joint power generation systems to be predicted. Meteorological factors corresponding to the power generation data of each power generation mode; a second input module, used to input the meteorological factors corresponding to the power generation data of each power generation mode into any one of the first aspect and the first aspect of the embodiment of the present invention.
  • the power prediction model of the multi-energy combined power generation system constructed by the power prediction model construction method of the multi-energy combined power generation system described in the item is obtained.
  • a fifth aspect of the embodiment of the present invention provides a computer-readable storage medium.
  • the computer-readable storage medium stores computer instructions.
  • the computer instructions are used to cause the computer to execute the first aspect and the first embodiment of the present invention.
  • a sixth aspect of the embodiment of the present invention provides an electronic device, including: a memory and a processor.
  • the memory and the processor are connected to each other for communication.
  • the memory stores computer instructions.
  • the processor executes the Computer instructions to execute the power prediction model construction method of a multi-energy combined power generation system as described in the first aspect and any one of the first aspects of the embodiments of the present invention, or as described in any of the second aspect and the second aspect of the embodiments of the present invention.
  • the power prediction model construction method of the multi-energy joint power generation system constructs a loss function composed of the Nash efficiency coefficient and the correlation between the power generation, taking into account the need to improve the accuracy of model prediction and the power of the multi-energy joint power generation system.
  • the correlation of predictions provides basic and accurate data support for the preparation of multi-energy complementary dispatch plans.
  • the power prediction method of the multi-energy joint power generation system uses the trained power prediction model of the multi-energy joint power generation system to perform prediction, and realizes the synchronization and joint prediction of the power of the multi-energy joint power generation system.
  • Figure 1 is a flow chart of a method for building a power prediction model of a multi-energy combined power generation system according to an embodiment of the present invention
  • Figure 2 is a schematic diagram of a box plot provided according to an embodiment of the present invention.
  • Figure 3 is a flow chart of a power prediction method of a multi-energy joint power generation system according to an embodiment of the present invention
  • Figure 4 is a structural block diagram of a device for building a power prediction model of a multi-energy combined power generation system according to an embodiment of the present invention
  • Figure 5 is a structural block diagram of a power prediction device of a multi-energy combined power generation system according to an embodiment of the present invention
  • Figure 6 is a schematic structural diagram of a computer-readable storage medium provided according to an embodiment of the present invention.
  • FIG. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
  • An embodiment of the present invention provides a method for building a power prediction model for a multi-energy combined power generation system. As shown in Figure 1, the method includes the following steps:
  • Step S101 Obtain the historical power generation data of each power generation mode in the multi-energy combined power generation system and the corresponding meteorological factors of the historical power generation data.
  • the meteorological factors represent the meteorological factors that affect the power generation.
  • a multi-energy combined power generation system refers to a power generation system that utilizes the complementarity between various energy sources.
  • the meteorological factors corresponding to the historical power generation data of each power generation method in the system can include precipitation, evaporation, runoff, air pressure, wind speed, wind direction, and direct radiation. , scattered radiation, temperature and other meteorological elements that affect power generation power.
  • the embodiment of the present application does not limit the type of meteorological factors. Those skilled in the art can select meteorological factors that can affect the power generation according to actual needs.
  • Step S102 Input the historical power generation data of each power generation method and the corresponding meteorological factors of the historical power generation data into the preset network model, so that the preset network model calculates various types of power generation data according to the preset correlation calculation method.
  • the preset network model can be a neural network model such as a long short-term memory neural network (LSTM) model.
  • LSTM long short-term memory neural network
  • the preset network model performs correlation calculation through the following formula:
  • r represents the correlation of power generation between two different power generation methods (y 1 , y 2 ); n represents the total number of samples of historical power generation data for each power generation method; y 1(i) , y 2( i) respectively represent the power data of the i-th sample in the historical power generation data of two different power generation methods; Represents the average value of historical power generation data of two different power generation methods.
  • the preset network model learns the correlation calculation method, when the historical power generation data of multiple power generation methods and the meteorological factors corresponding to the historical power generation data are input into the model, the preset network model can calculate multiple power generation methods.
  • Step S103 Construct a loss function based on the power generation correlation between any two power generation modes and the Nash efficiency coefficient corresponding to each power generation mode, and use the historical power generation data of each power generation mode and the corresponding historical The meteorological factors of the power generation data are used as training data, and the preset network model is trained until the preset training conditions are met. software and obtain the corresponding power prediction model of the multi-energy combined power generation system. The Nash efficiency coefficient is used to verify the quality of model simulation results.
  • NSE Nash-Sutcliffe efficiency coefficient
  • T represents the total number of samples of historical power generation data of each power generation method used for model training, such as the total number of samples of 70% of the historical power generation data of each power generation method
  • y 1( t) represents the power data of the t-th sample in the historical power generation data of each power generation method used for model training
  • through the actually obtained power generation data y 1(t) and the predicted power data Comparing calculations to calculate the Nash efficiency coefficient can verify the quality of the model prediction results.
  • loss NSE represents the loss function corresponding to the Nash efficiency coefficient
  • loss r represents the loss function corresponding to the power generation correlation between any two power generation methods
  • NSE 1 , NSE 2 , and NSE 3 respectively represent the Nash efficiency coefficients corresponding to different power generation methods.
  • r 12 , r 13 , and r 23 respectively represent any two power generation methods input by the model.
  • the correlation between power generation is calculated from the historical power generation.
  • r 12 is the power generation correlation between power generation mode 1 and power generation mode 2
  • r 13 is the power generation power correlation between power generation mode 1 and power generation mode 3
  • r 23 is the power generation power correlation between power generation mode 2 and power generation mode 3.
  • Generation power correlation Represents the correlation of power generation between any two power generation modes corresponding to the model output, calculated from the predicted power generation; specifically, r 12 , r 13 , r 23 are all calculated with reference to formula (1);
  • ⁇ 12 , ⁇ 13 , and ⁇ 23 respectively represent The corresponding penalty parameter has the following values:
  • ⁇ r represents the relative difference threshold of the correlation between historical power generation and predicted power generation.
  • the value range is [0.1, 0.3]. The specific value needs to be further determined according to the preferences of the decision maker;
  • Use the constructed loss function to train the model use the historical power generation data of each power generation method and the meteorological factors corresponding to the historical power generation data as training data to train the preset network model until the loss function takes the minimum value (L min ), the training ends and the corresponding power prediction model of the multi-energy combined power generation system is obtained.
  • the default network models are three different LSTM models, and the corresponding loss functions are all formula 3 The loss function described above; after inputting the historical power generation data of the three power generation methods into the corresponding LSTM models for training, the power prediction model ⁇ LSTM 1 , LSTM 2 , LSTM 3 ⁇ of the multi-energy combined power generation system is obtained.
  • the power prediction model of the multi-energy combined power generation system is composed of three well-trained power generation power prediction models LSTM 1 , LSTM 2 , and LSTM 3 , which are used to predict power data corresponding to different power generation methods.
  • the parameters of the power prediction model of the multi-energy combined power generation system can also be optimized.
  • the hyperparameters corresponding to the power prediction model of the multi-energy joint power generation system may include: the number of memory units ⁇ MC 1 , MC 2 , MC 3 ⁇ , number of network layers ⁇ La 1 , La 2 , La 3 ⁇ , learning rate ⁇ R 1 , R 2 , R 3 ⁇ , batch size ⁇ B 1 , B 2 , B 3 ⁇ , time expansion The number of steps ⁇ ts 1 , ts 2 , ts 3 ⁇ , and the choice of gradient descent algorithm ⁇ G 1 , G 2 , G 3 ⁇ .
  • the training set (70% of the historical power generation data in the historical power generation data) is used to train the corresponding power generation power prediction models LSTM 1 , LSTM 2 , and LSTM for three different power generation methods. 3. Use the remaining 30% of the historical power generation data as the verification set to calculate the value of the corresponding objective function at this time.
  • the power prediction model construction method of the multi-energy joint power generation system constructs a loss function composed of the Nash efficiency coefficient and the correlation between the power generation, taking into account the need to improve the accuracy of model prediction and the power of the multi-energy joint power generation system.
  • the correlation of predictions provides basic and accurate data support for the preparation of multi-energy complementary dispatch plans.
  • the method before step S102, further includes: performing an identification operation on abnormal data and/or missing data on the acquired historical power generation data; Exception processing is performed on abnormal power generation data and missing data on historical power generation.
  • box plots to identify abnormal data means treating data that is greater than or less than the upper bound (UB) and lower bound (LB) set by the box plot as abnormal data.
  • the box plot is shown in Figure 2.
  • U is the upper quartile, indicating that only 1/4 of the historical power generation data corresponding to a certain type of power generation method is greater than U;
  • L is the lower quartile, indicating the historical power generation data corresponding to a certain type of power generation method. Only 1/4 of the generated power data is smaller than U.
  • a missing data identification operation on the acquired historical power generation data, including: when the time interval corresponding to any two adjacent historical power generation data acquired is greater than a preset time length, determine that the two adjacent The historical power generation data of is missing data of historical power generation. For example, data that is missing for more than 16 consecutive moments in the historical power generation data is regarded as missing data of historical power generation.
  • abnormality processing is performed on the identified historical power generation abnormal data and historical power generation missing data, including: deleting the historical power generation missing data in the acquired historical power generation data and using a preset supervised learning method to Process historical power generation abnormal data.
  • the preset supervised learning method may include K nearest neighbor complementary method, naive Bayes, decision tree, EM algorithm, etc.
  • the present invention does not specifically limit this, as long as it meets the requirements.
  • the historical power generation missing data is deleted from the historical power generation data.
  • the calculation formula for processing the historical power generation abnormal data using the K nearest neighbor complementary method is:
  • x j indicates that the j-th sample data in the historical power generation data corresponding to a certain type of power generation method is abnormal data x j ; x jk indicates the k-th data before x j ; x j+k indicates the j-th data after x j k data; the k value is generally 2-5, depending on actual needs OK.
  • the historical power generation data of each power generation method in the multi-energy joint power generation system is obtained. , including: obtaining historical data of the hydropower station and calculating corresponding historical hydropower power data based on the historical data of the hydropower station.
  • the historical data of the hydropower station include but are not limited to the inflow flow, water level-storage capacity relationship curve, discharge flow-tail water level relationship curve, upper and lower limits of hydropower station reservoir water level, maximum power generation flow value of hydropower station reservoir, minimum allowable discharge flow, hydropower station output coefficient, Hydropower station installed capacity, hydropower station guaranteed output, hydropower station dispatching procedures, etc.
  • P max represents the installed capacity
  • eta represents the output coefficient
  • QE t represents the power generation flow of the hydropower station, which is the smaller value of the maximum allowable power generation flow QE max and the outflow flow Q t
  • ⁇ Z t represents the power generation head difference
  • V t+1 V t +(I t -Q t ) ⁇ turban..(17)
  • An embodiment of the present invention provides a power prediction method for a multi-energy combined power generation system, as shown in Figure 3.
  • the method includes the following steps:
  • Step S201 Obtain the meteorological factors corresponding to the power generation data of each power generation mode in the multi-energy joint power generation system to be predicted.
  • Step S202 Input the meteorological factors corresponding to the power generation data of each power generation method into the power prediction of the multi-energy combined power generation system constructed by the power prediction model construction method of the multi-energy combined power generation system according to the embodiment of the present invention. model to obtain the power of the multi-energy combined power generation system to be predicted. Specifically, the trained power prediction model of the multi-energy combined power generation system can be used to obtain the power of the multi-energy combined power generation system to be predicted.
  • the water, wind, and light meteorological factors of the day to be predicted are input into the power prediction of the multi-energy combined power generation system.
  • the output of the model ⁇ P 1 , P 2 , P 3 ⁇ is the final power of the water, wind and photovoltaic combined power generation system. forecast result.
  • the power prediction method of the multi-energy joint power generation system uses the trained power prediction model of the multi-energy joint power generation system to perform prediction, and realizes the synchronization and joint prediction of the power of the multi-energy joint power generation system.
  • An embodiment of the present invention also provides a device for building a power prediction model of a multi-energy combined power generation system. As shown in Figure 4, the device includes:
  • the first acquisition module 401 is used to acquire the historical power generation data of each power generation method in the multi-energy combined power generation system and the meteorological factors corresponding to the historical power generation data.
  • the meteorological factors represent the meteorological factors that affect the power generation; details Please refer to the relevant description of step S101 in the above method embodiment.
  • the first input module 402 is used to input the historical power generation data of each power generation method and the corresponding meteorological factors of the historical power generation data into the preset network model, so that the preset network model calculates according to the preset correlation
  • the method calculates the power generation power correlation between any two power generation modes among multiple power generation modes; for details, please refer to the relevant description of step S102 in the above method embodiment.
  • the training module 403 is used to construct a loss function based on the power generation correlation between any two power generation methods and the Nash efficiency coefficient corresponding to each power generation method, and use the historical power generation data of each power generation method and Meteorological factors corresponding to historical power generation data are used as training data, and the preset network model is trained until the preset training conditions are met and the corresponding power prediction model of the multi-energy combined power generation system is obtained; for details, see the above method embodiments Relevant description of step S103.
  • the power prediction model construction device of the multi-energy joint power generation system constructs a loss function composed of the Nash efficiency coefficient and the correlation between the power generation, taking into account the need to improve the accuracy of model prediction and the power of the multi-energy joint power generation system.
  • the correlation of predictions provides basic and accurate data support for the preparation of multi-energy complementary dispatch plans.
  • the device further includes: a first identification module, configured to identify abnormal data and/or missing data on the acquired historical power generation data; a first The processing module is used for abnormal processing of the identified historical power generation abnormal data and historical power generation missing data.
  • the first identification module includes: a first determination sub-module, used when the time interval corresponding to any two adjacent historical power generation data obtained is greater than a predetermined time interval. Assuming the time length, it is determined that the two adjacent historical power generation data are missing data of historical power generation.
  • the first processing module includes: a first processing sub-module, configured to delete the missing data of historical power generation in the acquired historical power generation data and use the preset The supervised learning method processes the historical power generation abnormal data.
  • the multi-energy combined power generation system includes a hydropower generation system.
  • the first acquisition module includes: a first calculation sub-module, used to acquire historical data of hydropower stations and calculate corresponding historical hydropower power data based on the historical data of hydropower stations.
  • An embodiment of the present invention also provides a power prediction device for a multi-energy combined power generation system. As shown in Figure 5, the device includes:
  • the second acquisition module 501 is used to acquire the meteorological factors corresponding to the power generation data of each power generation mode in the multi-energy joint power generation system to be predicted; for details, please refer to the relevant description of step S201 in the above method embodiment.
  • the second input module 502 is used to input the meteorological factors corresponding to the power generation data of each power generation mode into the multi-energy combined power generation system constructed by the power prediction model construction method of the multi-energy combined power generation system according to the embodiment of the present invention.
  • the power prediction model of the power generation system is used to obtain the power of the multi-energy combined power generation system to be predicted; for details, please refer to the relevant description of step S202 in the above method embodiment.
  • the power prediction device of the multi-energy joint power generation system uses the trained power prediction model of the multi-energy joint power generation system to perform prediction, thereby realizing the synchronization and joint prediction of the power of the multi-energy joint power generation system.
  • An embodiment of the present invention also provides a storage medium, as shown in Figure 6, on which a computer program 601 is stored.
  • the storage media can be magnetic disks, optical disks, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), flash memory (Flash Memory), hard disk (Hard Disk Drive). , abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the above types of memories.
  • the program can be stored in a computer-readable storage medium.
  • the program can be stored in a computer-readable storage medium.
  • the process may include the processes of the embodiments of each of the above methods.
  • the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (Hard disk). Disk Drive (abbreviation: HDD) or solid-state drive (Solid-State Drive, SSD), etc.; the storage medium may also include a combination of the above types of memories.
  • an embodiment of the present invention also provides an electronic device.
  • the electronic device may include a processor 71 and Memory 72, where the processor 71 and the memory 72 can be connected through a bus or other means.
  • the connection through a bus is taken as an example.
  • the processor 71 may be a central processing unit (Central Processing Unit, CPU).
  • the processor 71 can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or Other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components and other chips, or combinations of the above types of chips.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field-Programmable Gate Array
  • Other programmable logic devices discrete gate or transistor logic devices, discrete hardware components and other chips, or combinations of the above types of chips.
  • the memory 72 can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as corresponding program instructions/modules in embodiments of the present invention.
  • the processor 71 executes the non-transient software programs, instructions and modules stored in the memory 72 to execute various functional applications and data processing of the processor, that is, to implement the power prediction of the multi-energy combined power generation system in the above method embodiment. Model building methods or power prediction methods for multi-energy combined power generation systems.
  • the memory 72 may include a program storage area and a data storage area, where the program storage area may store an operating device and an application program required for at least one function; the storage data area may store data created by the processor 71 and the like.
  • memory 72 may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device.
  • memory 72 optionally includes memory located remotely relative to processor 71 , and these remote memories may be connected to processor 71 through a network. Examples of the above-mentioned networks include but are not limited to the Internet, intranets, local area networks, mobile communication networks and combinations thereof.
  • the one or more modules are stored in the memory 72, and when executed by the processor 71, execute the power prediction model building method or multiple methods of the multi-energy combined power generation system in the embodiment shown in Figures 1-3. Power prediction method for energy cogeneration systems.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Power Engineering (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Educational Administration (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

提供一种多能源联合发电系统的功率预测模型构建及预测方法，获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应的气象因子并输入预设网络模型计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；以该相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型，兼顾了模型预测的精度提高需求和多能源联合发电系统的功率预测的相关性，为多能源互补调度计划编制提供了基础性、精准的数据支撑。

Description

一种多能源联合发电系统的功率预测模型构建及预测方法 技术领域
本发明涉及电力能源技术领域，具体涉及一种多能源联合发电系统的功率预测模型构建及预测方法。
背景技术
大力发展水能、风能、光能等可再生能源的发电技术可有效解决由于常规化石能源发电过程中引起的环境污染问题。从当前可再生能源的资源状况和技术发展水平看，将水能、风能、太阳能等多种可再生能源进行聚合，形成多能互补发电系统是一条有效的途径。但光伏发电和风力发电的随机性和波动性是这两种可再生能源固有的缺陷，大规模的光伏和风电并网势必会对发电系统的安全和稳定运行造成威胁。因此，提高多种可再生能源的联合发电预测的精度，对发电系统的安全稳定运行、提高电能质量和提高可再生能源的有效利用意义重大。但现有的可再生能源发电功率预测大都借助深度学习等人工智能算法、针对单一能源进行分析，尚未基于多种可再生能源开展功率联合预测。
发明内容
有鉴于此，本发明实施例提供了涉及一种多能源联合发电系统的功率预测模型构建及预测方法，以实现对多种可再生能源的联合预测。
本发明提出的技术方案如下：
本发明实施例第一方面提供一种多能源联合发电系统的功率预测模型构建方法，该多能源联合发电系统的功率预测模型构建方法包括：获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素；将所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型。
可选地，将所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子输入预设网络模型之前，所述方法还包括：对获取到的所述历史发电功率数据进行异常数据和/或缺失数据的识别操作；对识别到的历史发电功率异常数据和历史发电功率缺失 数据进行异常处理。
可选地，对获取到的所述历史发电功率数据进行缺失数据的识别操作，包括：当获取的任意两个相邻的所述历史发电功率数据对应的时间间隔大于预设时长，判定所述两个相邻的历史发电功率数据为历史发电功率缺失数据。
可选地，对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理，包括：在获取到的历史发电功率数据中删除所述历史发电功率缺失数据并利用预设监督学习方法对所述历史发电功率异常数据进行处理。
可选地，所述多能源联合发电系统包括水能发电系统。
可选地，获取多能源联合发电系统中每一种发电方式的历史发电功率数据，包括：获取水电站历史数据并根据所述水电站历史数据计算对应的水能历史发电功率数据。
本发明实施例第二方面提供一种多能源联合发电系统的功率预测方法，该多能源联合发电系统的功率预测方法包括：获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子；将所述每一种发电方式的发电功率数据对应的气象因子输入如本发明实施例第一方面及第一方面任一项所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到所述待预测多能源联合发电系统的功率。
本发明实施例第三方面提供一种多能源联合发电系统的功率预测模型构建装置，该多能源联合发电系统的功率预测模型构建装置包括：第一获取模块，用于获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素；第一输入模块，用于将所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；训练模块，用于以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型。
本发明实施例第四方面提供一种多能源联合发电系统的功率预测装置，该多能源联合发电系统的功率预测装置包括：第二获取模块，用于获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子；第二输入模块，用于将所述每一种发电方式的发电功率数据对应的气象因子输入如本发明实施例第一方面及第一方面任一项所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到 所述待预测多能源联合发电系统的功率。
本发明实施例第五方面提供一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，所述计算机指令用于使所述计算机执行如本发明实施例第一方面及第一方面任一项所述的多能源联合发电系统的功率预测模型构建方法，或者如本发明实施例第二方面及第二方面任一项所述的多能源联合发电系统的功率预测方法。
本发明实施例第六方面提供一种电子设备，包括：存储器和处理器，所述存储器和所述处理器之间互相通信连接，所述存储器存储有计算机指令，所述处理器通过执行所述计算机指令，从而执行如本发明实施例第一方面及第一方面任一项所述的多能源联合发电系统的功率预测模型构建方法，或者如本发明实施例第二方面及第二方面任一项所述的多能源联合发电系统的功率预测方法。
本发明提供的技术方案，具有如下效果：
本发明实施例提供的多能源联合发电系统的功率预测模型构建方法，构建了由纳什效率系数与发电功率相关性组成的损失函数，兼顾了模型预测的精度提高需求和多能源联合发电系统的功率预测的相关性，为多能源互补调度计划编制提供了基础性、精准的数据支撑。
本发明实施例提供的多能源联合发电系统的功率预测方法，利用训练好的多能源联合发电系统的功率预测模型进行预测，实现了多能源联合发电系统功率的同步、联合预测。
附图说明
为了更清楚地说明本发明具体实施方式或现有技术中的技术方案，下面将对具体实施方式或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本发明的一些实施方式，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是根据本发明实施例的多能源联合发电系统的功率预测模型构建方法的流程图；
图2是根据本发明实施例提供的箱线图的示意图；
图3是根据本发明实施例的多能源联合发电系统的功率预测方法的流程图；
图4是根据本发明实施例的多能源联合发电系统的功率预测模型构建装置的结构框图；
图5是根据本发明实施例的多能源联合发电系统的功率预测装置的结构框图；
图6是根据本发明实施例提供的计算机可读存储介质的结构示意图；
图7是根据本发明实施例提供的电子设备的结构示意图。
具体实施方式
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附 图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明实施例提供一种多能源联合发电系统的功率预测模型构建方法，如图1所示，该方法包括如下步骤：
步骤S101：获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应的历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素。具体地，多能源联合发电系统表示利用各种能源之间的互补性所组成的发电系统。以多能源联合发电系统为由水能、风能、太阳能为例，则系统每一种发电方式的历史发电功率数据对应的气象因子可以包括降水量、蒸发、径流、气压、风速、风向、直射辐射、散射辐射、气温等影响发电功率的气象元素。本申请实施例对该气象因子的类型不作限定，本领域技术人员可以根据实际需要选择可以影响发电功率的气象因素。
步骤S102：将所述每一种发电方式的历史发电功率数据以及对应的历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性。具体地，预设网络模型可以为长短期记忆神经网络(LSTM)模型等神经网络模型，本发明对此不做具体限定，只要满足需求即可。
本申请实施例中预设网络模型通过下式进行相关性计算：
式中，r表示两种不同发电方式(y1、y2)之间的发电功率相关性；n表示每一种发电方式的历史发电功率数据的样本总数；y1(i)、y2(i)分别表示两种不同发电方式的历史发电功率数据中第i个样本的功率数据；分别表示两种不同发电方式的历史发电功率数据的平均值。
该预设网络模型学习该相关性计算方法后，当该模型中输入多种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子后，该预设网络模型可以计算得到多种发电方式中任意两种发电方式之间的发电功率相关性。
步骤S103：以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条 件并得到对应的多能源联合发电系统的功率预测模型。纳什效率系数用于验证模型模拟结果的好坏。
具体地，首先，通过下式计算每一种发电方式对应的纳什效率系数(Nash-Sutcliffe efficiency coefficient，NSE)：
式中，T表示用于模型训练的每一种发电方式的历史发电功率数据的样本总数，比如每一种发电方式的历史发电功率数据中70％的历史发电功率数据的样本总数；y1(t)表示用于模型训练的每一种发电方式的历史发电功率数据中第t个样本的功率数据；表示用于模型训练的每一种发电方式的历史发电功率数据中第t个样本的预测功率数据；通过实际获取的发电功率数据y1(t)与预测功率数据进行运算比较来计算纳什效率系数，可以验证模型预测结果的好坏。
然后，根据多能源中任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建下述损失函数，以多能源联合发电系统包含3种发电方式为例，则损失函数计算公式如下式所示：
L＝lossNSE+lossr…………………………………(3)
式中，lossNSE表示纳什效率系数对应的损失函数；lossr表示任意两种发电方式之间的发电功率相关性对应的损失函数；
其中：
lossNSE＝3-(NSE1+NSE2+NSE3)…………………………………(4)
式中，NSE1、NSE2、NSE3分别表示不同发电方式对应的纳什效率系数，具体参考公式(2)进行计算；r12、r13、r23分别表示模型输入的任意两种发电方式之间的发电功率相关性，由历史发电功率计算得出。其中r12为发电方式1和发电方式2之间的发电功率相关性、r13为发电方式1和发电方式3之间的发电功率相关性、r23为发电方式2和发电方式3之间的发电功率相关性；分别表示模型输出的对应的任意两种发电方式之间的发电功率相关性，由预测发电功率计算得到；具体地，r12、r13、r23均参考公式(1)进行计算；
λ12、λ13、λ23分别表示对应的惩罚项参数，取值方式如下：


式中，αr表示历史发电功率与预测发电功率相关性的相对差异阈值，取值范围为[0.1,0.3]，具体取值需根据决策者偏好进一步确定；
利用构建的损失函数对模型进行训练：以每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据对该预设网络模型进行训练，直至该损失函数取值最小(Lmin)时训练结束并得到对应的多能源联合发电系统的功率预测模型。
在一实施例中，当多能源联合发电系统包括三种不同的发电方式(方式1、2、3)时，预设网络模型为三个不同的LSTM模型，且对应的损失函数均为公式3所述的损失函数；将三种发电方式的历史发电功率数据分别输入对应的LSTM模型中进行训练后，得到多能源联合发电系统的功率预测模型{LSTM1，LSTM2，LSTM3}。其中，该多能源联合发电系统的功率预测模型为三个训练好的发电功率预测模型LSTM1、LSTM2、LSTM3组成，分别用于预测不同发电方式对应的功率数据。
作为本发明实施例一种可选的实施方式，还可以对多能源联合发电系统的功率预测模型的参数进行优化。
具体地，当多能源联合发电系统包括三种不同的发电方式(方式1、2、3)时，该多能源联合发电系统的功率预测模型对应的超参数可以包括：记忆单元数{MC1，MC2，MC3}，网络层数{La1，La2，La3}，学习率{R1，R2，R3}，批次大小{B1，B2，B3}，时间展开步数{ts1，ts2，ts3}，以及梯度下降算法的选择{G1，G2，G3}。
分别在给定一组超参数的情况下，使用训练集(历史发电功率数据中70％的历史发电功率数据)训练得到三种不同发电方式各自对应的发电功率预测模型LSTM1、LSTM2、LSTM3，将剩余30％的历史发电功率数据作为验证集计算出此时对应的目标函数的取值。
获取目标函数取值最小时每一种发电方式对应的发电功率预测模型的超参数组合，并最终获取得到该超参数组合下对应的最佳多能源联合发电系统的功率预测模型 {LSTM1(best)，LSTM2(best)，LSTM3(best)}。
本发明实施例提供的多能源联合发电系统的功率预测模型构建方法，构建了由纳什效率系数与发电功率相关性组成的损失函数，兼顾了模型预测的精度提高需求和多能源联合发电系统的功率预测的相关性，为多能源互补调度计划编制提供了基础性、精准的数据支撑。
作为本发明实施例一种可选的实施方式，步骤S102之前，所述方法还包括：对获取到的所述历史发电功率数据进行异常数据和/或缺失数据的识别操作；对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理。
首先，利用箱线图对该历史发电功率数据中的异常数据进行识别处理。
具体地，利用箱线图识别异常数据表示将大于或小于箱线图设定的上界(UB)和下界(LB)的数据视为异常数据。其中，箱线图如图2所示。
其中，UB和LB的计算公式为：
UB＝U+1.5(U-L)…………………(9)
LB＝L-1.5(U-L)…………………(10)
式中，U为上四分位数，表示某一类发电方式对应的历史发电功率数据中只有1/4的数据大于U；L为下四分位数，表示某一类发电方式对应的历史发电功率数据中只有1/4的数据小于U。
其次，对获取到的历史发电功率数据进行缺失数据的识别操作，包括：当获取的任意两个相邻的所述历史发电功率数据对应的时间间隔大于预设时长，判定所述两个相邻的历史发电功率数据为历史发电功率缺失数据。比如，将该历史发电功率数据中连续缺失16个时刻以上的数据视为历史发电功率缺失数据。
然后，对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理，包括：在获取到的历史发电功率数据中删除所述历史发电功率缺失数据并利用预设监督学习方法对所述历史发电功率异常数据进行处理。
预设监督学习方法可以包括K近邻互补法、朴素贝叶斯、决策树、EM算法等，本发明对此不做具体限定，只要满足需求即可。
具体地，在该历史发电功率数据中删除该历史发电功率缺失数据。
在一实施例中，利用K近邻互补法处理该历史发电功率异常数据的计算公式为：
式中，xj表示某一类发电方式对应的历史发电功率数据中第j个样本的数据为异常数据xj；xj-k表示xj前第k个数据；xj+k表示xj后第k个数据；k值一般取值为2-5，根据实际需求 确定即可。
作为本发明实施例一种可选的实施方式，当多能源联合发电系统为水、风、光多能互补联合发电系统时，获取多能源联合发电系统中每一种发电方式的历史发电功率数据，包括：获取水电站历史数据并根据所述水电站历史数据计算对应的水能历史发电功率数据。其中，水电站历史数据包括但不限于入库流量、水位-库容关系曲线、下泄流量-尾水位关系曲线、水电站水库水位上下限值、水电站水库最大发电流量值、最小允许下泄流量、水电站出力系数、水电站装机容量、水电站保证出力、水电站调度规程等。
然后，通过下式计算对应的水能历史发电功率数据：
Pt＝min(Pmax，ηQEtΔZt)…………………(13)
式中，Pmax表示装机容量；η表示出力系数；QEt表示水电站发电流量，为最大允许发电流量QEmax和出库流量Qt两者中的较小值；ΔZt表示发电水头差；
其中：


Vt+1＝Vt+(It-Qt)Δt…………………(17)
式中，表示上游水位，根据当前平均库容和水电站水位-库容关系曲线Zup～f(V)计算得到；表示尾水位，根据当前出库流量Qt和水电站尾水位-泄流曲线Zdown～g(Q)计算得到；Vt+1表示第t+1时段的库容值；Vt表示第t时段的库容值；It表示入库流量；Qt表示出库流量，可根据调度规程或既定调度方案Qt＝f(θt，It，Vt)计算得到；其中，水电站调度方案参数θt＝{θ1，t，…，θM，t}取决于研究对象、采用的调度规则描述型式及调度规则呈现方式。
本发明实施例提供一种多能源联合发电系统的功率预测方法，如图3所示，该方法包括如下步骤：
步骤S201：获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子。
步骤S202：将所述每一种发电方式的发电功率数据对应的气象因子输入如本发明实施例所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到所述待预测多能源联合发电系统的功率。具体地，可以利用训练好的多能源联合发电系统的功率预测模型获取待预测多能源联合发电系统的功率。
在一实施例中，将待预测日的水、风、光气象因子输入多能源联合发电系统的功率预测 模型{LSTM1(best)，LSTM2(best)，LSTM3(best)}中，该模型的输出{P1，P2，P3}即为水、风、光联合发电系统的功率的最终预测结果。
本发明实施例提供的多能源联合发电系统的功率预测方法，利用训练好的多能源联合发电系统的功率预测模型进行预测，实现了多能源联合发电系统功率的同步、联合预测。
本发明实施例还提供一种多能源联合发电系统的功率预测模型构建装置，如图4所示，该装置包括：
第一获取模块401，用于获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素；详细内容参见上述方法实施例中步骤S101的相关描述。
第一输入模块402，用于将所述每一种发电方式的历史发电功率数据以及对应的历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；详细内容参见上述方法实施例中步骤S102的相关描述。
训练模块403，用于以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型；详细内容参见上述方法实施例中步骤S103的相关描述。
本发明实施例提供的多能源联合发电系统的功率预测模型构建装置，构建了由纳什效率系数与发电功率相关性组成的损失函数，兼顾了模型预测的精度提高需求和多能源联合发电系统的功率预测的相关性，为多能源互补调度计划编制提供了基础性、精准的数据支撑。
作为本发明实施例一种可选的实施方式，所述装置还包括：第一识别模块，用于对获取到的所述历史发电功率数据进行异常数据和/或缺失数据的识别操作；第一处理模块，用于对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理。
作为本发明实施例一种可选的实施方式，所述第一识别模块包括：第一判定子模块，用于当获取的任意两个相邻的所述历史发电功率数据对应的时间间隔大于预设时长，判定所述两个相邻的历史发电功率数据为历史发电功率缺失数据。
作为本发明实施例一种可选的实施方式，所述第一处理模块包括：第一处理子模块，用于在获取到的历史发电功率数据中删除所述历史发电功率缺失数据并利用预设监督学习方法对所述历史发电功率异常数据进行处理。
作为本发明实施例一种可选的实施方式，所述多能源联合发电系统包括水能发电系统。
作为本发明实施例一种可选的实施方式，所述第一获取模块包括：第一计算子模块，用于获取水电站历史数据并根据所述水电站历史数据计算对应的水能历史发电功率数据。
本发明实施例提供的多能源联合发电系统的功率预测模型构建装置的功能描述详细参见上述实施例中多能源联合发电系统的功率预测模型构建方法描述。
本发明实施例还提供一种多能源联合发电系统的功率预测装置，如图5所示，该装置包括：
第二获取模块501，用于获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子；详细内容参见上述方法实施例中步骤S201的相关描述。
第二输入模块502，用于将所述每一种发电方式的发电功率数据对应的气象因子输入如本发明实施例所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到所述待预测多能源联合发电系统的功率；详细内容参见上述方法实施例中步骤S202的相关描述。
本发明实施例提供的多能源联合发电系统的功率预测装置，利用训练好的多能源联合发电系统的功率预测模型进行预测，实现了多能源联合发电系统功率的同步、联合预测。
本发明实施例提供的多能源联合发电系统的功率预测装置的功能描述详细参见上述实施例中多能源联合发电系统的功率预测方法描述。
本发明实施例还提供一种存储介质，如图6所示，其上存储有计算机程序601，该指令被处理器执行时实现上述实施例中多能源联合发电系统的功率预测模型构建方法或多能源联合发电系统的功率预测方法的步骤。其中，存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)、随机存储记忆体(Random Access Memory，RAM)、快闪存储器(Flash Memory)、硬盘(Hard Disk Drive，缩写：HDD)或固态硬盘(Solid-State Drive，SSD)等；所述存储介质还可以包括上述种类的存储器的组合。
本领域技术人员可以理解，实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)、随机存储记忆体(Random Access Memory，RAM)、快闪存储器(Flash Memory)、硬盘(Hard Disk Drive，缩写：HDD)或固态硬盘(Solid-State Drive，SSD)等；所述存储介质还可以包括上述种类的存储器的组合。
本发明实施例还提供了一种电子设备，如图7所示，该电子设备可以包括处理器71和 存储器72，其中处理器71和存储器72可以通过总线或者其他方式连接，图7中以通过总线连接为例。
处理器71可以为中央处理器(Central Processing Unit，CPU)。处理器71还可以为其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等芯片，或者上述各类芯片的组合。
存储器72作为一种非暂态计算机可读存储介质，可用于存储非暂态软件程序、非暂态计算机可执行程序以及模块，如本发明实施例中的对应的程序指令/模块。处理器71通过运行存储在存储器72中的非暂态软件程序、指令以及模块，从而执行处理器的各种功能应用以及数据处理，即实现上述方法实施例中的多能源联合发电系统的功率预测模型构建方法或多能源联合发电系统的功率预测方法。
存储器72可以包括存储程序区和存储数据区，其中，存储程序区可存储操作装置、至少一个功能所需要的应用程序；存储数据区可存储处理器71所创建的数据等。此外，存储器72可以包括高速随机存取存储器，还可以包括非暂态存储器，例如至少一个磁盘存储器件、闪存器件、或其他非暂态固态存储器件。在一些实施例中，存储器72可选包括相对于处理器71远程设置的存储器，这些远程存储器可以通过网络连接至处理器71。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
所述一个或者多个模块存储在所述存储器72中，当被所述处理器71执行时，执行如图1-3所示实施例中的多能源联合发电系统的功率预测模型构建方法或多能源联合发电系统的功率预测方法。
上述电子设备具体细节可以对应参阅图1至图3所示的实施例中对应的相关描述和效果进行理解，此处不再赘述。

Claims (9)

  1. 一种多能源联合发电系统的功率预测模型构建方法，其特征在于，包括如下步骤：
    获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素；
    将所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；
    以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型。
  2. 根据权利要求1所述的方法，其特征在于，将所述每一种发电方式的历史发电功率数据以及对应的历史发电功率数据的气象因子输入预设网络模型之前，所述方法还包括：
    对获取到的所述历史发电功率数据进行异常数据和/或缺失数据的识别操作；
    对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理。
  3. 根据权利要求2所述的方法，其特征在于，对获取到的所述历史发电功率数据进行缺失数据的识别操作，包括：
    当获取的任意两个相邻的所述历史发电功率数据对应的时间间隔大于预设时长，判定所述两个相邻的历史发电功率数据为历史发电功率缺失数据。
  4. 根据权利要求2所述的方法，其特征在于，对识别到的历史发电功率异常数据和历史发电功率缺失数据进行异常处理，包括：
    在获取到的历史发电功率数据中删除所述历史发电功率缺失数据并利用预设监督学习方法对所述历史发电功率异常数据进行处理。
  5. 根据权利要求1所述的方法，其特征在于，所述多能源联合发电系统包括水能发电系统。
  6. 根据权利要求5所述的方法，其特征在于，获取多能源联合发电系统中每一种发电方式的历史发电功率数据，包括：
    获取水电站历史数据并根据所述水电站历史数据计算对应的水能历史发电功率数据。
  7. 一种多能源联合发电系统的功率预测方法，其特征在于，包括如下步骤：
    获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子；
    将所述每一种发电方式的发电功率数据对应的气象因子输入如权利要求1-6任一项所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到所述待预测多能源联合发电系统的功率。
  8. 一种多能源联合发电系统的功率预测模型构建装置，其特征在于，包括：
    第一获取模块，用于获取多能源联合发电系统中每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子，所述气象因子表征影响发电功率的气象因素；
    第一输入模块，用于将所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子输入预设网络模型，使得所述预设网络模型按照预设相关性计算方法计算得到多种发电方式中任意两种发电方式之间的发电功率相关性；
    训练模块，用于以所述任意两种发电方式之间的发电功率相关性以及每一种发电方式对应的纳什效率系数构建损失函数、以所述每一种发电方式的历史发电功率数据以及对应于历史发电功率数据的气象因子作为训练数据，对所述预设网络模型进行训练直至满足预设训练条件并得到对应的多能源联合发电系统的功率预测模型。
  9. 一种多能源联合发电系统的功率预测装置，其特征在于，包括：
    第二获取模块，用于获取待预测多能源联合发电系统中每一种发电方式的发电功率数据对应的气象因子；
    第二输入模块，用于将所述每一种发电方式的发电功率数据对应的气象因子输入如权利要求1-6任一项所述的多能源联合发电系统的功率预测模型构建方法构建得到的多能源联合发电系统的功率预测模型，得到所述待预测多能源联合发电系统的功率。
PCT/CN2023/113555 2022-08-31 2023-08-17 一种多能源联合发电系统的功率预测模型构建及预测方法 WO2024046137A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211059975.2 2022-08-31
CN202211059975.2A CN115425680B (zh) 2022-08-31 2022-08-31 一种多能源联合发电系统的功率预测模型构建及预测方法

Publications (1)

Publication Number Publication Date
WO2024046137A1 true WO2024046137A1 (zh) 2024-03-07

Family

ID=84199577

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/113555 WO2024046137A1 (zh) 2022-08-31 2023-08-17 一种多能源联合发电系统的功率预测模型构建及预测方法

Country Status (2)

Country Link
CN (1) CN115425680B (zh)
WO (1) WO2024046137A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117852422A (zh) * 2024-03-08 2024-04-09 宏华海洋油气装备（江苏）有限公司 一种基于组合优化的海上风电浮式基础主尺度优化方法
CN117908456A (zh) * 2024-03-20 2024-04-19 国能大渡河检修安装有限公司 基于深度学习的水电站启闭机监控方法及系统
CN118100171A (zh) * 2024-04-19 2024-05-28 阿里巴巴（中国）有限公司 能源转换系统的功率预测方法、系统和电子设备

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115425680B (zh) * 2022-08-31 2023-07-18 中国长江三峡集团有限公司 一种多能源联合发电系统的功率预测模型构建及预测方法
CN116581755B (zh) * 2023-07-12 2023-09-29 长江水利委员会水文局 功率预测方法、装置、设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110295610A1 (en) * 2010-05-26 2011-12-01 Yuan Ze University Method for configuring installation capacities of hybrid energy generation system
CN113497445A (zh) * 2021-09-08 2021-10-12 国网江西省电力有限公司电力科学研究院 一种区域多尺度新能源电站出力联合预测方法及系统
CN115425680A (zh) * 2022-08-31 2022-12-02 中国长江三峡集团有限公司 一种多能源联合发电系统的功率预测模型构建及预测方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170286838A1 (en) * 2016-03-29 2017-10-05 International Business Machines Corporation Predicting solar power generation using semi-supervised learning
CN108832663A (zh) * 2018-07-18 2018-11-16 北京天诚同创电气有限公司 微电网光伏发电系统的发电功率的预测方法及设备
CN112234657B (zh) * 2020-09-24 2022-08-30 国网山东省电力公司电力科学研究院 基于新能源联合出力和需求响应的电力系统优化调度方法
CN113496311A (zh) * 2021-06-25 2021-10-12 国网山东省电力公司济宁供电公司 光伏电站发电功率预测方法及系统
CN114548509A (zh) * 2022-01-18 2022-05-27 湖南大学 一种多能源系统多类型负荷联合预测方法及系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110295610A1 (en) * 2010-05-26 2011-12-01 Yuan Ze University Method for configuring installation capacities of hybrid energy generation system
CN113497445A (zh) * 2021-09-08 2021-10-12 国网江西省电力有限公司电力科学研究院 一种区域多尺度新能源电站出力联合预测方法及系统
CN115425680A (zh) * 2022-08-31 2022-12-02 中国长江三峡集团有限公司 一种多能源联合发电系统的功率预测模型构建及预测方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117852422A (zh) * 2024-03-08 2024-04-09 宏华海洋油气装备（江苏）有限公司 一种基于组合优化的海上风电浮式基础主尺度优化方法
CN117852422B (zh) * 2024-03-08 2024-05-10 宏华海洋油气装备（江苏）有限公司 一种基于组合优化的海上风电浮式基础主尺度优化方法
CN117908456A (zh) * 2024-03-20 2024-04-19 国能大渡河检修安装有限公司 基于深度学习的水电站启闭机监控方法及系统
CN118100171A (zh) * 2024-04-19 2024-05-28 阿里巴巴（中国）有限公司 能源转换系统的功率预测方法、系统和电子设备

Also Published As

Publication number Publication date
CN115425680A (zh) 2022-12-02
CN115425680B (zh) 2023-07-18

Similar Documents

Publication Publication Date Title
WO2024046137A1 (zh) 一种多能源联合发电系统的功率预测模型构建及预测方法
Zhang et al. Short-term forecasting and uncertainty analysis of wind turbine power based on long short-term memory network and Gaussian mixture model
US11204591B2 (en) Modeling and calculating normalized aggregate power of renewable energy source stations
CN110556820B (zh) 用于确定能量系统操作场景的方法和设备
Capizzi et al. Advanced and adaptive dispatch for smart grids by means of predictive models
Liao et al. Ultra-short-term interval prediction of wind power based on graph neural network and improved bootstrap technique
CN110380444B (zh) 一种基于变结构Copula的多场景下分散式风电有序接入电网的容量规划方法
CN113988477A (zh) 基于机器学习的光伏功率短期预测方法、装置及存储介质
CN116581755B (zh) 功率预测方法、装置、设备及存储介质
Fusco et al. Knowledge-and data-driven services for energy systems using graph neural networks
Wang et al. Short‐Term Wind Speed Forecasting Using the Data Processing Approach and the Support Vector Machine Model Optimized by the Improved Cuckoo Search Parameter Estimation Algorithm
CN108233357A (zh) 基于非参数概率预测及风险期望的风电日前消纳优化方法
Ibrahim et al. LSTM neural network model for ultra-short-term distribution zone substation peak demand prediction
CN113112085A (zh) 一种基于bp神经网络的新能源场站发电负荷预测方法
Lai et al. Short‐term passenger flow prediction for rail transit based on improved particle swarm optimization algorithm
CN116937565A (zh) 一种分布式光伏发电功率预测方法、系统、设备及介质
CN107016470A (zh) 风力发电场风能预测方法和装置
CN115952921A (zh) 一种光伏能源功率预测方法、装置、电子设备及存储介质
CN115965134A (zh) 一种区域电网风力发电功率预测优化方法
CN112821456B (zh) 基于迁移学习的分布式源-储-荷匹配方法及装置
CN115275975A (zh) 一种光储充电站电力数据匹配度的确定方法及装置
CN114857656B (zh) 一种供热负荷预测方法和装置
CN112215383A (zh) 一种分布式光伏发电功率预测方法和系统
CN112001518A (zh) 一种基于云计算的预测和能量管理方法及系统
Bâra et al. Intelligent systems for predicting and analyzing data in power grid companies

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23859171

Country of ref document: EP

Kind code of ref document: A1