CN117595384B - Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence - Google Patents
Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117595384B CN117595384B CN202410071689.0A CN202410071689A CN117595384B CN 117595384 B CN117595384 B CN 117595384B CN 202410071689 A CN202410071689 A CN 202410071689A CN 117595384 B CN117595384 B CN 117595384B
- Authority
- CN
- China
- Prior art keywords
- power generation
- historical
- acquiring
- power
- real
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 22
- 230000002457 bidirectional effect Effects 0.000 title claims abstract description 22
- 238000010248 power generation Methods 0.000 claims abstract description 152
- 230000005611 electricity Effects 0.000 claims abstract description 120
- 238000004146 energy storage Methods 0.000 claims abstract description 65
- 238000012544 monitoring process Methods 0.000 claims abstract description 41
- 230000002159 abnormal effect Effects 0.000 claims abstract description 39
- 230000005540 biological transmission Effects 0.000 claims abstract description 37
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000004364 calculation method Methods 0.000 claims abstract description 31
- 238000004458 analytical method Methods 0.000 claims abstract description 21
- 238000007726 management method Methods 0.000 claims abstract description 14
- 230000005855 radiation Effects 0.000 claims description 93
- 238000013528 artificial neural network Methods 0.000 claims description 24
- 239000000428 dust Substances 0.000 claims description 19
- 238000010276 construction Methods 0.000 claims description 15
- 238000004088 simulation Methods 0.000 claims description 12
- 238000013507 mapping Methods 0.000 claims description 11
- 238000009825 accumulation Methods 0.000 claims description 9
- 230000003287 optical effect Effects 0.000 claims description 9
- 230000003213 activating effect Effects 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 description 10
- 230000006870 function Effects 0.000 description 5
- 238000005457 optimization Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000009472 formulation Methods 0.000 description 2
- RVRCFVVLDHTFFA-UHFFFAOYSA-N heptasodium;tungsten;nonatriacontahydrate Chemical compound O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.O.[Na+].[Na+].[Na+].[Na+].[Na+].[Na+].[Na+].[W].[W].[W].[W].[W].[W].[W].[W].[W].[W].[W] RVRCFVVLDHTFFA-UHFFFAOYSA-N 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- 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
- H02J3/28—Arrangements for balancing of the load in a network by storage of 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
- 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
- H02J2300/22—The renewable source being solar energy
- H02J2300/24—The renewable source being solar energy of photovoltaic origin
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Power Engineering (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Marketing (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a photovoltaic power station source network load bidirectional prediction method and a system based on artificial intelligence, which relate to the technical field of photovoltaic power stations and comprise the following steps: constructing a sunlight calculation branch; acquiring historical power generation information; training to obtain a power generation amount prediction branch, and obtaining a power generation amount prediction model; acquiring real-time environment monitoring data, inputting a power generation amount prediction model to predict power generation amount, and acquiring a predicted power generation amount sequence; acquiring historical electricity utilization information, and carrying out electricity utilization trend analysis to obtain a first predicted electricity utilization sequence; acquiring line loss, acquiring a first energy storage scheme and acquiring a second predicted electricity consumption sequence; comparing to obtain an abnormal time node set; and optimizing the first energy storage scheme, and carrying out energy storage management. The method solves the technical problems that the traditional method can not accurately predict the generated energy of the photovoltaic power station according to meteorological data, and lacks accurate calculation of line loss in the power transmission process of the power grid, so that the prediction accuracy is poor and the utilization rate of an energy system is low.
Description
Technical Field
The invention relates to the technical field of photovoltaic power stations, in particular to a photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence.
Background
Photovoltaic power plants are one of the important sources of renewable energy sources, but due to the uncertainty of solar conditions, the prediction of the power production of photovoltaic power plants becomes crucial. On one hand, the power generation of the photovoltaic power station is influenced by meteorological conditions, and the problems of the traditional method in power generation prediction are caused by uncertainty and variability of weather; on the other hand, the line loss exists in the power transmission process, the efficiency of the energy storage system is also affected, and the factors are difficult to comprehensively consider by the traditional method, so that the utilization rate of the energy system is not high.
Disclosure of Invention
The invention provides a photovoltaic power station source network load bidirectional prediction method based on artificial intelligence, which aims to solve the technical problems that the traditional method can not accurately predict the generated energy of a photovoltaic power station according to meteorological data, and the prediction accuracy is poor and the utilization rate of an energy system is not high due to the lack of accurate calculation of line loss in the power transmission process of a power grid.
In view of the above problems, the application provides a photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence.
In a first aspect of the disclosure, a photovoltaic power station source network load bidirectional prediction method based on artificial intelligence is provided, and the method comprises:
acquiring climate information according to the geographic position of a photovoltaic power station, and constructing a sunlight quantity calculation branch according to the climate information, wherein the sunlight quantity calculation branch comprises sunlight radiation intensity and cloud layer thickness;
acquiring historical power generation information of a photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
according to the historical power generation information, training to obtain a power generation amount prediction branch, and integrating the sunlight amount calculation branch and the power generation amount prediction branch to obtain a power generation amount prediction model;
performing real-time environment monitoring on the photovoltaic power station, acquiring real-time environment monitoring data, inputting the real-time environment monitoring data into the power generation amount prediction model for power generation amount prediction, and acquiring a predicted power generation amount sequence of a first preset period, wherein the real-time environment monitoring data comprises real-time solar radiation intensity and real-time cloud layer thickness;
acquiring historical electricity consumption information of an electricity consumption area, and carrying out electricity consumption trend analysis according to the historical electricity consumption information to obtain a first predicted electricity consumption sequence of a first preset period;
Acquiring line loss, acquiring a first energy storage scheme, and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the first energy storage scheme and the predicted electricity generation sequence;
comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence to obtain an abnormal time node set, wherein the abnormal time node set is a time node with a deviation value of the first predicted electricity consumption and the second predicted electricity consumption larger than or equal to a preset deviation value;
and optimizing the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set, and carrying out energy storage management according to the optimized first energy storage scheme.
In another aspect of the disclosure, a photovoltaic power station source network load bidirectional prediction system based on artificial intelligence is provided, the system is used in the above method, and the system includes:
the solar radiation system comprises a first branch construction module, a second branch construction module and a third branch construction module, wherein the first branch construction module is used for acquiring climate information according to the geographic position of a photovoltaic power station and constructing a solar radiation calculation branch according to the climate information, and the solar radiation calculation branch comprises solar radiation intensity and cloud layer thickness;
The power generation information acquisition module is used for acquiring historical power generation information of the photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
the second branch construction module is used for training to obtain a power generation amount prediction branch according to the historical power generation information, integrating the solar radiation amount calculation branch and the power generation amount prediction branch and obtaining a power generation amount prediction model;
the power generation amount prediction module is used for carrying out real-time environment monitoring on the photovoltaic power station, acquiring real-time environment monitoring data, inputting the real-time environment monitoring data into the power generation amount prediction model for carrying out power generation amount prediction, and acquiring a predicted power generation amount sequence of a first preset period, wherein the real-time environment monitoring data comprises real-time solar radiation intensity and real-time cloud layer thickness;
the power consumption trend analysis module is used for obtaining historical power consumption information of a power consumption area, carrying out power consumption trend analysis according to the historical power consumption information and obtaining a first predicted power consumption sequence of a first preset period;
The predicted electricity consumption acquisition module is used for acquiring line loss, acquiring a first energy storage scheme and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the first energy storage scheme and the predicted electricity generation sequence;
the abnormal node acquisition module is used for comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence to acquire an abnormal time node set, wherein the abnormal time node set is a time node with a deviation value of the first predicted electricity consumption and the second predicted electricity consumption being greater than or equal to a preset deviation value;
the energy storage management module is used for optimizing the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set and carrying out energy storage management according to the optimized first energy storage scheme.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
the climate information is acquired through the geographical position of the photovoltaic power station, and a sunlight calculation branch is constructed, wherein the branch comprises sunlight radiation intensity and cloud layer thickness, so that the generation capacity of the photovoltaic power station can be predicted more accurately; the method comprises the steps of obtaining a mapping data set of historical sunshine and generated energy, training a generated energy prediction branch through a neural network and other methods, and capturing a complex relationship between the historical sunshine and the generated energy better through a deep learning model, so that the prediction precision is improved; the method comprises the steps of acquiring real-time environment monitoring data, including real-time solar radiation intensity and cloud layer thickness, inputting a prediction model, acquiring a predicted generating capacity sequence of a first preset period, and improving the response capability to real-time environment changes, so that the generating capacity prediction is more flexible and accurate; historical electricity consumption of the electricity consumption area is obtained, electricity consumption trend analysis is carried out, a first predicted electricity consumption sequence of a first preset period is obtained, a basis for predicting electricity consumption is provided for power grid planning, and better matching of electricity generation and electricity consumption is facilitated; calculating the line loss, acquiring a first energy storage scheme, and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the energy storage scheme and the predicted electricity generation amount, wherein the line loss is comprehensively considered, more comprehensive information is provided for the formulation of the energy storage scheme, and the improvement of the electric energy utilization rate is facilitated; comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence, acquiring an abnormal time node, optimizing a first energy storage scheme according to the abnormal node and the corresponding deviation value, improving the sensitivity to abnormal conditions, and being beneficial to realizing more stable and efficient energy supply by optimizing the energy storage scheme. In summary, the method comprehensively considers various factors such as the power generation condition, the real-time environmental condition, the power consumption trend, the line loss and the like of the photovoltaic power station, supports the artificial intelligence technology, realizes the comprehensive optimization of the bidirectional prediction of the source network load of the photovoltaic power station, and improves the intellectualization and the sustainability of an energy system.
The foregoing description is only an overview of the technical solutions of the present application, and may be implemented according to the content of the specification in order to make the technical means of the present application more clearly understood, and in order to make the above-mentioned and other objects, features and advantages of the present application more clearly understood, the following detailed description of the present application will be given.
Drawings
Fig. 1 is a schematic flow chart of a bidirectional prediction method for source network load of a photovoltaic power station based on artificial intelligence according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a photovoltaic power station source network load bidirectional prediction system based on artificial intelligence according to an embodiment of the present application.
Reference numerals illustrate: the system comprises a first branch construction module 10, a power generation information acquisition module 20, a second branch construction module 30, a power generation amount prediction module 40, a power consumption trend analysis module 50, a predicted power consumption acquisition module 60, an abnormal node acquisition module 70 and an energy storage management module 80.
Detailed Description
According to the photovoltaic power station source network load bidirectional prediction method based on artificial intelligence, the technical problems that the traditional method cannot accurately predict the power generation amount of a photovoltaic power station according to meteorological data, and the prediction accuracy is poor and the utilization rate of an energy system is low due to the fact that accurate calculation of line loss in the power transmission process of a power grid is lacked are solved.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, an embodiment of the present application provides an artificial intelligence-based bidirectional prediction method for a photovoltaic power station source network load, where the method includes:
acquiring climate information according to the geographic position of a photovoltaic power station, and constructing a sunlight quantity calculation branch according to the climate information, wherein the sunlight quantity calculation branch comprises sunlight radiation intensity and cloud layer thickness;
the method comprises the steps of obtaining accurate geographic coordinates of a photovoltaic power station through a GPS or other geographic positioning tools, obtaining relevant weather information from weather data sources such as a weather station, a weather satellite and the like by utilizing geographic position information of the photovoltaic power station, wherein the obtained weather information comprises temperature, humidity, wind speed and the like, sunlight radiation intensity and cloud cover thickness are particularly concerned, and constructing a sunlight quantity calculation branch by utilizing the obtained weather information, wherein the sunlight quantity calculation branch comprises an expression for calculating sunlight quantity according to the sunlight radiation intensity and the cloud cover thickness.
Further, the expression of the sunlight amount calculation branch is:
；
wherein Q is the sunlight quantity, And k is a constant related to optical characteristics of cloud, h is cloud layer thickness, and T is a sunlight scale factor which represents the ratio of actual sunlight time to total daytime time.
In this expression, Q represents the amount of solar radiation received by the photovoltaic power plant at a given moment,indicating the maximum intensity of solar radiation in the ideal case without cloud shielding, +.>The larger the thickness of the cloud layer is, the larger the attenuation factor is, so that the sunlight radiation intensity is reduced, the sunlight proportion factor T is the ratio of the actual sunlight time to the total daytime time, and the T is zero at night or in cloudy days, so that no sunlight exists. From this expression, it can be seen that the insolation quantity Q is subjected to the intensity of insolation radiation when cloudlessAnd the influence of the cloud layer thickness h and the sunlight scale factor T, the expression can be used for simulating the actual sunlight quantity received by the photovoltaic power station under different meteorological conditions, so that the generating capacity of the photovoltaic power station can be predicted more accurately.
Acquiring historical power generation information of a photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
and collecting historical power generation information of the photovoltaic power station through a data recording system, monitoring equipment, a weather station and the like, wherein the historical power generation information comprises historical solar radiation amount and historical power generation amount data. And carrying out time correspondence on the historical sunshine amount and the historical power generation amount data, establishing a mapping relation, and ensuring that the sunshine amount and the power generation amount at each time point are mutually corresponding, wherein the time resolution can be selected according to requirements, such as daily, hourly and every minute, and depends on the acquisition frequency of the data and the modeling requirement. And forming a data pair by the corresponding historical sunshine amount and the historical power generation amount to form a mapping data set of a historical sunshine amount set and a historical power generation amount set.
According to the historical power generation information, training to obtain a power generation amount prediction branch, and integrating the sunlight amount calculation branch and the power generation amount prediction branch to obtain a power generation amount prediction model;
selecting an appropriate machine learning model, such as a neural network, for predicting the generated energy, taking the historical solar energy as an input characteristic, taking the historical solar energy and the corresponding historical generated energy as target output of the model by each data point, carrying out model training by using the historical solar energy and the corresponding historical generated energy data, and training to obtain a generated energy prediction branch.
The solar energy calculation branch and the trained power generation amount prediction branch are integrated together to obtain a power generation amount prediction model, and the model can predict future power generation amount of the photovoltaic power station according to real-time solar conditions and provide important information for power grid dispatching and energy management.
Further, according to the historical power generation information, training to obtain a power generation amount prediction branch comprises the following steps:
acquiring a pre-trained fully-connected neural network;
and training the fully-connected neural network by taking the historical solar radiation set as input data and the historical power generation set as output data to obtain the power generation prediction branch.
The method comprises the steps of obtaining a pre-trained fully-connected neural network, wherein the number of layers of the neural network, the number of neurons of each layer and an activation function are determined, determining the dimension of an input layer, the dimension is matched with the feature quantity of training data, and the dimension of an output layer is matched with the category quantity of tasks. The weights and biases of the network are initialized, for example, using a random initialization method, and pre-trained weights may also be used. An appropriate loss function is established that will be used to measure the performance of the model on the training data, and an optimization algorithm is selected to minimize the loss function, such as random gradient descent, etc.
Loading a fully-connected neural network which is trained in advance, taking the historical solar radiation amount set as input data, taking the historical generating capacity set as output data, forming a training data set by the input data and the output data, and training the fully-connected neural network, specifically dividing the training data set into a training set and a testing set according to a preset proportion, for example, 8:2, inputting the prepared training set into the loaded fully-connected neural network for forward propagation to obtain the predicted output of a model to the training set, carrying out model performance test according to the predicted output of the model, carrying out iterative optimization on model parameters according to test results, and obtaining the generating capacity predicted branch when the model converges, for example, the preset iterative times are reached or the preset accuracy is met;
Performing real-time environment monitoring on the photovoltaic power station, acquiring real-time environment monitoring data, inputting the real-time environment monitoring data into the power generation amount prediction model for power generation amount prediction, and acquiring a predicted power generation amount sequence of a first preset period, wherein the real-time environment monitoring data comprises real-time solar radiation intensity and real-time cloud layer thickness;
for photovoltaic power stations, environmental data are collected in real time through equipment such as weather stations and optical sensors, and real-time environmental monitoring data comprise real-time sunlight radiation intensity and real-time cloud layer thickness. And transmitting the obtained real-time solar radiation intensity and the real-time cloud layer thickness as input to a trained generating capacity prediction model, and predicting the generating capacity of the photovoltaic power station by using the generating capacity prediction model, wherein specifically, when the real-time solar radiation intensity and the real-time cloud layer thickness are input, the real-time solar radiation is obtained through a solar radiation calculation branch and then is input into a generating capacity prediction branch to obtain a generating capacity prediction result in a first preset period, and the specific definition of the first preset period depends on specific requirements of application, such as one day, one week and the like.
Acquiring historical electricity consumption information of an electricity consumption area, and carrying out electricity consumption trend analysis according to the historical electricity consumption information to obtain a first predicted electricity consumption sequence of a first preset period;
Historical electricity usage for electricity usage areas, including electricity usage data at daily, hourly, or other time granularity, is collected from sources such as utility companies, electricity meters, energy monitoring systems, and the like. And carrying out trend analysis on historical electricity consumption by utilizing a statistical method or a time sequence analysis and other technologies, identifying and analyzing the characteristics of periodicity, seasonality, trend and the like of the electricity consumption, and establishing an electricity consumption trend model according to the result of the trend analysis, wherein the method such as linear regression, ARIMA (autoregressive integral sliding average) model and the like can be adopted to capture the change trend of the historical electricity consumption. And predicting future power consumption based on the power consumption trend model to obtain a first predicted power consumption sequence of a first preset period, wherein the first predicted power consumption sequence corresponds to the predicted power generation sequence.
Acquiring line loss, acquiring a first energy storage scheme, and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the first energy storage scheme and the predicted electricity generation sequence;
the power line is monitored or simulated to obtain the line loss, for example, the line loss is calculated by monitoring parameters such as current, voltage and the like in real time, and the simulation can also be carried out by adopting a power system simulation tool. And selecting a first energy storage scheme according to factors such as system requirements and cost consideration, determining parameters such as capacity, charge and discharge rate and the like of the energy storage equipment, wherein the first energy storage scheme is an initial energy storage scheme.
According to the first energy storage scheme, a first energy storage sequence is obtained, the influence of line loss on electric quantity transportation is considered, the electric quantity sequence which can be received by the electricity utilization end under the first energy storage scheme is predicted from the electricity generation end, specifically, the predicted electric quantity sequence is subtracted by the predicted electric quantity sequence, the line loss is correspondingly subtracted by each time node, and the predicted electric quantity which can be received by the electricity utilization end under the first energy storage scheme at each node can be obtained and used as a second predicted electric quantity sequence.
Further, obtaining the line loss includes:
acquiring a power transmission line from a photovoltaic power station to a power utilization area;
generating a first power generation amount in a photovoltaic power station, and carrying out power transmission simulation according to the power transmission line to obtain a simulated power transmission amount;
and acquiring real-time power transmission quantity in a power utilization area, calculating the difference value between the real-time power transmission quantity and the analog power transmission quantity, and acquiring the line loss quantity.
The method comprises the steps of collecting geographic information related to a power transmission line, including coordinates of a starting point and an ending point of the line, the length of the line, the type of the line and other information related to characteristics of the line, integrating and converting the collected geographic information data into a power grid topological structure by using a geographic information system tool, and acquiring technical parameters of the power transmission line, such as resistance, reactance, capacitance and the like, which can be obtained through field measurement.
And selecting a proper electric power system simulation tool, such as MATLAB, for carrying out electric simulation of the electric transmission line, introducing the established power grid topological structure into the selected simulation tool, and setting simulation initial conditions, including the generated power, the electric load and the like of the photovoltaic power station. And running an electrical simulation, simulating a process of transmitting power to a power utilization area along a power transmission line when the photovoltaic power station generates the first power generation amount, and extracting the simulated power transmission amount from the simulation result from the photovoltaic power station to the power utilization area.
Under the condition that the photovoltaic power station generates the first power generation amount, the corresponding real-time power transmission amount data is obtained through the power monitoring system of the power consumption area, the real-time power transmission amount is compared with the simulated power transmission amount obtained through simulation, the comparison of the values of the real-time power transmission amount and the simulated power transmission amount at the same time point is ensured through time synchronization, the difference between the real-time power transmission amount and the simulated power transmission amount is calculated, namely, the line loss amount represents the energy loss condition in the power grid. Wherein the difference between the simulated power output and the actual power output may be affected by a variety of factors including actual changes in line impedance, actual load fluctuations in the power usage area, etc.
Comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence to obtain an abnormal time node set, wherein the abnormal time node set is a time node with a deviation value of the first predicted electricity consumption and the second predicted electricity consumption larger than or equal to a preset deviation value;
For each time node, a deviation value between the first predicted power consumption and the second predicted power consumption is calculated, the deviation value being obtainable by calculating a difference between the two. A preset deviation value is set, the value can be adjusted according to actual requirements and system characteristics, and the preset deviation value represents a threshold value of a power consumption prediction error which can be accepted under normal conditions. For each time node, checking whether the calculated deviation value is larger than or equal to a preset deviation value, and if so, identifying the time node as an abnormal time node. And forming an abnormal time node set by all the time points marked as abnormal time nodes, wherein the set comprises the time nodes with larger deviation between the first predicted power consumption and the second predicted power consumption.
And optimizing the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set, and carrying out energy storage management according to the optimized first energy storage scheme.
And for the abnormal time node set, checking a corresponding deviation value set, knowing a difference value between the predicted and actual electricity consumption when the abnormal time node occurs, wherein the larger the difference value is the larger the electricity consumption requirement and the higher the inaccuracy of the abnormal time node, and according to the analysis result of the abnormal time node, optimizing a first energy storage scheme to improve the adaptability of the system to emergency conditions, wherein an optimization strategy comprises the steps of releasing energy storage, adjusting a charging and discharging strategy, optimizing the performance parameters of energy storage equipment and the like.
According to the optimized first energy storage scheme, an energy storage management strategy is adjusted, and when an abnormal time node occurs, the system can more intelligently adjust the operation mode of the energy storage equipment so as to cope with the situation that the prediction deviation of the power consumption is large. The optimized first energy storage scheme and the adjusted energy storage management strategy are implemented into an actual system, so that the system can be flexibly adjusted according to real-time conditions and abnormal conditions, and the overall performance of the energy storage system is improved.
Further, inputting the real-time environmental monitoring data into the power generation amount prediction model to perform power generation amount prediction, and further comprising:
extracting a historical solar radiation intensity set and a historical cloud cover thickness set based on the historical solar quantity set;
clustering and dividing the historical solar radiation intensity set and the historical cloud layer thickness set respectively to obtain N historical solar radiation intensity clusters, M historical cloud layer thickness clusters and corresponding historical solar radiation intensity clustersThe N historical solar radiation intensity clusters and the M historical cloud layer thickness clusters are provided with grade marks, and M, N is an integer greater than or equal to 2;
based on a neural network, constructing a second power generation amount prediction model, and adopting the N historical solar radiation intensity clusters, the M historical cloud layer thickness clusters and the corresponding historical solar radiation intensity clusters Performing supervision training on the second power generation amount prediction model until a preset standard is met to obtain the second power generation amount prediction model;
and when the level difference between the first solar radiation intensity cluster and the first cloud layer thickness cluster is larger than a preset level difference, activating the second generated energy prediction model to predict generated energy.
Extracting solar radiation intensity data corresponding to each time point from the historical solar quantity collection, wherein the solar radiation intensity data can come from an illumination sensor, a weather station or other data sources; and extracting cloud thickness data corresponding to each point in time, the cloud thickness data may be from meteorological satellites, radar or other equipment dedicated to cloud monitoring.
And proper clustering algorithms, such as K-means clustering, hierarchical clustering and the like, are selected to perform standardized processing on the historical solar radiation intensity set and the historical cloud layer thickness set, so that the historical solar radiation intensity set and the historical cloud layer thickness set are ensured to be on the same scale, and the influence of a certain feature on a clustering result is avoided.
And carrying out cluster division on the standardized historical solar radiation intensity set and the historical cloud layer thickness set, dividing the data into N historical solar radiation intensity clusters and M historical cloud layer thickness clusters according to a selected clustering algorithm, wherein the values of M and N respectively represent the numbers of the historical solar radiation intensity clusters and the historical cloud layer thickness clusters and can be adjusted according to the characteristics and actual requirements of the data. And the data pair formed by any historical solar radiation intensity cluster and any historical cloud layer thickness cluster corresponds to one historical generating capacity cluster. A rank identification is added to each cluster of historical solar radiation intensities and clusters of historical cloud layer thicknesses, which can be determined based on characteristics of the clustering results, such as an average value of each cluster.
An appropriate neural network architecture is selected, such as a fully connected neural network, a Recurrent Neural Network (RNN), or the like. And taking the historical solar radiation intensity cluster and the historical cloud layer thickness cluster as input features of the neural network, and taking the corresponding historical generating capacity cluster as an output label of the neural network. Depending on the architecture chosen, a suitable number and type of layers are added to the neural network, including input, hidden and output layers, setting activation functions, loss functions, etc.
Adopting the N historical solar radiation intensity clusters, the M historical cloud layer thickness clusters and the corresponding historical solar radiation intensity clustersThe historical generating capacity cluster is used as a data set of supervision training, input characteristics and output labels are paired, and the data set is divided into a training set and a verification set. And (3) performing supervised training on the neural network by using a training set, continuously adjusting network parameters through a back propagation algorithm to reduce errors between predicted output and actual output, performing multiple rounds of training until a preset standard is met, monitoring performance on a verification set in each round of training, ensuring good model generalization, avoiding overfitting, and obtaining a second generated energy prediction model when the model meets the preset standard, for example, reaching a preset iteration number, wherein the model can be used for future generated energy prediction.
Comparing the real-time solar radiation intensity data with solar radiation intensity clusters in the historical data to find a first solar radiation intensity cluster to which the real-time data belongs, and similarly comparing the real-time cloud layer thickness data with cloud layer thickness clusters in the historical data to find a first cloud layer thickness cluster to which the real-time data belongs.
And calculating the grade difference between the first solar radiation intensity cluster and the first cloud layer thickness cluster, and when the grade difference is larger than the preset grade difference, activating the second power generation amount prediction model to conduct power generation amount prediction, wherein the preset grade difference is set according to actual requirements and is used for avoiding the situation that model prediction accuracy is poor due to overlarge real-time solar radiation intensity and real-time cloud layer thickness difference at a certain moment. And predicting the future power generation amount by using a second power generation amount prediction model, and obtaining a predicted power generation amount sequence under special conditions by taking the clustering information of the real-time solar radiation intensity and the cloud layer thickness into consideration according to the cluster mapping relation learned by training by the second power generation amount prediction model.
Through the steps, whether the second power generation amount prediction model needs to be activated or not can be judged according to the real-time meteorological data and the clustering information, so that the accuracy of power generation amount prediction is improved when the power generation environment of the photovoltaic power station is obviously changed.
Further, the method further comprises the following steps:
generating an electricity consumption early warning value according to the second predicted electricity consumption sequence;
the power consumption data of the power consumption area is monitored in real time, and a real-time power consumption peak value is obtained;
and when the real-time electricity consumption peak reaches the electricity consumption early warning value, generating early warning information and sending the early warning information to a user.
The second predicted power consumption sequence is the power consumption that the user terminal predicted from the power generating terminal can receive, if the real-time power consumption data reaches the value, the potential safety hazard such as power failure may be caused, therefore, based on the second predicted power consumption sequence, a threshold lower than the predicted value is set for limiting the real-time power consumption data, the threshold can be determined according to factors such as system requirements, user behaviors, equipment capacity and the like, for example, the threshold is set to 80% of the second predicted power consumption sequence, which is a set standard, and when the real-time power consumption data reaches the threshold, early warning is triggered, so that power consumption protection is performed.
The electric energy consumption data of the electricity consumption area is obtained in real time by using an ammeter or monitoring equipment, the electricity consumption data is continuously monitored, the electricity consumption peak value is calculated in real time, and the instantaneous electricity consumption peak value is captured in a mode of recording the maximum electric energy consumption value in a period of time or using a sliding window.
When the real-time electricity peak reaches a set electricity consumption early warning value, an early warning mechanism is triggered, a notification containing relevant electricity consumption early warning information is generated, the early warning information can be sent to a user in various modes, such as an email, a short message, a push notification, a mobile phone App message and the like, and the generated early warning information is sent to the user, so that the user can quickly know the current electricity consumption condition and respond to the current electricity consumption condition, and electricity consumption protection is realized.
Therefore, the situation that the electricity consumption peak value exceeds the early warning value can be timely notified to the user, the user is helped to take appropriate measures to optimize electricity consumption behavior, and the sustainability and user experience of the electricity consumption system are improved.
Further, the method further comprises the following steps:
carrying out surface dust monitoring on equipment of the photovoltaic power station through an optical sensor to acquire dust accumulation data;
carrying out power generation loss analysis according to the dust accumulation data to obtain a power generation loss index;
and correcting the predicted power generation amount sequence according to the power generation loss index.
Optical sensors are installed on the equipment of the photovoltaic power station to ensure accurate sensing of dust conditions on the equipment surface, the sensors may be of a type that is suitable for environmental conditions and monitoring purposes, such as light scattering sensors and the like. The dust condition of the surface of the photovoltaic power station equipment is monitored in real time by using an optical sensor, and the sensor performs measurement at fixed time intervals or under the triggering of specific events. Real-time data acquired by the sensor are collected and recorded, and the data comprise dust thickness, dust distribution and other information.
Based on the dust accumulation data, a relation model between the power generation loss and the dust accumulation is established, including analysis and field test of historical data, and the influence of the dust accumulation on the power generation efficiency of the photovoltaic power station is analyzed, for example, quantitative analysis is performed by comparing the power generation efficiency at different dust levels. Based on the result of the power generation loss analysis, a power generation loss index, which is a percentage indicating the degree of reduction in power generation efficiency due to dust, is calculated and obtained.
And (3) correlating the calculated power generation loss index with a predicted power generation amount sequence, and according to the size and trend of the power generation loss index, formulating a corresponding correction strategy, including correcting by reducing the predicted power generation amount, and according to the formulated correction strategy, correspondingly adjusting the predicted power generation amount sequence, for example, subtracting a correction value from each time point.
Therefore, the predicted power generation sequence can be timely adjusted according to the actual power generation loss condition, and the accuracy and reliability of prediction are improved, so that the reduction of power generation efficiency caused by dust and other factors can be better dealt with.
In summary, the photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence provided by the embodiment of the application have the following technical effects:
1. the climate information is acquired through the geographical position of the photovoltaic power station, and a sunlight calculation branch is constructed, wherein the branch comprises sunlight radiation intensity and cloud layer thickness, so that the generation capacity of the photovoltaic power station can be predicted more accurately;
2. the method comprises the steps of obtaining a mapping data set of historical sunshine and generated energy, training a generated energy prediction branch through a neural network and other methods, and capturing a complex relationship between the historical sunshine and the generated energy better through a deep learning model, so that the prediction precision is improved;
3. The method comprises the steps of acquiring real-time environment monitoring data, including real-time solar radiation intensity and cloud layer thickness, inputting a prediction model, acquiring a predicted generating capacity sequence of a first preset period, and improving the response capability to real-time environment changes, so that the generating capacity prediction is more flexible and accurate;
4. historical electricity consumption of the electricity consumption area is obtained, electricity consumption trend analysis is carried out, a first predicted electricity consumption sequence of a first preset period is obtained, a basis for predicting electricity consumption is provided for power grid planning, and better matching of electricity generation and electricity consumption is facilitated;
5. calculating the line loss, acquiring a first energy storage scheme, and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the energy storage scheme and the predicted electricity generation amount, wherein the line loss is comprehensively considered, more comprehensive information is provided for the formulation of the energy storage scheme, and the improvement of the electric energy utilization rate is facilitated;
6. comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence, acquiring an abnormal time node, optimizing a first energy storage scheme according to the abnormal node and the corresponding deviation value, improving the sensitivity to abnormal conditions, and being beneficial to realizing more stable and efficient energy supply by optimizing the energy storage scheme.
In summary, the method comprehensively considers various factors such as the power generation condition, the real-time environmental condition, the power consumption trend, the line loss and the like of the photovoltaic power station, supports the artificial intelligence technology, realizes the comprehensive optimization of the bidirectional prediction of the source network load of the photovoltaic power station, and improves the intellectualization and the sustainability of an energy system.
Example two
Based on the same inventive concept as the photovoltaic power station source network load bidirectional prediction method based on artificial intelligence in the foregoing embodiment, as shown in fig. 2, the present application provides a photovoltaic power station source network load bidirectional prediction system based on artificial intelligence, where the system includes:
the first branch construction module 10 is configured to obtain climate information according to a geographic location of the photovoltaic power station, and construct a solar radiation calculation branch according to the climate information, wherein the solar radiation calculation branch includes solar radiation intensity and cloud layer thickness;
the power generation information acquisition module 20 is used for acquiring historical power generation information of the photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
the second branch construction module 30 is used for training to obtain a power generation amount prediction branch according to the historical power generation information, integrating the solar radiation amount calculation branch and the power generation amount prediction branch to obtain a power generation amount prediction model;
The generating capacity prediction module 40 is configured to perform real-time environmental monitoring on the photovoltaic power station, obtain real-time environmental monitoring data, input the real-time environmental monitoring data into the generating capacity prediction model to perform generating capacity prediction, and obtain a predicted generating capacity sequence of a first predetermined period, where the real-time environmental monitoring data includes real-time solar radiation intensity and real-time cloud layer thickness;
the electricity consumption trend analysis module 50 is used for obtaining historical electricity consumption information of an electricity consumption area, and carrying out electricity consumption trend analysis according to the historical electricity consumption information to obtain a first predicted electricity consumption sequence of a first preset period;
the predicted electricity consumption acquisition module 60 is configured to acquire a line loss, acquire a first energy storage scheme, and acquire a second predicted electricity consumption sequence of a first predetermined period in combination with the line loss, the first energy storage scheme, and the predicted electricity generation sequence;
an abnormal node obtaining module 70, where the abnormal node obtaining module 70 is configured to compare the first predicted power consumption sequence and the second predicted power consumption sequence, and obtain an abnormal time node set, where the abnormal time node set is a time node in which a deviation value of the first predicted power consumption and the second predicted power consumption is greater than or equal to a preset deviation value;
The energy storage management module 80 is configured to optimize the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set, and perform energy storage management according to the optimized first energy storage scheme.
Further, the expression of the sunlight amount calculation branch is:
；
wherein Q is the sunlight quantity,and k is a constant related to optical characteristics of cloud, h is cloud layer thickness, and T is a sunlight scale factor which represents the ratio of actual sunlight time to total daytime time.
Further, the system also includes a power generation amount prediction branch construction module to execute the following operation steps:
acquiring a pre-trained fully-connected neural network;
and training the fully-connected neural network by taking the historical solar radiation set as input data and the historical power generation set as output data to obtain the power generation prediction branch.
Further, the system also includes a second power generation amount prediction module to perform the following operation steps:
extracting a historical solar radiation intensity set and a historical cloud cover thickness set based on the historical solar quantity set;
Clustering and dividing the historical solar radiation intensity set and the historical cloud layer thickness set respectively to obtain N historical solar radiation intensity clusters, M historical cloud layer thickness clusters and corresponding historical solar radiation intensity clustersThe N historical solar radiation intensity clusters and the M historical cloud layer thickness clusters are provided with grade marks, and M, N is largeAn integer of 2 or less;
based on a neural network, constructing a second power generation amount prediction model, and adopting the N historical solar radiation intensity clusters, the M historical cloud layer thickness clusters and the corresponding historical solar radiation intensity clustersPerforming supervision training on the second power generation amount prediction model until a preset standard is met to obtain the second power generation amount prediction model;
and when the level difference between the first solar radiation intensity cluster and the first cloud layer thickness cluster is larger than a preset level difference, activating the second generated energy prediction model to predict generated energy.
Further, the system further comprises a line loss amount acquisition module to perform the following operation steps:
acquiring a power transmission line from a photovoltaic power station to a power utilization area;
generating a first power generation amount in a photovoltaic power station, and carrying out power transmission simulation according to the power transmission line to obtain a simulated power transmission amount;
and acquiring real-time power transmission quantity in a power utilization area, calculating the difference value between the real-time power transmission quantity and the analog power transmission quantity, and acquiring the line loss quantity.
Further, the system also comprises an early warning module for executing the following operation steps:
generating an electricity consumption early warning value according to the second predicted electricity consumption sequence;
the power consumption data of the power consumption area is monitored in real time, and a real-time power consumption peak value is obtained;
and when the real-time electricity consumption peak reaches the electricity consumption early warning value, generating early warning information and sending the early warning information to a user.
Further, the system also comprises a predicted power generation amount sequence correction module for executing the following operation steps:
carrying out surface dust monitoring on equipment of the photovoltaic power station through an optical sensor to acquire dust accumulation data;
carrying out power generation loss analysis according to the dust accumulation data to obtain a power generation loss index;
and correcting the predicted power generation amount sequence according to the power generation loss index.
The foregoing detailed description of the photovoltaic power station source network load bidirectional prediction method based on artificial intelligence will be clear to those skilled in the art, and the photovoltaic power station source network load bidirectional prediction system based on artificial intelligence in this embodiment is relatively simple for the device disclosed in the embodiment, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (6)
1. The photovoltaic power station source network load bidirectional prediction method based on artificial intelligence is characterized by comprising the following steps of:
acquiring climate information according to the geographic position of a photovoltaic power station, and constructing a sunlight quantity calculation branch according to the climate information, wherein the sunlight quantity calculation branch comprises sunlight radiation intensity and cloud layer thickness;
Acquiring historical power generation information of a photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
according to the historical power generation information, training to obtain a power generation amount prediction branch, and integrating the sunlight amount calculation branch and the power generation amount prediction branch to obtain a power generation amount prediction model;
performing real-time environment monitoring on the photovoltaic power station, acquiring real-time environment monitoring data, inputting the real-time environment monitoring data into the power generation amount prediction model for power generation amount prediction, and acquiring a predicted power generation amount sequence of a first preset period, wherein the real-time environment monitoring data comprises real-time solar radiation intensity and real-time cloud layer thickness;
acquiring historical electricity consumption information of an electricity consumption area, and carrying out electricity consumption trend analysis according to the historical electricity consumption information to obtain a first predicted electricity consumption sequence of a first preset period;
acquiring line loss, acquiring a first energy storage scheme, and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the first energy storage scheme and the predicted electricity generation sequence;
comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence to obtain an abnormal time node set, wherein the abnormal time node set is a time node with a deviation value of the first predicted electricity consumption and the second predicted electricity consumption larger than or equal to a preset deviation value;
Optimizing the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set, and carrying out energy storage management according to the optimized first energy storage scheme;
the expression of the sunlight quantity calculation branch is as follows:
；
wherein Q is the sunlight quantity,the solar radiation intensity is the solar radiation intensity without cloud, k is a constant related to the optical characteristics of the cloud, h is the thickness of the cloud layer, T is a solar scale factor, and the ratio of the actual solar time to the total daytime time is represented;
inputting the real-time environment monitoring data into the power generation amount prediction model to predict the power generation amount, and further comprising:
extracting a historical solar radiation intensity set and a historical cloud cover thickness set based on the historical solar quantity set;
clustering and dividing the historical solar radiation intensity set and the historical cloud layer thickness set respectively to obtain N historical solar radiation intensity clusters, M historical cloud layer thickness clusters and corresponding historical solar radiation intensity clustersThe N historical solar radiation intensity clusters and the M historical cloud layer thickness clusters are provided with grade marks, and M, N is an integer greater than or equal to 2;
based on a neural network, constructing a second power generation amount prediction model, and adopting the N historical solar radiation intensity clusters, the M historical cloud layer thickness clusters and the corresponding historical solar radiation intensity clusters Performing supervision training on the second power generation amount prediction model until a preset standard is met to obtain the second power generation amount prediction model;
and when the level difference between the first solar radiation intensity cluster and the first cloud layer thickness cluster is larger than a preset level difference, activating the second generated energy prediction model to predict generated energy.
2. The method of claim 1, wherein training a derived power generation prediction branch based on the historical power generation information comprises:
acquiring a pre-trained fully-connected neural network;
and training the fully-connected neural network by taking the historical solar radiation set as input data and the historical power generation set as output data to obtain the power generation prediction branch.
3. The method of claim 1, wherein obtaining the line loss amount comprises:
acquiring a power transmission line from a photovoltaic power station to a power utilization area;
generating a first power generation amount in a photovoltaic power station, and carrying out power transmission simulation according to the power transmission line to obtain a simulated power transmission amount;
And acquiring real-time power transmission quantity in a power utilization area, calculating the difference value between the real-time power transmission quantity and the analog power transmission quantity, and acquiring the line loss quantity.
4. The method as recited in claim 1, further comprising:
generating an electricity consumption early warning value according to the second predicted electricity consumption sequence;
the power consumption data of the power consumption area is monitored in real time, and a real-time power consumption peak value is obtained;
and when the real-time electricity consumption peak reaches the electricity consumption early warning value, generating early warning information and sending the early warning information to a user.
5. The method as recited in claim 1, further comprising:
carrying out surface dust monitoring on equipment of the photovoltaic power station through an optical sensor to acquire dust accumulation data;
carrying out power generation loss analysis according to the dust accumulation data to obtain a power generation loss index;
and correcting the predicted power generation amount sequence according to the power generation loss index.
6. An artificial intelligence based photovoltaic power station source network load bidirectional prediction system, which is characterized by being used for implementing the artificial intelligence based photovoltaic power station source network load bidirectional prediction method as claimed in any one of claims 1-5, and comprising:
the solar radiation system comprises a first branch construction module, a second branch construction module and a third branch construction module, wherein the first branch construction module is used for acquiring climate information according to the geographic position of a photovoltaic power station and constructing a solar radiation calculation branch according to the climate information, and the solar radiation calculation branch comprises solar radiation intensity and cloud layer thickness;
The power generation information acquisition module is used for acquiring historical power generation information of the photovoltaic power station, wherein the historical power generation information is a mapping data set of a historical solar radiation set and a historical power generation set;
the second branch construction module is used for training to obtain a power generation amount prediction branch according to the historical power generation information, integrating the solar radiation amount calculation branch and the power generation amount prediction branch and obtaining a power generation amount prediction model;
the power generation amount prediction module is used for carrying out real-time environment monitoring on the photovoltaic power station, acquiring real-time environment monitoring data, inputting the real-time environment monitoring data into the power generation amount prediction model for carrying out power generation amount prediction, and acquiring a predicted power generation amount sequence of a first preset period, wherein the real-time environment monitoring data comprises real-time solar radiation intensity and real-time cloud layer thickness;
the power consumption trend analysis module is used for obtaining historical power consumption information of a power consumption area, carrying out power consumption trend analysis according to the historical power consumption information and obtaining a first predicted power consumption sequence of a first preset period;
The predicted electricity consumption acquisition module is used for acquiring line loss, acquiring a first energy storage scheme and acquiring a second predicted electricity consumption sequence of a first preset period by combining the line loss, the first energy storage scheme and the predicted electricity generation sequence;
the abnormal node acquisition module is used for comparing the first predicted electricity consumption sequence with the second predicted electricity consumption sequence to acquire an abnormal time node set, wherein the abnormal time node set is a time node with a deviation value of the first predicted electricity consumption and the second predicted electricity consumption being greater than or equal to a preset deviation value;
the energy storage management module is used for optimizing the first energy storage scheme according to the abnormal time node set and the corresponding deviation value set and carrying out energy storage management according to the optimized first energy storage scheme.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410071689.0A CN117595384B (en) | 2024-01-18 | 2024-01-18 | Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410071689.0A CN117595384B (en) | 2024-01-18 | 2024-01-18 | Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117595384A CN117595384A (en) | 2024-02-23 |
| CN117595384B true CN117595384B (en) | 2024-03-26 |
Family
ID=89911911
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410071689.0A Active CN117595384B (en) | 2024-01-18 | 2024-01-18 | Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117595384B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106909985A (en) * | 2017-01-11 | 2017-06-30 | 沃太能源南通有限公司 | A kind of photovoltaic generation forecasting system and Forecasting Methodology |
| CN115145983A (en) * | 2022-08-04 | 2022-10-04 | 陕西理工大学 | Method for determining daily generated energy of photovoltaic power generation system of new energy user |
| CN116111656A (en) * | 2022-12-28 | 2023-05-12 | 杭州中恒电气股份有限公司 | Micro-grid dispatching method and device |
| CN116436077A (en) * | 2023-04-04 | 2023-07-14 | 重庆跃达新能源有限公司 | Flexible power supply system and method based on photovoltaic energy storage |
| CN117293882A (en) * | 2023-04-28 | 2023-12-26 | 重庆跃达新能源有限公司 | Light energy storage load prediction system and method |
| CN117318111A (en) * | 2023-11-29 | 2023-12-29 | 南通沃太新能源有限公司 | Weather prediction-based dynamic adjustment method and system for light energy storage source |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113128793A (en) * | 2021-05-19 | 2021-07-16 | 中国南方电网有限责任公司 | Photovoltaic power combination prediction method and system based on multi-source data fusion |
-
2024
- 2024-01-18 CN CN202410071689.0A patent/CN117595384B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106909985A (en) * | 2017-01-11 | 2017-06-30 | 沃太能源南通有限公司 | A kind of photovoltaic generation forecasting system and Forecasting Methodology |
| CN115145983A (en) * | 2022-08-04 | 2022-10-04 | 陕西理工大学 | Method for determining daily generated energy of photovoltaic power generation system of new energy user |
| CN116111656A (en) * | 2022-12-28 | 2023-05-12 | 杭州中恒电气股份有限公司 | Micro-grid dispatching method and device |
| CN116436077A (en) * | 2023-04-04 | 2023-07-14 | 重庆跃达新能源有限公司 | Flexible power supply system and method based on photovoltaic energy storage |
| CN117293882A (en) * | 2023-04-28 | 2023-12-26 | 重庆跃达新能源有限公司 | Light energy storage load prediction system and method |
| CN117318111A (en) * | 2023-11-29 | 2023-12-29 | 南通沃太新能源有限公司 | Weather prediction-based dynamic adjustment method and system for light energy storage source |
Non-Patent Citations (2)
| Title |
|---|
| Evaluating the Weather Forecasting Models and the Impact to PV Generation Forecasting;Spyros Theocharides等;《2023 IEEE 50th Photovoltaic Specialists Conference (PVSC)》;20230616;1-5 * |
| 基于改进DDPG算法的中短期光伏发电功率预测;苏诗慧;雷勇;李永凯;朱英伟;;半导体光电;20201015(05);116-122 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117595384A (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bae et al. | Hourly solar irradiance prediction based on support vector machine and its error analysis | |
| Liu et al. | An improved photovoltaic power forecasting model with the assistance of aerosol index data | |
| Yang et al. | Classification and summarization of solar irradiance and power forecasting methods: A thorough review | |
| CN113496311A (en) | Photovoltaic power station generated power prediction method and system | |
| de Guia et al. | Using stacked long short term memory with principal component analysis for short term prediction of solar irradiance based on weather patterns | |
| Perveen et al. | An ANFIS‐based model for solar energy forecasting and its smart grid application | |
| CN117318111B (en) | Weather prediction-based dynamic adjustment method and system for light energy storage source | |
| Mo et al. | Impact of aging and performance degradation on the operational costs of distributed generation systems | |
| Javed et al. | Predicting solar irradiance using machine learning techniques | |
| Kayri et al. | Predicting power production from a photovoltaic panel through artificial neural networks using atmospheric indicators | |
| Mayer et al. | Probabilistic modeling of future electricity systems with high renewable energy penetration using machine learning | |
| Sadeghi et al. | Improving PV power plant forecast accuracy: A hybrid deep learning approach compared across short, medium, and long-term horizons | |
| Ehsan et al. | Artificial neural network predictor for grid-connected solar photovoltaic installations at atmospheric temperature | |
| CN117595384B (en) | Photovoltaic power station source network load bidirectional prediction method and system based on artificial intelligence | |
| Panjwani et al. | Short-term solar and wind generation forecasting for the western region of india | |
| Ioakimidis et al. | Short-term wind speed forecasting model based on ANN with statistical feature parameters | |
| Zhou et al. | Ultra-short term hybrid power forecasting model for photovoltaic power station with meteorological monitoring data | |
| Onaolapo | Reliability study under the smart grid paradigm using computational intelligent techniques and renewable energy sources. | |
| Zhou et al. | PV power characteristic modeling based on multi-scale clustering and its application in generation prediction | |
| Cho et al. | Application of Parallel ANN-PSO to Hourly Solar PV Estimation | |
| Gil-Vera et al. | Predictive modeling of photovoltaic solar power generation | |
| Rao et al. | Bright sunshine duration index-based prediction of solar PV power using ANN approach | |
| Gersnoviez et al. | Neuro-fuzzy systems for daily solar irradiance classification and PV efficiency forecasting | |
| Zeljković et al. | Probabilistic Techno-Economic Assessment of Wind-PV-Diesel-Battery Power Supplies for Mobile Telephony Base Stations | |
| Alanqar et al. | Solar Power Monitoring and Forecasting System using Artificial Intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |