CN110705760A - Photovoltaic power generation power prediction method based on deep belief network - Google Patents
Photovoltaic power generation power prediction method based on deep belief network Download PDFInfo
- Publication number
- CN110705760A CN110705760A CN201910887578.6A CN201910887578A CN110705760A CN 110705760 A CN110705760 A CN 110705760A CN 201910887578 A CN201910887578 A CN 201910887578A CN 110705760 A CN110705760 A CN 110705760A
- Authority
- CN
- China
- Prior art keywords
- power generation
- photovoltaic power
- prediction model
- meteorological
- generation power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010248 power generation Methods 0.000 title claims abstract description 80
- 238000000034 method Methods 0.000 title claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims abstract description 48
- 238000012549 training Methods 0.000 claims abstract description 28
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 239000002245 particle Substances 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000004364 calculation method Methods 0.000 claims description 16
- 239000000126 substance Substances 0.000 claims description 10
- 238000005457 optimization Methods 0.000 claims description 8
- 230000004913 activation Effects 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 230000007613 environmental effect Effects 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000001556 precipitation Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 description 3
- 238000000053 physical method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000002688 persistence Effects 0.000 description 1
- 238000005293 physical law Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
Classifications
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a photovoltaic power generation power prediction method based on a deep belief network, which comprises the following steps of: acquiring historical meteorological data and photovoltaic power generation output data of an area to be predicted to construct a training data set; dividing the historical meteorological data into a plurality of samples, preprocessing the samples and constructing a meteorological factor data matrix; clustering the meteorological factor data matrix by adopting a fuzzy C-means clustering algorithm; establishing a photovoltaic power generation power prediction model by using a depth confidence network; optimizing the initial weight of the prediction model by adopting a particle swarm algorithm; inputting historical meteorological data in the training data set into the prediction model for training to obtain the optimal weight of each layer of the prediction model and storing the optimal weight; and judging the meteorological factor category of the day to be predicted by adopting a fuzzy C-means clustering algorithm, and inputting all samples in the meteorological factor data matrix, which belong to the same category as the day to be predicted, into the prediction model to obtain the predicted photovoltaic power generation power.
Description
Technical Field
The invention relates to the technical field of power systems and automation thereof, in particular to a photovoltaic power generation power prediction method based on a deep belief network.
Background
Due to uncertain factors such as randomness, fluctuation and intermittence of photovoltaic power generation, a large-scale photovoltaic power plant brings a series of problems such as voltage and frequency deviation, voltage fluctuation and grid disconnection to the safe and stable operation of a power system when being merged into a power grid. Therefore, accurate photovoltaic power generation power prediction is beneficial to large-scale photovoltaic power generation grid connection, and harm brought by photovoltaic power generation grid connection is reduced.
The existing photovoltaic power generation power prediction method comprises a physical method and a statistical method. The physical method comprises the steps of establishing a fluid mechanics and thermodynamic equation set for describing a weather evolution process according to an inherent physical law, gradually solving meteorological elements such as wind speed, wind direction, air pressure and temperature, and establishing a relevant model to obtain a predicted value of the photovoltaic power generation power; the statistical method is to establish a mapping relation between historical data and predicted values and substitute measured values for prediction. Common methods in the physical method for predicting the photovoltaic power generation power include a persistence method, a time series method, a neural network method, Kalman filtering and spatial correlation. At present, a main neural network method is used for predicting photovoltaic power generation power, but the neural network structure is relatively simple, and along with randomness and instability of meteorological factors, the defect of low prediction accuracy exists, for example, in a deep confidence network, the initial weight of the network is randomly determined, so that the situation that the prediction accuracy of photovoltaic power generation prediction is influenced due to the fact that the initial weight of the network easily falls into a local optimal solution exists.
Disclosure of Invention
The invention provides a photovoltaic power generation power prediction method based on a deep belief network, aiming at overcoming the defect of low prediction precision in the prior art.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a photovoltaic power generation power prediction method based on a deep belief network comprises the following steps:
s1: acquiring historical meteorological data and photovoltaic power generation output data of an area to be predicted to construct a training data set;
s2: dividing the historical meteorological data into a plurality of samples by taking a natural day as a unit, preprocessing the samples and constructing a meteorological factor data matrix B;
s3: clustering the meteorological factor data matrix B by adopting a fuzzy C-means clustering algorithm to obtain the membership degree of each sample in the meteorological factor data matrix B to all class centers;
s4: establishing a photovoltaic power generation power prediction model by using a depth confidence network;
s5: optimizing the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm;
s6: inputting historical meteorological data in the training data set into the prediction model for training to obtain the optimal weight of each layer of the prediction model and storing the optimal weight;
s7: and judging the meteorological factor category of the day to be predicted by adopting a fuzzy C-means clustering algorithm, and inputting all samples, which are in the same category as the day to be predicted, in the clustered meteorological factor data matrix B into the photovoltaic power generation power prediction model in the step S6 to obtain the predicted photovoltaic power generation power.
In the technical scheme, randomness and instability of photovoltaic power generation output are considered, historical output data recorded in a photovoltaic power generation place and meteorological factor data responded are collected for analysis, specifically, a fuzzy C-means clustering algorithm is adopted to cluster the preprocessed meteorological data, new membership degrees of each sample point to all mines are obtained, and therefore the category of the sample points is determined so as to achieve the purpose of automatically classifying sample data; aiming at the condition that the initial weight of the deep belief network is easy to have a local optimal solution, optimizing the initial weight of a photovoltaic power generation power prediction model established by the deep belief network by adopting a particle swarm algorithm to obtain the optimal weight, then training each clustered sample category through the prediction model optimized by the particle swarm to obtain the optimal weight, and obtaining the photovoltaic power generation power prediction model with the optimal weight. In practical application, the meteorological factor category of a day to be predicted is judged through a fuzzy C-means clustering algorithm, and then all samples in the meteorological factor data matrix B, which are the same as the category of the day to be predicted, are input into a photovoltaic power generation power prediction model with optimal weight to predict the photovoltaic power generation power of the day to be predicted.
Preferably, in step S2, the training data set is preprocessed by mean interpolation, which has the following formula:
wherein the content of the first and second substances,denotes the normalized value, xiRepresenting original meteorological data, wherein i represents ith meteorological data; x is the number ofminRepresenting the minimum, x, in the sequence of raw meteorological data in the same samplemaxRepresenting the maximum in the sequence of raw meteorological data in the same sample.
Preferably, in step S2, the meteorological factor data matrix B is:
where m represents the sampling time in the sample, n represents a meteorological data factor, bmnRepresenting meteorological data of the m moments in the sample under the condition of n meteorological factors; the meteorological data factors include, but are not limited to, cloud cover, temperature, humidity, wind speed, precipitation.
Preferably, the specific steps of the step S3 are as follows:
s31: confirming a fuzzy C mean value cluster center initialization value for clustering according to the meteorological factor data matrix B, and constructing a cluster center value matrix, wherein the calculation formula is as follows:
wherein, CxyA cluster center value of a y-th class environmental factor attribute representing an x-th class environmental factor; r is the neighborhood radius, c is the number of categories; n is in the neighborhood radius rangeThe number of samples of (a);
s32: initializing a membership matrix U, and normalizing elements in the membership matrix U to ensure that the elements of the membership matrix U are all in an interval [0,1 ];
s33: and iteratively updating the membership matrix U, wherein an iterative formula is as follows:
wherein u isijThe element of the membership degree matrix U represents the degree to which the ith object belongs to the jth class, wherein i is 1, 2.
S34: and iteratively updating the clustering center value matrix, wherein the iterative formula is as follows:
wherein K is a clustering center value matrix coefficient, and alpha represents a coefficient weight; xiRepresenting a single sample;
s35: judging whether the requirements are metOr reaching the preset iteration times, if so, ending the iteration and outputting the optimal membership degree, otherwise, skipping to execute the step S33 to carry out the next iteration; wherein epsilon represents a minimum error value, which is a minimum value close to 0; t represents the number of iterations.
Preferably, the specific steps of the step S4 are as follows:
s41: and (3) expressing the state E by using a restricted Boltzmann machine, wherein the expression formula is as follows:
θ=(γij,ai,bj)
wherein v is visible layer unit weight, h is hidden layer unit weight, and theta is limited Boltzmann machine parameter; n and m represent the number of neurons in the visible layer unit and the hidden layer unit, respectively;
s42: determining a limited Boltzmann machine probability distribution p (v, h θ) for state E, which is formulated as follows:
s43: respectively calculating the activation probability p (h) of the jth hidden layer unit of each layer of restricted Boltzmann machinej| v, θ) and activation probability p (v) of the ith visible layer uniti| h, θ), the calculation formula is as follows:
s44: training the states of the explicit element and the implicit element according to the contrast divergence calculation criterion, the parameter variation calculation criterion and the parameter updating criterion of the learning rate, then repeatedly executing the steps S41-S44 until the visible layer unit weight and the implicit layer unit weight are converged, finishing the training of the multi-layer Boltzmann machine firstly, and finishing the establishment of the photovoltaic power generation prediction model.
Preferably, the specific steps of the step S5 are as follows:
s51: calculating the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the calculation formula is as follows:
wherein, omega represents the initial weight of the photovoltaic power generation power prediction model, and omegaidRepresenting initial weights of the optimized photovoltaic power generation power prediction model; n isiIs a fixed value determined by the number of samples;
s52: optimizing the weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the optimization formula is as follows:
wherein the content of the first and second substances,represents the updated photovoltaic power generation power prediction model weight value,representing an initial weight value of the photovoltaic power generation power prediction model;representing the updated model weight velocity values,representing a model initial weight velocity value; x is the number ofdAs an initial position, pbestidIndicating the optimal position of a single particle, gbestidRepresents the optimal position of the whole population of particles, and M and ψ represent model weight velocity value coefficients.
Preferably, the specific steps of the step S6 include: and preprocessing historical meteorological data in the training data set by adopting an average interpolation method, inputting the preprocessed historical meteorological data into the photovoltaic power generation power prediction model which completes the initial weight optimization, and obtaining the optimal solution of each layer of the photovoltaic power generation power prediction model through training of the limited Boltzmann machine and adjustment of back propagation of the BP neural network.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that: the weather factor category of the day to be predicted is judged based on the fuzzy C-means clustering, so that the randomness and the instability of weather factor data can be effectively solved; the photovoltaic power generation power prediction model with the optimal weight is constructed by utilizing the particle swarm optimization deep belief network, and the problem that the initial weight of the conventional deep belief network prediction model is easy to fall into a local optimal solution can be solved.
Drawings
Fig. 1 is a flowchart of a photovoltaic power generation power prediction method based on a deep belief network according to this embodiment.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Fig. 1 is a flowchart of a photovoltaic power generation power prediction method based on a deep belief network according to this embodiment.
The embodiment provides a photovoltaic power generation power prediction method based on a deep belief network, which comprises the following steps:
s1: historical meteorological data and photovoltaic power generation output data of an area to be predicted are collected to construct a training data set.
In the embodiment, meteorological data and historical output data recorded by monitoring a photovoltaic power plant in a certain area are collected, data of a certain year is used as a training set, and a day of each month in the next year is randomly selected as a day to be predicted.
S2: and dividing the historical meteorological data into a plurality of samples by taking the natural day as a unit, preprocessing the samples and constructing a meteorological factor data matrix B.
In this embodiment, the historical meteorological data is divided into a plurality of samples in a unit of natural day, and meanwhile, in consideration of no illumination at night, the time interval of 06:00-19:00 per day is selected as 1 hour as sampling time, that is, each sample comprises 19 meteorological data.
In this embodiment, to ensure that the meteorological data can be clustered, the gas phase data is preprocessed by using an average interpolation method, which has the following formula:
wherein the content of the first and second substances,denotes the normalized value, xiRepresenting original meteorological data, wherein i represents ith meteorological data; x is the number ofminRepresenting the minimum, x, in the sequence of raw meteorological data in the same samplemaxRepresenting the maximum value in the sequence of raw meteorological data in the same sample; the meteorological factor data matrix B constructed by the method is as follows:
wherein m represents the sampling time in the sample, n represents meteorological data factors, and n is 1,2,3,4,5, corresponding to cloud cover, temperature, humidity, wind speed, precipitation respectively; bmnRepresenting meteorological data at time m in the sample under conditions of meteorological factors n.
S3: clustering the meteorological factor data matrix B by adopting a fuzzy C-means clustering algorithm to obtain the membership degree of each sample in the meteorological factor data matrix B to all class centers; the method comprises the following specific steps:
s31: confirming a fuzzy C mean value cluster center initialization value for clustering according to the meteorological factor data matrix B, and constructing a cluster center value matrix, wherein the calculation formula is as follows:
wherein, CxyA cluster center value of a y-th class environmental factor attribute representing an x-th class environmental factor; r is the neighborhood radius, c is the number of categories; n is the number of samples in the neighborhood radius range;
s32: initializing a membership matrix U, and normalizing elements in the membership matrix U to enable the elements of the membership matrix U to be in an interval [0,1 ];
s33: and iteratively updating the membership matrix U, wherein the iterative formula is as follows:
wherein u isijThe element of the membership degree matrix U represents the degree to which the ith object belongs to the jth class, wherein i is 1, 2.
S34: and iteratively updating the clustering center value matrix, wherein the iterative formula is as follows:
wherein K is a clustering center value matrix coefficient, and alpha represents the weight of the clustering center value matrix coefficient; xiRepresenting a single sample;
s35: judging whether the requirements are metOr reaching the preset iteration times, if so, ending the iteration and outputting the optimal membership degree, otherwise, skipping to execute the step S33 to carry out the next iteration; it is composed ofIn (e), ε represents the minimum error value, and ε is the minimum value close to 0; t represents the number of iterations.
S4: and establishing a photovoltaic power generation power prediction model by utilizing the depth confidence network.
In the embodiment, a deep confidence network is used to establish a prediction model of the corresponding component, wherein the deep confidence network is a neural network model stacked by a plurality of limited boltzmann machines. The method comprises the following specific steps:
s41: and (3) expressing the state E by using a restricted Boltzmann machine, wherein the expression formula is as follows:
θ=(γij,ai,bj)
wherein v is visible layer unit weight, h is hidden layer unit weight, and theta is limited Boltzmann machine parameter; n and m represent the number of neurons in the visible layer unit and the hidden layer unit, respectively;
s42: determining a finite boltzmann machine probability distribution p (v, h | θ) for state E, which is formulated as follows:
s43: respectively calculating the activation probability p (h) of the jth hidden layer unit of each layer of restricted Boltzmann machinej| v, θ) and activation probability p (v) of the ith visible layer uniti| h, θ), the calculation formula is as follows:
s44: training the states of the explicit element and the implicit element according to the contrast divergence calculation criterion, the parameter variation calculation criterion and the parameter updating criterion of the learning rate, then repeatedly executing the steps S41-S44 until the visible layer unit weight and the implicit layer unit weight are converged, finishing the training of the multi-layer Boltzmann machine firstly, and finishing the establishment of the photovoltaic power generation prediction model.
Wherein, the formula of the updating criterion of the contrast divergence is as follows:
wherein, < vihj>dataFor the expectation of data distribution, < vihj>recon(ii) a desire defined for a constrained boltzmann machine;
the formula of each parameter variation calculation criterion is as follows:
wherein, Δ wij、Δai、ΔbjRespectively representing limited Boltzmann machine component parameters;
the formula of the parameter update criterion of the learning rate ξ is as follows:
wherein the content of the first and second substances,respectively, the restricted boltzmann machine component learning rates.
S5: optimizing the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm; the method comprises the following specific steps:
s51: calculating the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the calculation formula is as follows:
wherein, omega represents the initial weight of the photovoltaic power generation power prediction model, and omegaidRepresenting initial weights of the optimized photovoltaic power generation power prediction model; n isiIs a fixed value determined by the number of samples;
s52: optimizing the weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the optimization formula is as follows:
wherein the content of the first and second substances,represents the updated photovoltaic power generation power prediction model weight value,representing an initial weight value of the photovoltaic power generation power prediction model;representing the updated model weight velocity values,representing a model initial weight velocity value; x is the number ofdAs an initial position, pbestidIndicating the optimal position of a single particle, gbestidRepresents the optimal position of the whole population of particles, and M and ψ represent model weight velocity value coefficients.
In this embodiment, considering that the initial weight of the deep belief network is randomly determined and a situation that the deep belief network is likely to fall into a locally optimal solution exists, the present embodiment uses a particle swarm algorithm to find a globally optimal solution by following a currently searched optimal solution, has the advantages of easy implementation, high precision and fast convergence, and can effectively solve the defect that the deep belief network is likely to fall into the locally optimal solution.
S6: and inputting historical meteorological data in the training data set into the prediction model for training to obtain the optimal weight of each layer of the prediction model and storing the optimal weight.
In the step, historical meteorological data in the training data set are preprocessed by an average interpolation method, then the preprocessed historical meteorological data are input into the photovoltaic power generation power prediction model which completes initial weight optimization, and the optimal weight of each layer of the photovoltaic power generation power prediction model is obtained and stored through training of a limited Boltzmann machine and adjustment of back propagation of a BP neural network, so that the photovoltaic power generation power prediction model with the optimal weight is obtained.
S7: and judging the meteorological factor category of the day to be predicted by adopting a fuzzy C-means clustering algorithm, and inputting all samples, which are the same as the category of the day to be predicted, in the clustered meteorological factor data matrix B into the photovoltaic power generation power prediction model in the step S6 to obtain the predicted photovoltaic power generation power.
In the embodiment, the meteorological factor category of the day to be predicted is judged based on the fuzzy C-means clustering, the randomness and the instability of meteorological factor data can be effectively solved, the particle swarm optimization deep belief network is used for constructing the photovoltaic power generation power prediction model with the optimal weight, the problem that the initial weight of the existing deep belief network prediction model is easy to fall into a local optimal solution can be solved, the prediction precision of the model can be effectively improved, and the method has important significance for improving the stability of a power grid and the operation management efficiency of a photovoltaic power generation field.
It should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (7)
1. A photovoltaic power generation power prediction method based on a deep belief network is characterized by comprising the following steps:
s1: acquiring historical meteorological data and photovoltaic power generation output data of an area to be predicted to construct a training data set;
s2: dividing the historical meteorological data into a plurality of samples by taking a natural day as a unit, preprocessing the samples and constructing a meteorological factor data matrix B;
s3: clustering the meteorological factor data matrix B by adopting a fuzzy C-means clustering algorithm to obtain the membership degree of each sample in the meteorological factor data matrix B to all class centers;
s4: establishing a photovoltaic power generation power prediction model by using a depth confidence network;
s5: optimizing the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm;
s6: inputting historical meteorological data in the training data set into the prediction model for training to obtain the optimal weight of each layer of the prediction model and storing the optimal weight;
s7: and judging the meteorological factor category of the day to be predicted by adopting a fuzzy C-means clustering algorithm, and inputting all samples, which are the same as the category of the day to be predicted, in the clustered meteorological factor data matrix B into the photovoltaic power generation power prediction model in the step S6 to obtain the predicted photovoltaic power generation power.
2. The photovoltaic power generation power prediction method according to claim 1, wherein in the step S2, the training data set is preprocessed by an average interpolation method, which is expressed by the following formula:
wherein the content of the first and second substances,denotes the normalized value, xiRepresenting original meteorological data, wherein i represents ith meteorological data; x is the number ofminRepresenting the minimum, x, in the sequence of raw meteorological data in the same samplemaxRepresenting the maximum in the sequence of raw meteorological data in the same sample.
3. The photovoltaic power generation power prediction method according to claim 2, wherein in the step S2, the meteorological factor data matrix B is:
where m represents the sampling time in the sample, n represents a meteorological data factor, bmnRepresenting meteorological data of the m moments in the sample under the condition of n meteorological factors; the meteorological data factors include, but are not limited to, cloud cover, temperature, humidity, wind speed, precipitation.
4. The photovoltaic power generation power prediction method according to claim 3, wherein the specific steps of the step S3 are as follows:
s31: confirming a fuzzy C mean value cluster center initialization value for clustering according to the meteorological factor data matrix B, and constructing a cluster center value matrix, wherein the calculation formula is as follows:
wherein, CxyRepresenting class x environmentsCluster center value of the y-th class environmental factor attribute of the factor; r is the neighborhood radius, c is the number of categories; n is the number of samples in the neighborhood radius range;
s32: initializing a membership matrix U, and normalizing elements in the membership matrix U to enable the elements of the membership matrix U to be in an interval [0,1 ];
s33: and iteratively updating the membership matrix U, wherein the iterative formula is as follows:
wherein u isijThe element of the membership degree matrix U represents the degree to which the ith object belongs to the jth class, wherein i is 1, 2.
S34: and iteratively updating the clustering center value matrix, wherein the iterative formula is as follows:
wherein K is a clustering center value matrix coefficient, and alpha represents the weight of the clustering center value matrix coefficient; xiRepresenting a single sample;
s35: judging whether the requirements are metOr reaching the preset iteration times, if so, ending the iteration and outputting the optimal membership degree, otherwise, skipping to execute the step S33 to carry out the next iteration; where ε represents the minimum error value and t represents the number of iterations.
5. The photovoltaic power generation power prediction method according to claim 4, wherein the specific steps of the step S4 are as follows:
s41: and (3) expressing the state E by using a restricted Boltzmann machine, wherein the expression formula is as follows:
θ=(γij,ai,bj)
wherein v is visible layer unit weight, h is hidden layer unit weight, and theta is limited Boltzmann machine parameter; n and m represent the number of neurons in the visible layer unit and the hidden layer unit, respectively;
s42: determining a finite boltzmann machine probability distribution p (v, h | θ) for state E, which is formulated as follows:
s43: respectively calculating the activation probability p (h) of the jth hidden layer unit of each layer of restricted Boltzmann machinej| v, θ) and activation probability p (v) of the ith visible layer uniti| h, θ), the calculation formula is as follows:
s44: training the states of the explicit element and the implicit element according to the contrast divergence calculation criterion, the parameter variation calculation criterion and the parameter updating criterion of the learning rate, then repeatedly executing the steps S41-S44 until the visible layer unit weight and the implicit layer unit weight are converged, finishing the training of the multi-layer Boltzmann machine firstly, and finishing the establishment of the photovoltaic power generation prediction model.
6. The photovoltaic power generation power prediction method according to claim 3, wherein the specific steps of the step S5 are as follows:
s51: calculating the initial weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the calculation formula is as follows:
wherein, omega represents the initial weight of the photovoltaic power generation power prediction model, and omegaidRepresenting initial weights of the optimized photovoltaic power generation power prediction model; n isiIs a fixed value determined by the number of samples;
s52: optimizing the weight of the photovoltaic power generation power prediction model by adopting a particle swarm algorithm, wherein the optimization formula is as follows:
wherein the content of the first and second substances,represents the updated photovoltaic power generation power prediction model weight value,representing an initial weight value of the photovoltaic power generation power prediction model;representing the updated model weight velocity values,representing a model initial weight velocity value; x is the number ofdAs an initial position, pbestidIndicating the optimal position of a single particle, gbestidRepresents the optimal position of the whole population of particles, and M and ψ represent model weight velocity value coefficients.
7. The photovoltaic power generation power prediction method according to claim 3, wherein the specific step of the step of S6 includes: and preprocessing historical meteorological data in the training data set by adopting an average interpolation method, inputting the preprocessed historical meteorological data into the photovoltaic power generation power prediction model which completes the initial weight optimization, and obtaining the optimal solution of each layer of the photovoltaic power generation power prediction model through training of the limited Boltzmann machine and adjustment of back propagation of the BP neural network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910887578.6A CN110705760A (en) | 2019-09-19 | 2019-09-19 | Photovoltaic power generation power prediction method based on deep belief network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910887578.6A CN110705760A (en) | 2019-09-19 | 2019-09-19 | Photovoltaic power generation power prediction method based on deep belief network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110705760A true CN110705760A (en) | 2020-01-17 |
Family
ID=69195652
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910887578.6A Pending CN110705760A (en) | 2019-09-19 | 2019-09-19 | Photovoltaic power generation power prediction method based on deep belief network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110705760A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111695666A (en) * | 2020-05-26 | 2020-09-22 | 河海大学 | Wind power ultra-short term conditional probability prediction method based on deep learning |
CN111695601A (en) * | 2020-05-15 | 2020-09-22 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
CN111881565A (en) * | 2020-07-17 | 2020-11-03 | 特变电工新疆新能源股份有限公司 | Photovoltaic module process data reconstruction method |
CN112633572A (en) * | 2020-12-21 | 2021-04-09 | 国网甘肃省电力公司电力科学研究院 | Day classification dual-model photovoltaic power generation combined prediction method based on prediction meteorological data |
CN112862216A (en) * | 2021-03-10 | 2021-05-28 | 广东电网有限责任公司 | Multi-industry energy demand prediction method based on deep belief network |
CN112906979A (en) * | 2021-03-19 | 2021-06-04 | 上海电机学院 | Power generation power prediction method |
CN113323821A (en) * | 2021-06-11 | 2021-08-31 | 中南大学 | Method for adjusting yaw control parameters of wind turbine model prediction |
CN116070798A (en) * | 2023-03-30 | 2023-05-05 | 国网山东省电力公司滨州供电公司 | Photovoltaic power generation power short-term prediction method and system |
CN116128150A (en) * | 2023-03-02 | 2023-05-16 | 南京邮电大学 | Photovoltaic intelligent output prediction method based on two-stage optimization |
CN116365519A (en) * | 2023-06-01 | 2023-06-30 | 国网山东省电力公司微山县供电公司 | Power load prediction method, system, storage medium and equipment |
CN116706907A (en) * | 2023-08-09 | 2023-09-05 | 深圳航天科创泛在电气有限公司 | Photovoltaic power generation prediction method based on fuzzy reasoning and related equipment |
WO2023201552A1 (en) * | 2022-04-20 | 2023-10-26 | Southeast University | County-wide photovoltaic prediction method based on cluster division and data enhancement |
CN117239745A (en) * | 2023-11-16 | 2023-12-15 | 北京弘象科技有限公司 | Photovoltaic power generation amount prediction method and device, electronic equipment and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105913151A (en) * | 2016-04-12 | 2016-08-31 | 河海大学常州校区 | Photovoltaic power station power generation amount predication method based on adaptive mutation particle swarm and BP network |
US20170161606A1 (en) * | 2015-12-06 | 2017-06-08 | Beijing University Of Technology | Clustering method based on iterations of neural networks |
CN108985323A (en) * | 2018-06-04 | 2018-12-11 | 江苏大学 | A kind of short term prediction method of photovoltaic power |
CN109636003A (en) * | 2018-11-14 | 2019-04-16 | 北京电力经济技术研究院有限公司 | A kind of power-system short-term load forecasting method of high precision of prediction |
CN109902743A (en) * | 2019-02-28 | 2019-06-18 | 广东工业大学 | A kind of Wind turbines output power predicting method |
CN110059871A (en) * | 2019-04-04 | 2019-07-26 | 广东工业大学 | Photovoltaic power generation power prediction method |
-
2019
- 2019-09-19 CN CN201910887578.6A patent/CN110705760A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170161606A1 (en) * | 2015-12-06 | 2017-06-08 | Beijing University Of Technology | Clustering method based on iterations of neural networks |
CN105913151A (en) * | 2016-04-12 | 2016-08-31 | 河海大学常州校区 | Photovoltaic power station power generation amount predication method based on adaptive mutation particle swarm and BP network |
CN108985323A (en) * | 2018-06-04 | 2018-12-11 | 江苏大学 | A kind of short term prediction method of photovoltaic power |
CN109636003A (en) * | 2018-11-14 | 2019-04-16 | 北京电力经济技术研究院有限公司 | A kind of power-system short-term load forecasting method of high precision of prediction |
CN109902743A (en) * | 2019-02-28 | 2019-06-18 | 广东工业大学 | A kind of Wind turbines output power predicting method |
CN110059871A (en) * | 2019-04-04 | 2019-07-26 | 广东工业大学 | Photovoltaic power generation power prediction method |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111695601A (en) * | 2020-05-15 | 2020-09-22 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
CN111695601B (en) * | 2020-05-15 | 2023-06-20 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
CN111695666A (en) * | 2020-05-26 | 2020-09-22 | 河海大学 | Wind power ultra-short term conditional probability prediction method based on deep learning |
CN111881565A (en) * | 2020-07-17 | 2020-11-03 | 特变电工新疆新能源股份有限公司 | Photovoltaic module process data reconstruction method |
CN111881565B (en) * | 2020-07-17 | 2024-03-22 | 特变电工新疆新能源股份有限公司 | Photovoltaic module process data reconstruction method |
CN112633572A (en) * | 2020-12-21 | 2021-04-09 | 国网甘肃省电力公司电力科学研究院 | Day classification dual-model photovoltaic power generation combined prediction method based on prediction meteorological data |
CN112862216A (en) * | 2021-03-10 | 2021-05-28 | 广东电网有限责任公司 | Multi-industry energy demand prediction method based on deep belief network |
CN112906979A (en) * | 2021-03-19 | 2021-06-04 | 上海电机学院 | Power generation power prediction method |
CN113323821A (en) * | 2021-06-11 | 2021-08-31 | 中南大学 | Method for adjusting yaw control parameters of wind turbine model prediction |
WO2023201552A1 (en) * | 2022-04-20 | 2023-10-26 | Southeast University | County-wide photovoltaic prediction method based on cluster division and data enhancement |
CN116128150A (en) * | 2023-03-02 | 2023-05-16 | 南京邮电大学 | Photovoltaic intelligent output prediction method based on two-stage optimization |
CN116070798A (en) * | 2023-03-30 | 2023-05-05 | 国网山东省电力公司滨州供电公司 | Photovoltaic power generation power short-term prediction method and system |
CN116365519B (en) * | 2023-06-01 | 2023-09-26 | 国网山东省电力公司微山县供电公司 | Power load prediction method, system, storage medium and equipment |
CN116365519A (en) * | 2023-06-01 | 2023-06-30 | 国网山东省电力公司微山县供电公司 | Power load prediction method, system, storage medium and equipment |
CN116706907A (en) * | 2023-08-09 | 2023-09-05 | 深圳航天科创泛在电气有限公司 | Photovoltaic power generation prediction method based on fuzzy reasoning and related equipment |
CN116706907B (en) * | 2023-08-09 | 2024-01-23 | 深圳航天科创泛在电气有限公司 | Photovoltaic power generation prediction method based on fuzzy reasoning and related equipment |
CN117239745A (en) * | 2023-11-16 | 2023-12-15 | 北京弘象科技有限公司 | Photovoltaic power generation amount prediction method and device, electronic equipment and storage medium |
CN117239745B (en) * | 2023-11-16 | 2024-01-23 | 北京弘象科技有限公司 | Photovoltaic power generation amount prediction method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110705760A (en) | Photovoltaic power generation power prediction method based on deep belief network | |
CN113962364B (en) | Multi-factor power load prediction method based on deep learning | |
CN113128793A (en) | Photovoltaic power combination prediction method and system based on multi-source data fusion | |
JP5888640B2 (en) | Photovoltaic power generation prediction apparatus, solar power generation prediction method, and solar power generation prediction program | |
CN110503256A (en) | Short-term load forecasting method and system based on big data technology | |
CN109583635B (en) | Short-term load prediction modeling method for operational reliability | |
CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
CN105069521A (en) | Photovoltaic power plant output power prediction method based on weighted FCM clustering algorithm | |
CN110929953A (en) | Photovoltaic power station ultra-short term output prediction method based on cluster analysis | |
CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
CN114298140A (en) | Wind power short-term power prediction correction method considering unit classification | |
CN111598225A (en) | Air conditioner cold load prediction method based on adaptive deep confidence network | |
Li et al. | Forecasting of short-term photovoltaic power generation using combined interval type-2 Takagi-Sugeno-Kang fuzzy systems | |
CN115099511A (en) | Photovoltaic power probability estimation method and system based on optimized copula | |
CN116029419A (en) | Deep learning-based long-term new energy daily average generation power prediction method and system | |
CN116169670A (en) | Short-term non-resident load prediction method and system based on improved neural network | |
CN112288157A (en) | Wind power plant power prediction method based on fuzzy clustering and deep reinforcement learning | |
CN116451821A (en) | Deep learning-based large-area new energy generation power prediction method and system | |
CN113449920A (en) | Wind power prediction method, system and computer readable medium | |
Ye et al. | Optimal Component IGSCV-SVR Ensemble Model Improved by VMD for Ultra-short-term Wind Speed Forecasting. | |
CN110163437A (en) | Day-ahead photovoltaic power generation power prediction method based on DPK-means | |
CN115907228A (en) | Short-term power load prediction analysis method based on PSO-LSSVM | |
CN110059871A (en) | Photovoltaic power generation power prediction method | |
Li et al. | A short-term wind power forecasting method based on NWP wind speed fluctuation division and clustering | |
CN115345245A (en) | Tropical cyclone rapid reinforcement forecasting method based on random forest and transfer learning |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200117 |
|
RJ01 | Rejection of invention patent application after publication |