226 results about "Numerical weather forecast" patented technology

The invention discloses a physical prediction method for wind power station power based on a computational fluid mechanics model. The physical prediction method comprises the following steps: establishing a computational fluid mechanics model; performing discretization on wind conditions of a wind power station, and taking the discretized wind as boundary conditions for conditions numerical simulation of the computational fluid to obtain space flow filed distribution of the wind power station at the discretized wind conditions; establishing data base of hub height, wind speed, wind direction and generated power of wind generation sets under the discretized wind conditions; and taking numerical weather prediction parameters as input data, and utilizing the data base to figure out the wind speeds and the wind directions of the wind generation sets so as to figure out the generated powers of the wind generation sets, and accordingly obtain predicted value of the wind power station power. According to the invention, the physical prediction method is applicable to multiple wind power stations, has small calculated amount in the power prediction stage and has short calculation time.

The invention discloses a clustering-analysis-based wind power short-term prediction system and prediction method. The prediction system comprises a short-term prediction server and a real-time data acquisition apparatus; and a power prediction function unit and a prediction database are installed at the short-term prediction server. According to the prediction method, a daily correlation analysis is carried out by using a pearson product moment correlation coefficient to determine consistency of the daily correlation of the wind power and daily similar situation of the available weather information; clustering analysis pretreatment is carried out on historical weather database by using a K mean value clustering method; historical day data similar to a weather feature parameter of a prediction day are selected by using a method using an Euclidean distance as a similarity measure and the data are used as the training samples for neural network prediction model establishment; after training based on the similar samples after clustering, a wind power prediction model based on the cluster analysis is obtained; and the prediction day NWP information is used as input parameter of the model and the wind power is used as the model output, so that prediction power data of the prediction day are obtained.

The invention relates to a wind electric power prediction method and a device thereof. The method comprises the following steps of: step one: extracting data from SCADA (Supervisory Control and Data Acquisition) relative to a numerical weather prediciton system or a power system, and carrying out smoothing processing on the extracted data; step two: determining input and output of training samples of a least squares support vector machine according to the processed data; step three: initializing relevant parameters of a smallest squares support vector machine and an improved self-adaptive particle swarm algorithm; step four: optimizing model parameters according to an optimization process; step five: acquiring a model of the smallest squares support vector machine according to the optimized parameters; and step six: carrying out prediction according to the model of the smallest squares support vector machine. According to the wind electric power prediction method disclosed by the invention, a modelling process is simple and practical, wind electric power can be rapidly and effectively predicted, and the wind electric power prediction method has an important significance on safety and stability, and scheduling and running of the electric power system, and therefore, the wind electric power prediction method has wide popularization and application values.

The invention discloses a photovoltaic power prediction method based on a convolutional neural network and meta-learning. The method belongs to the technical field of renewable energy development andutilization, and comprises the steps of establishing a deep convolutional neural network model, taking historical photovoltaic power data, historical meteorological data and numerical meteorological forecast data as model input, and taking photovoltaic power of a day to be predicted as output to form a model training sample; analyzing the weather type of the day to be predicted based on the radiation data in the numerical weather forecast, and selecting a similar day training sample; based on a meta-learning strategy and similar day training samples, training a neural network model by using eight loss function indexes, and outputting eight prediction results to realize point prediction and probability prediction of photovoltaic power. The invention also discloses a prediction system of themethod. The method can adapt to photovoltaic power prediction conditions of different seasons and different weathers, has extremely high prediction precision, and can effectively improve the operation stability of the photovoltaic grid-connected system.

The present invention provides a wind power forecasting system, which utilizes the meteorological element forecast value output by a numerical weather forecast system and works out the forecast value of generation power of a wind power field in the determined future time through the calculation by a calculation processing unit. The wind power forecasting system can work out the generation power of wind power field at the future time ahead, thus providing a reliable basis for absorbing wind power by a grid, reducing peak regulation cost and improving wind power quality.

The invention belongs to the technical field of a power system and particularly relates to an integrated platform system for remote management and control of a wind power field cluster for cross-regional multiple-wind-field unified management and control. The integrated platform system for the remote management and control of the wind power field cluster comprises a remote wind power field monitoring subsystem, a wind power prediction sub system, a video image monitoring subsystem and a large screen projection display subsystem, wherein subsystems are in communication connection with another other through a communication network; the remote wind power field monitoring subsystem is used for acquiring the operation data of wind power field booster station monitoring, box transformer substation monitoring and real-time fan monitoring; the wind power prediction subsystem is used for downloading numerical weather prediction information, receiving the data of a wind power field anemometer tower, performing wind power field output prediction on each wind power field in an ultrashort period of future 0-4 hours and short period of 0-72 hours; and the wind power prediction subsystem is further used for giving early warning on disaster weather. The integrated platform system for the remote management and control of the wind power field cluster is capable of realizing cross-regional, multi-wind field state monitoring and operation management and realizing the wind power prediction, state detection and fault treatment of the full-digital wind power fields.

The invention provides a whole-network balance adjustment method and a system thereof for a heat supply pipe network. The order and the form of a model for a heat supply system are obtained accordingto the heat balance analysis. An expanded state observer of the heat supply system is constructed to obtain a system model, wherein the system model is not clear in internal structure and is not easyto model. Meanwhile, a time-lag performance parameter of the heat supply system is obtained. Based on the numerical weather forecast service, the outdoor temperature after the time lag tau of the system is predicted from a specific moment. Furthermore, a user heat supply load after the tau moment is predicted, and then the advanced control is carried out. Through constructing an auto-disturbance rejection controller module, the coupling between heat stations is overcome and the operation state of the heat supply system is regulated in real time. The action of an electrically operated valve ofa primary network is controlled, so that the indoor temperature of a secondary network user reaches a preset value. Based on the self-learning function of a BP neural network, the parameters of the auto-disturbance rejection controller are set. Based on the above steps, the coupling between heat stations is overcome, and the thermodynamic balance of the whole network is achieved. The indoor temperatures of all users are balanced and consistent. The purposes of reducing the heat loss of the pipe network and saving the energy are achieved. The method can be widely applied to the centralized heating control.

The invention discloses a photovoltaic power generation power prediction method based on support vector machine regression, and the method comprises the steps: firstly, obtaining the historical outputdata and numerical weather forecast data of a target station; Screening out meteorological factors with high correlation from the meteorological factors; Secondly, preprocessing the historical data set, selecting appropriate input parameters, and performing data normalization to construct an input vector of a support vector machine; Calculating correlation degrees between the historical data setand four typical days day by day by using a grey correlation coefficient method; Clustering correlation degree calculation results so as to divide the historical data into four training sets accordingto weather types; Carrying out training modeling on the classified historical samples by adopting a support vector machine regression algorithm to obtain a prediction model; Determining the weather type of the to-be-predicted day through correlation calculation, and calling a corresponding prediction model; And finally, prediction day value weather forecast parameters are input, and a power prediction result is obtained based on a support vector machine regression algorithm and a prediction model.

The invention discloses a wind power climbing event probability prediction method and system based on a Bayesian network, and the method comprises the steps: mining the dependency relationship betweena wind power climbing event and related meteorological influence factors such as wind speed, wind direction, temperature, air pressure, humidity, and the like, and building a Bayesian network topological structure with the highest fitting degree with sample data; quantitatively describing a conditional dependency relationship between the climbing event and each meteorological factor, estimating the value of each conditional probability in a conditional probability table at each node of the Bayesian network, and forming a Bayesian network model for predicting the wind power climbing event together with a Bayesian network topological structure; deducing the probability of occurrence of each state of the climbing event according to the numerical weather forecast information of the mastered prediction time; the value of the corresponding conditional probability at each node is adaptively adjusted, so that the inferred conditional probability result of each state of the climbing event is optimized, and the compromise between the reliability and the sensitivity of the prediction result is realized.

The invention discloses a wind speed modeling method considering fan positions of a plateau mountain area. The method comprises the steps of S1, collecting geographical coordinates of n fans of a wind power plant, a coordinate position-containing topographic map of a wind power plant region, wind speed data of numerical weather forecast of the wind power plant region, and historical wind speed data and a wind direction rose map of the wind power plant region; S2, classifying the positions of the n fans into three types including a flat region, a mountain slope front region and a mountain slope back region; S3, performing canopy clustering on the historical wind speed data of the wind power plant region to obtain m wind speed sections; and S4, obtaining a wind speed model considering the fan positions of the plateau mountain area through a linear regression analysis method by utilizing the historical wind speed data of the wind power plant region. The technical problems of large description error and relatively large prediction precision error during wind speed data description and power prediction of the wind power plant of the plateau mountain area by utilizing a small amount of real-time data of a wind measurement tower and numerical weather forecast in the prior art are solved.

The invention discloses a method for predicting short-term power of a wind farm on the basis of a BP (back propagation) neural network. The method includes the steps of a, acquiring historical records of meteorological element data of a location of the wind farm and output power corresponding to each record; b, correcting the meteorological element data into fan hub height data; c, applying the corrected meteorological element data as input data to be input into the BP neural network, and applying the output power corresponding to the meteorological element data as input of the BP neural network to train the BP neural network; d, acquiring meteorological element data of the location of the wind farm according to numerical weather prediction data in a prediction period, correcting the meteorological element data into fan hub height data and generating corrected meteorological element data; and e, inputting the corrected meteorological element data obtained in step d into the BP neural network, and outputting data which is generation output power of the wind farm in the prediction period. The method is simple, easy and highly accurate.

The invention discloses a photovoltaic power station short-term power prediction method, and the method comprises the steps: selecting meteorological factors such as season types, weather types and irradiation intensity / temperature as an input data set according to the historical output power data and numerical weather prediction data of a photovoltaic power station; preprocessing the input data set, extracting a day type feature vector, and clustering the day type feature vector by adopting a K-Means clustering method to obtain K different day type results; according to the numerical weatherprediction data of the prediction day, determining a day type to which the prediction day belongs, obtaining a data sample set of the most similar day of n days in the day type to which the predictionday belongs based on a similar day theory, taking the data sample set as prediction model training data, performing training modeling on the training data set by adopting a random forest regression prediction algorithm, and establishing a photovoltaic power station short-term power prediction model; and calling a photovoltaic power station short-term power prediction model based on the predictionday numerical weather prediction data, and obtaining a short-term power prediction result of the photovoltaic power station in the prediction day.

A neural network wind power forecasting method and system includes: collecting numerical weather forecasting data at forecasting time; the wind power prediction value is obtained by substituting the numerical weather forecast data into the prediction model constructed in advance. The prediction model is based on principal component analysis and neural network. The technical scheme of the inventioneffectively improves the prediction accuracy, shows that the method has certain feasibility and advancement in wind power prediction based on numerical weather prediction, and has certain advantagesin processing large sample data.

The invention provides a method and a device for power predication of a wind power station. The method for power predication of the wind power station includes: acquiring weather forecast data outputted by a numerical weather forecast system; using CFD (computational fluid dynamics) software for optimization calculation of the weather forecast data to obtain weather predication data of the wind power station; using a preset corresponding relation between an output power and weather data in a statistic model to obtain power predication data of the wind power station according to the weather predication data; and outputting the power predication data. Using the method and the device for power predication of the wind power station realizes data coupling of the numerical weather forecast system and a CFD system and obtains more precise point positions of fans and weather information of wind speed, wind direction and the like at the heights of the fans, the power predication data are obtained by means of the statistic model according to the weather predication data, inaccuracy caused by a power curve iterative computation mode is avoided, difference of power curves of different fans is eliminated, and accordingly precision in power predication of the wind power station is improved.

The invention provides a wind power cluster power interval prediction method and system based on deep learning. The method comprises steps that numerical weather forecast and historical wind power ofeach wind power station are obtained as original input data; mutual information between an interpretation variable and a target variable in a region is extracted by calculating the mutual informationof the interpretation variable so as to extract associated information; interpretation variables conforming to relevancy are selected; data reconstruction and dimension reduction are carried out by using a principal component analysis method. According to the method, the interval constraint condition is constructed, the prediction model is constructed by using deep learning, the reconstructed anddimensionality-reduced data is input into the model to be trained, model optimization is carried out by combining a particle swarm optimization method, the final prediction model is determined, and power interval prediction is carried out by using the final prediction model, so that the method has relatively high accuracy.

The invention provides a wind power plant power predication method. The method comprises the following steps: step A, collecting wind speed historical data of mesoscale numerical weather forecast and actually measured wind speed historical data matching the wind speed historical data in terms of time; B, matching wind speed historical data of mesoscale numerical weather forecast of a prediction day with the wind speed historical data of the mesoscale numerical weather forecast to obtain historical data with greatest similarity; C, determining a wind measurement tower actually measured wind speed matching the historical data with the greatest similarity in terms of time, and replacing the wind speed historical data of the mesoscale numerical weather forecast of the prediction day with the wind measurement tower actually measured wind speed; and D, establishing a fitting wind speed-power feature curve of a wind power plant area, and through combination with the wind measurement tower actually measured wind speed after replacement in the step C, obtaining predicted power of the wind power plant area at the predication day. According to the invention, compared to a conventional statistical scale-reducing method applying a nerve network, the wind power plant power predication method has the following advantages: the logic structure is optimized, and besides, a curve matching model also has the advantage of high execution efficiency.

The invention provides an uncertainty analysis method and device for short-term wind power prediction and relates to the technical field of wind farm power prediction. The method includes a step of taking normalized and preprocessed numerical weather forecast data and wind farm single wind turbine power data as an input and an output respectively and training a radial basis neural network model byinput data and output data in a training data set, a step of predicting short-term wind power in the test data set according to the trained radial basis neural network model and generating short-termwind power prediction output data, a step of determining a short-term wind power prediction error according to the short-term wind power prediction output data and test output data in the test data set and establishing a segmented cloud model of a prediction error of each short-term wind power segment, and a step of calculating upper and lower interval limit values of the short-term wind power prediction output data according to the segmented cloud model to be an uncertainty analysis result of short-term wind power prediction.

The invention discloses a low-cost photovoltaic power prediction method based on city weather forecasts. The low-cost photovoltaic power prediction method comprises the steps of first obtaining penetration coefficients of extrasolar solar radiation through the atmosphere by calculating the extrasolar radiation intensity through a Sun-Earth model; then obtaining a weather forecast type classifier model of statistic penetration coefficients which can be obtained; then obtaining the statistic penetration coefficients of 14 days ago; conducting subtraction of the penetration coefficients and the obtained statistic penetration coefficients to obtain a difference value sequence; conducting multiply-accumulate of the different value sequence with a corresponding weight sequence to obtain a penetration coefficient error; correcting the statistic penetration coefficients under a weather type predicted on the same day through the penetration coefficient error; meanwhile, using a solar radiation intensity curve on the prediction day to reversely obtain a prediction solar radiation intensity curve; and obtaining a generated power prediction curve of a power station through the light intensity-power relation. The low-cost photovoltaic power prediction method based on the city weather forecasts can achieve real-time prediction based on the release of weather state forecasts and does not rely on high-cost numerical weather forecast products, and the cost of the generated power forecast is reduced.

The invention discloses a wind power prediction system for an extreme scene in the field of wind power prediction, and the system comprises a data collection server, a database server, a switch, a PCworkstation, a wind power prediction server, and a power grid dispatching center. The data acquisition server is connected with the numerical weather forecasting system, the anemometer tower, the booster station SCADA server, the fan SCADA server and the three-dimensional laser radar measurement system through a communication network, and the data acquisition server is connected with the PC workstation, the wind power prediction server and the database server through a switch through the communication network; and the switch is connected with the power grid dispatching center through the datainterface server. By increasing the input information amount, reducing the prediction deviation and adopting a physical, statistical and learning combined prediction method, a combined prediction model with small prediction error and high calculation efficiency is established, the calculation of economic dispatching is enhanced by selecting an extreme scene of a load, and a central scene sample isconsidered, so that the economy of the system is ensured.

The invention discloses a sine normalization method for a power forecast model of a wind power plant, and belongs to the field of wind power forecast. The method comprises the following steps of: 1), obtaining n groups of numerical weather forecast data and output power data of the wind power plant; 2), initializing a BP (Back Propagation) neural network; 3), carrying out linear normalization processing on wind speed, wind direction sine, wind direction cosine, temperature, atmospheric pressure and humidity respectively, carrying out sine normalization processing on the output power data of the wind power plant; and 4), forecasting by using x'new as an input value of the BP neural network, and carrying out backward normalization on the obtained forecast result. The sine normalization method for the power forecast model of the wind power plant, disclosed by the invention, has the beneficial effects that the strong universality is realized; the forecast precision of a neural network power forecast model is improved remarkably; and the method is simple and feasible, and can be implemented without varying out modification on the original neural network power forecast model.

The invention discloses a multi-time step wind power prediction method based on dynamic feature selection. According to the method, an intelligent hybrid model method for mining a historical power time sequence and disclosed numerical weather prediction (NWP) data by utilizing a dynamic feature extraction algorithm is provided, so that the challenge that wind power generation is difficult to predict under different time steps is solved. According to the method, on the basis of available original data, a dynamic filtering method with minimum redundancy and maximum correlation (mRMR) is adoptedto automatically select input variables with different prediction step lengths; secondly, supervised learning is conducted on the input data with the optimal characteristics through an adaptive neuralfuzzy inference system (ANFIS); ANFIS parameters are trained by the model by using a particle swarm optimization (PSO) algorithm so as to achieve an optimal prediction effect; and finally, the proposed hybrid intelligent model is evaluated through the operation data of the actual distributed wind turbine generator, and the effectiveness of the model is verified through experimental results.

The invention provides a short-term wind power prediction method and system based on a deep learning network. The short-term wind power prediction method comprises the steps of obtaining numerical weather forecast data of an area where to-be-predicted wind power is located; inputting the numerical weather forecast data into a pre-trained deep learning mapping model to obtain a predicted value of the wind power, wherein the deep learning mapping model comprises a corresponding relationship between the numerical weather forecast data and the wind power prediction value; and forming a grid by thenumerical weather forecast data according to positions, wherein each grid point in the grid comprises a plurality of weather parameters. According to the invention, the short-term power prediction precision of the wind power plant can be improved, meanwhile, the modeling time of the regional wind power prediction model can be shortened, and required computing resources and manpower resources arereduced.

The invention discloses an ultra-short-term wind power prediction method considering the historical sample similarity, and the method comprises the following steps: analyzing the correlation between acurrent power value and a historical power value and the correlation between the current power value and a meteorological factor historical value, screening attributes with higher correlation, constructing historical samples, and reflecting the information of the power of a fan at the current moment. After dimension reduction of a historical sample matrix is conducted through a principal component analysis method, K-means clustering is carried out, and an appropriate clustering category K is selected according to a prediction effect, wherein K different clustering categories represent power generation conditions of different wind conditions; according to the category labels, historical numerical weather forecast information is adopted as input, the wind power value at the current moment is adopted as output, corresponding K support vector machine prediction models are established, and hyper-parameters such as the penalty coefficient and the kernel function bandwidth of the support vector machine are determined through a cuckoo search algorithm. According to the method, the problems that all external information cannot be reflected and overfitting are solved, the prediction precision can be effectively improved, and therefore the wind power absorption capacity is improved.

The invention discloses a short-term wind power prediction method fusing multi-source information, and the method comprises the following steps: (1) a sample generation step: constructing a training sample according to numerical weather forecast data and wind power sequence data; (2) a feature extraction and feature fusion step: performing feature extraction and fusion on the training sample constructed in the sample generation step; and (3) a power prediction step: obtaining the prediction power output of the feature code obtained in the feature extraction and feature fusion step at the corresponding moment through a multi-layer perceptron, i.e., a final prediction result. Compared with a traditional wind power prediction method, the wind power prediction method has the advantages that the weather forecast data and the historical wind power data are fused, so that the periodic characteristics implied in the historical power data are captured, the time sequence characteristics of the numerical weather forecast data are mined, the difference characteristics of different fans are modeled, and the prediction precision is higher.

The embodiment of the invention discloses a method and device for optimizing numerical weather forecast through assimilated inversion of water vapor content. The method comprises: calculating the layered water vapor content of each atmospheric layer through inversion according to the radiance received by an observation satellite and generated by ground radiation, optimizing the initial field of the numerical weather forecast mode according to the calculated layered water vapor content of each atmospheric layer to obtain an optimized initial field, and carrying out new weather forecast throughthe optimized initial field. The optimized initial field considers the influence of water vapor in the atmosphere on the field, so that the weather forecast of the optimized initial field can reflectthe real weather condition more accurately compared with the weather forecast based on the original initial field. According to the method, the optimization of the initial field by combining satelliteobservation data is realized, and the weather forecast closer to the real condition can be obtained through the optimized target initial field.

The invention relates to a multi-source meteorological data integration method and system. The method comprises: collecting weather meteorological data and electric power meteorological data to form original data; unifying the original data format into a preset standard format; and carrying out integrated processing on all meteorological data which are unified into a preset standard format to forma meteorological monitoring data set. The scheme is used for improving the numerical weather forecast precision and realizing the establishment and application of the customized numerical mode of thepower industry.

The invention provides a regional wind power prediction method and system based on space-time quantile regression. The method comprises the following steps: collecting the operation and numerical weather prediction data of a plurality of wind power plants in a preset time period, converting the collected data into a feature map, and building a training set, a verification set and a test set; establishing a space-time quantile regression model, and training and optimizing the model by utilizing the training set, the training set, the verification set and the test set; acquiring operation data and environment data of each wind power plant in real time, and predicting regional wind power generation in a certain time period in the future according to the optimized space-time quantile regression model. According to the invention, short-term non-parameterized probability prediction is carried out on regional wind power through the space-time quantile regression model; the selection problem of explanatory variables in regional wind power prediction with large input information is solved, the prediction accuracy and reliability are greatly improved, and a specific solution is provided forregional wind power generation probability prediction with big data.

The invention provides an EWT-PDBN combination-based short-term wind power prediction method. The method comprises the following steps of A, collecting numerical weather forecast data and historical wind power data of a wind power plant; b, performing preprocessing and normalization processing of all the acquired data; c, decomposing the normalized historical average wind power data by using an empirical wavelet transform signal decomposition technology; d, performing correlation screening of the decomposed different intrinsic mode component function sub-sequences, respectively taking the screened group sub-sequences and other data subjected to normalization processing as input data, and inputting the input data into a particle swarm optimization deep belief network model for prediction toobtain group prediction data; and E, superposing a group of prediction data to reconstruct a group of data, and then performing reverse normalization processing of the group of data to obtain the result as the final wind power prediction result. The method is advantaged in that through EWT-PDBN combined prediction, the wind power prediction result with high precision and small error is obtained.

The invention discloses a quantitative analysis method and system for an atmospheric pollution process, and the method comprises the steps: determining the start and stop time of a heavy pollution period according to observation data, and constructing a three-dimensional space grid for a region needing to be analyzed in a heavy pollution process; performing simulating by using a mesoscale numerical weather forecast mode to obtain hourly meteorological element values of the three-dimensional space grid in the heavy pollution period; and simulating and calculating the normal concentration valueof each atmospheric pollutant normally simulated by the three-dimensional space grid in the heavy pollution time period by utilizing a third-generation air quality mode, and simultaneously simulatingthe accompanying concentration value of each atmospheric pollutant which is not influenced by the chemical conversion and emission process in the same time period in an accompanying manner. The quantitative analysis method and system have the advantages that the influence of the physical process, the chemical conversion and the emission process on various pollutants in the atmosphere in the heavypollution process is quantitatively analyzed, the quantitative analysis results of the weather, the chemical conversion and the emission process are simultaneously output in the air quality simulationforecasting process, the calculation precision is high, and the error is small.

The invention discloses a refined assessment model of regional wind energy resources, which comprises the following steps: S1, using anemometer tower observation data, SCADA data of wind turbines and terrain parameters respectively belonging in small-scale wind field ranges to establish separate multi-reference-point wind energy resource assessment models of small-scale regions, wherein the small-scale wind field ranges are 1-20 km; S2, using weather station observation data included in regional ranges of 20-200km and the existing models in the step S1 to establish correlation models, obtaining corresponding weight coefficients, and establishing dynamic models; S3, using numerical weather forecast data to establish mesoscale models of regional ranges of 200-500km; S4, using a combination of the mesoscale models, a wind farm power prediction system, a GIS geographic information model and the existing models in the step S1 and the step S2 to establish a refined self-adaptive model; and S5, using a combination of the refined self-adaptive model and a power load dispatching system to realize refined regional wind energy assessment and visual dynamic dispatching management.