74results about How to "Improve Load Forecasting Accuracy" patented technology

The invention provides a cyclic neural network short-period load predication method based on information entropy clustering and an attenuation mechanism. The method comprises the following steps of analyzing characteristics which affact the power load; calculating the information entropies of all characteristics to the load by means of an xgboost algorithm; performing cluster analysis based on theinformation entropy of each characteristic as the weight on the historical data of a predicated area by means of a clustering algorithm; selecting a cluster which is nearest to a predicating day weight from the clustering results, and forming a time sequence T according to a sequence that the time to the predicating time reduces from longest to shortest; using the time sequence T as an encoder ofthe attention cyclic neural network, and obtaining a predication result by a decoder. Compared with the prior art, the cyclic neural network short-period load predication method has advantages of high predication precision and high self-adaptability.

The invention discloses an intelligent power grid short-term load prediction method based on an improved RBF neural network, relates to the technical field of intelligent power grid, and is used for determining the basis function center and improving the load prediction precision of the intelligent power grid. The prediction method includes: S1, performing network initialization; S2, calculating the basis function center ci; S3, calculating the variance [zeta]i according to the basis function center ci; S4, calculating the output Ri of a hidden layer according to the basis function center ci and the variance [zeta]i; S5, calculating the output of an output layer according to the output Ri of the hidden layer; S6, calculating a prediction error E according to a mean squared error and the function; S7, updating connecting weights of neurons of the hidden layer and neurons of the output layer in the neural network; and S8, determining the prediction error E, if the prediction error E is expected, ending iterative calculation, and otherwise, returning to step S4, and re-performing iterative calculation on the prediction error E. The method is used for predicting the load of the power grid.

The invention relates to a short-term load predicting method of a power grid. The method comprises the steps: step 1, acquiring historical data and pre-treating the data; step2, decomposing the historical load sample data into a plurality of different-frequency sub-sequences by using wavelet decomposition; step 3, performing single-branch reconstruction to each sub-sequence; step 4, dynamically choosing training samples and establishing a neural network predicting model optimized by a vertical and horizontal intersection algorithm; step 5, predicting each sub-sequence 24 hours in advance by using the optimal neural network predicting model; and step 6, superposing the predicted value of each sub-sequence to obtain a whole prediction result. The inherent defects of the neutral network can be overcome by optimizing BP neutral network parameters by a brand-new swarm intelligence algorithm, that is, the vertical and horizontal intersection algorithm instead of the traditional algorithm; the burr problem caused by the impact load processing is solved by the wavelet decomposition, the precision declining resulting from the removal of the effective load in the burr pre-treatment is solved and the predicted value of the hybrid algorithm is more approximate to the actual measured load value.

The invention discloses a big data based power load prediction method. The method comprises the steps of step one, providing data information of N periods, obtaining a first power load predictive value of the (N+1) periods through a reinforcement learning load prediction data model directed at same data information and obtaining a second power load predictive value of the (N+1) periods in a data driving mode; step two, performing information fusion on the first power load predictive value and the second power load predictive value through a D-S evidence theory to obtain a final predictive result of the (N+1) periods. By the aid of the method, directed at a power load prediction system containing multiple dimensions and multiple stages of space, time, attributes and the like, a data driving theory based non-model load prediction controller and wavelet neural network based accumulative learning prediction are combined, information fusion is performed on the predictive values through the information fusion technology to obtain an optimal predictive value, and accordingly, the accuracy and the timeliness of load prediction are improved greatly.

The invention discloses a partition power grid bus load prediction system based on weather information. Real-time weather information and prediction weather information are used in the system, load prediction of all buses of converting stations of 500 kV and 220 kV is achieved, and recognition of power grid partition and partition load prediction are achieved. A prediction algorithm used by the system comprises a classical algorithm and an intelligent prediction algorithm, the classical algorithm comprises unary linear regression, quadratic polynomial regression, self-adaptation index prediction, index prediction, increasing rate prediction, nonhomogeneous index prediction, a B. Compertz model and a logistic model, the intelligent prediction algorithm comprises an optimized BP neural network algorithm and an optimized particle swarm algorithm, and the system selects a prediction algorithm in a preferential mode during a prediction process. The system is a day-ahead bus load prediction system, bus load and partition load of each time interval from morrow to multiple days in future are predicted, and prediction content is active load of 96 points of a predicted day.

A method and system of optimizing microgrid operations is provided. One or more intelligent microgrid coordinators interface with the microgrid such that those microgrid coordinators are able to measure and control all microgrid asset activities. The microgrid coordinator is used to forecast the microgrid's demanded load more accurately, and assign asset commands so as to optimize microgrid consumption, generation, and storage of load. The method and system incorporate a valuation of dispatchable load in optimization functions. The microgrid coordinator is further used to protect the microgrid assets from over-current situations when the microgrid is connected to the bulk electric system and when islanded. The method and system provide a means to test the microgrid controller prior to implementation on the microgrid in order to assure proper operation.

The invention relates to a host load prediction method based on multi-sequence combination, which belongs to the field of computer application technology. From the perspective of nonlinear and non-stationary information processing, the present invention constructs multiple sequences and utilizes various data relationships related to each sequence and load, and combines the wavelet-AR-SVR-MA model and the AR model to predict the host load. To improve the host load prediction accuracy. This method is suitable for the host load prediction of key periods or key moments.

The invention belongs to the technical field of short-term load forecast of an electric power system and discloses a short-term load forecast method for an electric power system based on a deeply recursive neural network. The short-term load forecast method includes (1), collecting historical power grid load and meteorological data and building a base for standby use; (2), getting rid of abnormaldata obtained in the step (1) and subjecting residual data to normalization processing; (3), determining a model structure with feedforward and feedback functions; (4), training a DRNN forecast modelbased on an IPSO algorithm through historical data; (5), using the DRNN forecast model based on the IPSO algorithm for forecast of actual load. The short-term load forecast method has the advantages that relevance layers are added on the basis of a multi-hidden-layer structure of a deep neural network, and an improved particle swarm algorithm is used as an optimized learning algorithm of the network to deeply optimize a model weight space; errors are decreased effectively, feedforward and feedback connection can be fused, the network generalization ability is improved and the load forecast precision is enhanced effectively.

The invention discloses a short-term predicating method suitable for typhoon weather. The short-term predicating method comprises the following steps: obtaining base similar day sample data and meteorological data with relatively great short-term load interferences, calculating meteorological correlation according to the meteorological data, obtaining optimal similar day according to the meteorological correlation, correcting the load of the optimal similar day according to the day type difference and the annual type difference; obtaining load for predicating a typhoon day according to the corrected load of the optimal similar day. The short-term load predicating method can be used for coping with load volatility under the typhoon weather, selecting the day with meteorological characteristic which is the most similar with to-be-predicated typhoon weather as the optimal similar day based on a grey correlation method, correcting the optimal similar day load according to the day type difference and the annular type difference, thereby greatly improving the load predicating precision under the typhoon weather, guiding a power grid to schedule the load predicating work on the typhoon day, making an ordered power consumption scheme for typhoon damages, and guaranteeing the safe and stable operation of the power grid.

The invention discloses an load prediction method and system based on LLSVM optimization. The load prediction method based on LSSVM optimization comprises steps of discriminating and correcting abnormal data after obtaining original history load data, constructing a characteristic vector, performing k average value clustering on characteristic vectors, choosing a input variable according to a clustering effect, 2) using a punishment factor C of an LSSVM model and a kernel function width parameter sigma as position coordinates of a particle in a searching space for a particle swarm algorithm, using a particle coordinate vector value [C,sigma] having a smallest fitness value as an output of the particle swam algorithm, (3) using optimized and processed input variable data as input and output of the LSSVM model, and using the [C,sigma] obtained from the step 2 to solve a load prediction regression equation and using the regression equation to perform load prediction. The load prediction method and system based on LSSVM optimization can correct abnormal data, find the most suitable punishment factor and the kernel function parameter and improve accuracy of the load prediction based on the LSSVM.

The invention relates to a multi-model integrated load prediction method based on wavelet transform, which is divided into four stages: 1, on the basis of considering multiple influence factors, historical load related data is subjected to a maximum information coefficient feature selection technology to obtain a feature candidate set with high correlation; 2, in order to obtain a stable load sequence and improve the prediction precision, multiple wavelet transforms are integrated into a multi-prediction model; and 3, each load correlation sequence after wavelet function decomposition is trained by an intelligent prediction sub-model, and the sub-models provide different predictions in the same hour; 4, combining the optimal prediction of each time period and providing a final prediction result by adopting online secondary learning in an integrated decision process. According to the method, the load prediction precision can be further improved on the basis of various single predictionmodels. The method is high in generalization ability, can adapt to various environments, has high applicability, and is beneficial to reducing the operation cost of a power system.

The invention relates to a networked comprehensive management method for the panoramic electricity consumption information of smart home. A unified smart home electricity consumption data platform can be constructed, kinds of advanced software application can be flexibly developed and expanded under the support of a unified data source, and a conventional electric energy meter can be replaced by networked metering and charging; a user can comprehensively learn about and analyze own electricity consumption process and way from all directions, define the structure of own energy charge and analyze the rationality and potential optimized trend of the structure; and the user autonomously plans and regulates an electricity consumption scheme by a home interaction terminal, so the pertinency of the electricity consumption scheme and the personality needs of the user can be ensured, and the simulation of 'virtual electricity consumption' can make the user clearly learn about the actual effects of a certain electricity consumption scheme.

The invention discloses a method for short-term load forecasting of comprehensive subnet accumulation based on forecasting credibility evaluation. The method includes the steps of historical forecasting error statistics and analysis, time-phased error credibility rating and load forecasting performed by adopting a city comprehensive subnet accumulation method according to the rating. In the method, various city load forecasting results are fully utilized, so that characteristic change rules of load compositions in different regions of a power grid are beneficially grasped by scheduling departments in a deep and refined manner, level-to-level management of the load forecasting is enhanced, and scientization and refinement level of the load forecasting is fully improved.

The present invention discloses a short-term load prediction method under a big data environment. The short-term load prediction method adopts a Hadoop architecture to carry out the distributed storage and processing on the mass data, thereby improving the load prediction speed. The short-term load prediction method of the present invention uses an improved particle swarm optimization algorithm tooptimize a conventional BP neural network, thereby improving the load prediction precision.

The invention discloses a training sample grouping construction method used for support vector regression (SVR) short-term load forecasting, and belongs to the field of intelligent computing and machine study. The training sample grouping construction method comprises a step of analyzing correlation, wherein the correlation degree of the load of each time interval and the loads of other time intervals is analyzed through the Tangs correlation degree of the grey correlation degree to form a correlation degree matrix; a step of grouping prediction problems, wherein the time intervals with high load correlation degree are divided into one group according to the correlation degree matrix; a step of constructing a reference load matrix; a step of selecting a reference load to construct a training sample, wherein linear function fitting is carried out on each row of the loads in a load variation rate matrix in a least square fit mode, and fitting variance is calculated; and a step of selecting the load of the time interval with small fitting variance to serve as the forecasting reference load of the group. The training sample grouping construction method used for the SVR short-term load forecasting is capable of improving the load forecasting accuracy, and avoids the problem of high time complexity. The experiment result shows that a short-term load forecasting model trained by the training sample constructed through the method has good performance in forecasting accuracy and time complexity.

The invention relates to a multi-sequence combination prediction method of host load based on wavelet packet decomposition, which belongs to the field of computer application technology. Starting from the perspective of nonlinear and non-stationary information processing, the present invention constructs multiple sequences and utilizes various data relationships related to each sequence and load, and combines the wavelet packet-SVR model and the AR model to predict the load of the host to improve the load of the host. load prediction accuracy. This method is suitable for the host load prediction of key periods or key moments.

The invention discloses a DBN power grid load prediction method and device based on generalized demand side resources. The method comprises: establishing a scheduling model based on an electricity price contract for the reducible load LC, the transferable load LS and the energy storage system ES, wherein the model determines an optimal scheduling plan of three generalized demand side resources participating in the power market by means of a load aggregator; on the basis, fusing generalized demand side resource influence factors into a DBN load prediction model, and establishing a DBN short-term load prediction model considering generalized demand side resources; and training and testing the prediction model in combination with historical load data and weather data to obtain a daily load prediction curve of the to-be-tested area. The method is high in prediction precision and good in stability, and can meet the power grid prediction requirements under load big data.

The invention discloses a novel deviation electric quantity assessment mechanism optimization design method based on PBR. According to the method, a novel assessment unit price piecewise linear deviation electric quantity assessment mechanism based on a PBR (performance assessment mechanism with a reward and punishment mechanism) is provided; a double-layer optimization model of deviation electricquantity assessment mechanism key parameter design is constructed by cooperatively considering the goals that an electric power transaction center maintains balance account stability and an electricity selling company pursues maximization of electricity purchasing and selling profits and risk comprehensive utility. An adjustable load is used as a measure for an electricity selling company to dealwith deviation assessment; based on the psychology of a user, the intention of the user for responding to economic incentives of an electricity selling company is simulated, an actual calling strategy for interruptible loads of the user under a deviation assessment mechanism is researched from the perspective of loss avoidance of the electricity selling company, and on this basis, an optimal operation decision model of the electricity selling company under renewable energy allocation is established. The method plays an important role in stimulating an electricity selling company to improve the load prediction precision and reduce the system deviation rate.

The invention discloses a central air conditioner load forecasting method. The central air conditioner load forecasting method comprises the following steps of: acquiring at least more than two load forecasting values at time t obtained by a central air conditioner system via a common load forecasting algorithm; acquiring an actual load measurement value of a central air conditioner at the time t;forming a forecasting matrix from the plurality of load forecasting values to obtain a combined load forecasting value at the time t; solving a deviation coefficient of load forecasting; and calculating a final load forecasting value of the central air conditioner system at time t+1 via the combined load forecasting value at the time t+1 and the deviation coefficient. The invention also providesan intelligent terminal and a storage medium containing the method. High-precision load forecasting is realized on a load of the central air conditioner, the load forecasting weight is automatically adjusted by utilizing a weight distribution principle, and the forecasting method with high load forecasting precision always has a high weight, so that the overall load forecasting precision of the system is always maintained at a high precision level, the operating condition of the air conditioning system is adjusted in time, and the operating energy consumption of the air conditioning system isreduced.

The invention provides a system load prediction method integrating an isolated forest and a long-term and short-term memory network, which includes: considering the problem that noise and abnormal points exist in original data, and adopting an isolated forest algorithm to eliminate the abnormal points in the data; decomposing the input data into intrinsic mode function (IMF (intrinsic mode function)) components with different frequencies by utilizing an EMD (empirical mode decomposition) (EMD) algorithm; adopting a separated long short term memory neural network (LSTM (long short term memory network)) to predict each IMF (intrinsic mode function) and residual error, and reconstructing a prediction value in each LSTM model. According to the method, system load trend prediction with time sequence characteristics is realized, and the early warning capability of the system for too high system load caused by external attack or internal disturbance is improved.

The invention discloses an enterprise gateway load prediction method based on operation load characteristics. The method comprises steps that (1), the sample data required for load prediction is acquired; (2), the influence factor information required for load prediction is acquired; (3), load characteristic analysis on an operation load is carried out, and prediction operation loads are divided on the basis of types according to a result; (4), prediction model matching is carried out, and sample set selection is carried out; (5), whether the model and the sample both have prediction conditions is determined, if yes, gateway component load prediction is carried out, prediction error statistics is further carried out, and component prediction results are corrected continuously according to prediction errors; and (6), all the operation component prediction results are superposed, whether the set conditions are satisfied is determined, if yes, the final gateway prediction result is outputted. The method is advantaged in that the final enterprises gateway load prediction result is acquired, taking consideration of load characteristics of each enterprise operation and load change trend, enterprise load prediction precision can be effectively improved.

The invention discloses an optimal operation strategy determination method of an electricity selling company under a piecewise linear deviation electric quantity assessment mechanism. The method comprises the steps of providing a demand side deviation electric quantity assessment mechanism with the assessment unit price being piecewise linear based on a PBR mechanism; considering the willingness of the user, establishing an adjustable load response model, and simulating the response of the user to the interrupt/increase instruction; researching an actual calling strategy of an electricity selling company for the adjustable load based on the established deviation assessment mechanism; analyzing renewable energy electricity purchasing business of the electricity selling company under the additional system, and calculating electricity purchasing and selling profits of the electricity selling company under the deviation assessment mechanism; and establishing an electricity purchasing and selling decision-making model taking the maximization of the comprehensive utility of the risk and the expected income of a single electricity selling company as a target, and solving the model. The operation strategy obtained through optimization of the method can enable an electricity selling company to effectively avoid market risks, reduce deviation electric quantity assessment and improve operation benefits, and has good economical efficiency and practical application value.