64 results about "Electricity price forecasting" patented technology

The invention discloses an information entropy-based self-adaptive integrated classification method of data streams. Concept drift can be detected, and duplicate concepts can also be identified. In asystem, a new classifier is reconstructed and put into a classifier pool only when existence of a new concept is detected, the problem of duplicate training caused by duplicate concept appearance is prevented, model updating frequency is reduced, and real-time classification ability and classification effect of a model are improved. Through carrying out performance analysis comparison with classical data stream algorithms on a synthetic dataset and a real dataset, experiments show that the method of the invention can cope with multiple types of concept drift, improves anti-noise ability of theclassification model, and also has lower time cost consumption on the premise of ensuring higher classification accuracy. The method of the invention can be applied to many practical problems of sensor network anomaly detection, credit-card fraud behavior detection, weather forecasting, electricity price prediction and the like.

The invention relates to a short-period electricity price prediction method based on a BP neural network and a Markov chain. The method comprises the following steps of (S1), establishing a BP neural network model for predicting a short-period electricity price and performing preliminary prediction, thereby obtaining a predicted electricity price; (S2), performing error analysis on the predicted electricity price obtained in the step (S1), and calculating an absolute percentage error APE; (S3), dividing the Markov state space of the APE by means of a fuzzy C-means algorithm, and obtaining the Markov chain state of the APE; (S4), according to the Markov chain state of the APE, calculating a Markov state transition probability matrix; (S5), according to the Markov state transition probability matrix, calculating an APE state probability vector at a predicated time; and (S6), correcting the predicted electricity price and obtaining a final prediction result. Compared with the prior art, the short-period electricity price prediction method has advantages such as high precision.

The invention provides a cuckoo algorithm-based virtual power plant optimized dispatching method. The method comprises the following steps of 1) collecting and inputting data, wherein the data comprises electricity price prediction data, fan output prediction data, contract parameters, penalty parameters, gas turbine parameters, pumped storage power station parameters and the like; 2) constructinga virtual power plant economic dispatching optimization model, and preliminarily solving the model by adopting the optimization software; 3) inputting network constraint conditions, including power distribution network grid parameters, power flow constraint parameters and the like, and constructing a virtual power plant safety dispatching model; 4) carrying out the solving operation by adopting the cuckoo algorithm; 5) outputting a result. Compared with the prior art, the method realizes the effective management of a large-scale distributed power supply, an energy storage device, a controllable load and the like based on the virtual power plant technology. Therefore, on the premise that the economical efficiency and the reliability of a dispatching scheme are ensured, the calculation complexity is reduced. Meanwhile, the calculation time is reduced.

The invention discloses an electricity selling company optimal profit method considering multiple types of loads, and the method comprises the steps of enabling an intelligent metering device to collect load data, and obtain a typical daily load curve of different types of participating loads in a current month; determining bilateral transaction electric quantity and electricity price of the totalload, and predicting spot market transaction electricity price according to the flat electricity price of different types of users; modeling the electricity price response behavior of the user basedon the peak-valley time-of-use electricity price, and constructing a profit model maximizing the profit of the electricity selling company; providing pricing strategies of three different types of users, and solving the pricing strategies by using optimization software to obtain an optimal profit strategy of the electricity selling company; responding optimal peak-valley time-of-use electricity price data, obtained by the model, of all types of users in the current month to change respective electricity utilization plans by the users through automatic demand response terminal equipment shutdown or operation power adjustment. The technical problem that existing e-commerce research content still lacks pricing strategies and income systems considering multiple types of user response behaviorsis solved.

The embodiment of the invention provides an electricity price prediction method and device for an electric power spot market, terminal equipment and a storage medium, and the method comprises the steps: obtaining historical electricity price data in a certain time range, and forming a historical electricity price data sequence; normalizing the historical electricity price data sequence, and carrying out wavelet decomposition on the normalized historical electricity price data sequence by adopting a set scale to decompose the normalized historical electricity price data sequence into sub-sequences with different frequencies; dividing the sub-sequences with different frequencies into a training set and a prediction set; training a CNN model by using the training set, and transmitting the prediction set to the trained CNN model for electricity price prediction to obtain prediction data; carrying out wavelet reconstruction and reverse normalization processing on the prediction data, and outputting an actual electricity price prediction sequence. The embodiment of the invention can improve the precision of electricity price prediction in the electric power spot market.

The invention relates to a bidding strategy for electricity purchase. The strategy comprises that a day similar to an electricity-price-to-predict day is selected via a grey correlation analysis method; according to the electricity price of the similar day, the electricity price of the electricity-price-to-predict day is predicted via a multivariate linear regression method; according to historical electricity prices of the similar days, the predicted electricity price of the similar day is obtained via the multivariate linear regression method, and a relative error probability density distribution function of electricity price prediction is established according to the historical and predicted electricity price; and according to the relative error probability density distribution function and the electricity cost of the electricity-price-to-predict day, bidding for electricity purchase is carried out via particle swarm optimization. According to embodiments of the invention, bidding for electricity purchase is carried out by utilizing the electricity cost of the day similar to the electricity-price-to-predict day, and the accuracy is high; and the established relative error probability density distribution function is used, and the electricity price of the electricity-price-to-predict day is adjusted via particle swarm optimization, so that the electricity purchasing price of the electricity-price-to-predict day is obtained, and bidding for electricity purchase of users is more accurate and more scientific.

The invention relates to the technical field of electricity price prediction, more specifically, to a short-term electricity price prediction method based on local mean decomposition and optimized RBFneural network. Firstly, historical electricity price data are obtained and pretreated. Then, the historical electricity price data is decomposed into a series of PF components with physical meaningby using local mean decomposition. Then, the prediction model of RBF neural network is optimized by crossover algorithm to predict all PF components in advance by 0.5 h. Finally, all the predicted values of PF components are superposed to obtain the actual predicted results. The invention utilizes local mean value decomposition to reduce non-stationarity and non-linearity of electricity price sequence, the hybrid model of radial basis function neural network is optimized by crossover algorithm, which makes up for the shortcoming that neural network is easy to fall into local optimum, improvesthe generalization ability of neural network, reduces the influence of complex characteristics of electricity price series on forecasting results, and improves the accuracy of short-term electricity price forecasting.

The invention provides a method and a system for electricity price prediction, and a computer readable storage medium. The method comprises the following steps: enabling historical electricity price data to be clustered into m types through a K-means algorithm, and obtaining historical electricity price data with a mode label; training and constructing an electricity price mode recognition model according to the historical electricity price data with the mode label; establishing n mutually independent day-ahead electricity price prediction models; respectively obtaining n electricity price prediction sequences of the prediction day through the n mutually independent day-ahead electricity price prediction models; inputting the n electricity price prediction sequences into the electricity price mode identification model, and outputting prediction results of n electricity price daily fluctuation modes; calculating the final voting score of each electricity price daily fluctuation mode according to the prediction results of the n electricity price daily fluctuation modes, and selecting the electricity price daily fluctuation mode with the highest score as the finally predicted electricity price daily fluctuation mode. The method can improve the prediction precision of the daily fluctuation mode of the electricity price, and further improves the prediction accuracy of the day-aheadelectricity price.

The invention provides a real-time electricity price forecasting system based on multi-density clustering and multi-core SVM, which is characterized in that a database management module is respectively connected with a data acquisition module, a power generation energy consumption statistics module, a real-time electricity price forecasting module and a data visualization module. The method and system is mainly used to analyze the time of real-time electricity price, electricity load, main energy generation quantity and generation cost Spatial distribution characteristics and summed up the law, forecasts the real-time tariff for the selected area, and considers the characteristics of the non-linearity of real-time electricity price, sparsity and volatility, as well as the load, the main energy production and the cost of power generation, which improves the prediction accuracy and adaptability of the system, avoids the over-fitting of the prediction model, improves the distributed processing ability, and reduces the computational complexity and time complexity. The real-time electricity price forecasting method based on multi-density clustering and multi-core SVM which is scientificand reasonable, and strong in is provided.

The invention discloses a method and device for predicting declared electricity prices. The method includes: calculating the average declared electricity prices of each historical month according to the historical transaction data information of each electricity selling enterprise, and calculating the month to be analyzed through the calculation of the average declared electricity prices The average declared electricity price of the electricity sales enterprise to be predicted; the price difference between the declared electricity prices of the electricity sales enterprise to be predicted is calculated through the data of the historical transaction declaration electricity price of the electricity sales enterprise to be predicted; According to the price difference between the declared electricity prices of the electricity selling enterprises, the lowest declared electricity price of the enterprise to be predicted is calculated; the power supply and demand ratio of each historical month is obtained, and the supply and demand ratio of the month to be analyzed is calculated; according to the supply and demand ratio of the month to be analyzed , adjusting the declared electricity price information of the enterprise to be predicted to generate a declared electricity price scheme. The invention realizes the purpose of improving the prediction accuracy of declared electricity price.

The invention relates to a real-time electricity price prediction method based on an improved Adam algorithm optimization deep neural network. The method comprises the following steps: (1) obtaining related data of electricity price and influence factors thereof in a power system as sample data; (2) carrying out normalization preprocessing on the related data of the electricity price and the influence factors thereof in the power system; (3) determining an input mode, an output mode, a hidden layer number, a hidden layer neuron number, a hidden layer transfer function, an output layer transfer function and a loss function of the neural network, improving the traditional Adam algorithm, and using the improved Adam algorithm to optimize the deep neural network model; (4) taking the influence factor which is highly correlated with the actual electricity price as an input quantity, taking the predicted electricity price as an output quantity, training a deep neural network model based on an improved Adam algorithm, and optimizing parameters of the deep neural network model; and (5) processing electricity price influence factor data of different nodes in the power system by using the finally optimized deep neural network model, and predicting real-time electricity prices of different nodes. According to the method, the data utilization sufficiency can be improved, the convergence speed of training is accelerated, and the accuracy of electricity price prediction is improved.

The invention discloses an electricity price prediction method. Through the method, historical electricity price data can be decomposed through a mode decomposition algorithm to obtain multiple discrete modes, then an optimized prediction model is utilized to perform prediction on each discrete mode, a prediction result corresponding to each discrete mode can be obtained, and a prediction result of electricity price can be obtained just by superposing all the prediction results. Since mode decomposition is performed on highly-nonlinear historical electricity price data and then mode decomposition results are predicted, for each decomposition result, the problem of high nonlinearity can be avoided in the prediction process, and consequently the prediction result under each discrete mode canbe more accurate; and the actual prediction result based on the historical electricity price data can be obtained just by superposing all the prediction results, and therefore the actual prediction result can be more accurate. The invention furthermore provides an electricity price prediction system and device and a computer readable storage medium, all of which also have the advantages.

The invention discloses a power grid operation management method based on electricity price prediction. The method comprises the steps of obtaining power grid operation data; establishing an electricpower spot market clearing model; determining uncertainty factors in the operation process of the power grid, wherein the uncertainty factors comprise uncertainty of a load, uncertainty of renewable energy sources, uncertainty of external electricity and uncertainty of market subject quotation; identifying an extreme scene in which the uncertainty factors influence clearing of the spot market model; inputting the power grid operation data into the electric power spot market clearing model based on the extreme scene, and performing clearing calculation to obtain a node electricity price changeinterval; and managing the power grid according to the node electricity price change interval. According to the method, the electricity price prediction of the electricity market can be accurately analyzed, the calculation complexity is greatly reduced, the electricity market and the power grid are reasonably managed by calculating the node electricity price change interval, and the safe and stable operation of the power grid is ensured. The invention further discloses a power grid operation management device based on electricity price prediction.

The invention provides a short-term electricity price prediction method and device based on an xgboost algorithm, terminal equipment and a storage medium. The method comprises the steps: obtaining training sample data according to a preset feature selection rule; carrying out normalization preprocessing on the training sample data; inputting the training sample data subjected to normalization preprocessing into a pre-constructed xgboost model for training to obtain an xgboost electricity price prediction model; obtaining prediction day feature data according to the feature selection rule, andinputting the prediction day feature data into the xgboost electricity price prediction model; and performing reverse normalization processing on the output value of the xgboost electricity price prediction model to obtain a predicted daily electricity price prediction value. According to the method, the own characteristics of the spot market clearing price data are considered, the key influence factors are screened out to construct the feature vectors, and the prediction model is constructed based on the xgboost algorithm to perform short-term electricity price prediction, so that the electricity price prediction precision is effectively improved.

The embodiment of the invention relates to a spot price prediction method and device, and the method comprises the steps: building a mean regression model which is provided with a seasonal part and ajump part; calibrating the mean regression model under an actual probability condition according to historical electricity price data to obtain a calibration model; and performing Monte Carlo simulation prediction of risk assessment according to the calibration model and the risk market price to obtain spot price. According to the embodiment of the invention, by establishing the regression model with the deterministic seasonal part and the jumping part, the comprehensive influence of deterministic factors and random factors on the electricity price can be considered during electricity price prediction, so spot electricity price prediction is more reasonable and accurate.

The invention aims to solve the problems that an existing electricity market electricity price prediction method consumes too long time, more samples cannot be trained to improve the prediction precision and mainly has limitations that research on point prediction, electricity price interval prediction and probability density estimation is less, and more effective information cannot be provided for market participants to make an optimal market strategy. An electricity price prediction method based on empirical mode decomposition and minimum gated memory network quantile regression is established, the electricity price sequence is decomposed into a plurality of modal components by adopting empirical mode decomposition; QR and MGM are combined to form a hybrid model QR-MGM to predict each modal component under different quantiles, and KDE is adopted for a reconstructed prediction result to obtain a probability density function of the electricity price, so that the uncertainty of prediction can be quantified, and the probability density function of the electricity price can also be obtained.

The invention provides a transaction electricity price prediction method and device. The method includes: acquiring transaction data; according to transaction data and a preset prediction algorithm, calculating a declaration electricity price predicted value, a minimum transaction declaration electricity price predicted value, a maximum transaction declaration electricity price predicted value anda monthly average transaction electricity price predicted value; generating a plurality of transaction electricity price ranges under conditions that the minimum transaction declaration electricity price predicted value is smaller than the monthly average transaction electricity price predicted value, the monthly average transaction electricity price predicted value is smaller than the declaration electricity price predicted value, and the declaration electricity price predicted value is smaller than the maximum transaction declaration electricity price predicted value. On the basis of the method, transaction electricity price ranges can be predicted on the basis of transaction data, so that influences of personal subjective thoughts are avoided, and the scientific and effective prediction of the transaction electricity price ranges can be realized.

The invention discloses a day-ahead electricity price prediction method based on an EEMD-CNN + SAE-RFR hybrid algorithm. The day-ahead electricity price prediction method comprises the steps of: A, combining historical electricity prices and original data of influence factors of the historical electricity prices to form an original feature time sequence matrix of a current prediction day d; B, performing data preprocessing on the original feature time sequence matrix; C, decomposing the preprocessed original feature sequence matrix into a plurality of multi-frequency modal components by using ensemble empirical mode decomposition, and combining the multi-frequency modal components into a plurality of multi-frequency two-dimensional feature matrixes according to high and low combined subsequences of frequencies; D, predicting the two-dimensional feature matrix of each frequency through a deep learning network model based on a convolutional neural network-stacked auto-encoder, and outputting a plurality of multi-frequency prediction time sequence subitems according to the frequency; and E, performing reconstruction fitting on all the prediction time sequence subitems by using a random forest regression algorithm to obtain a final electricity price prediction value. The method has the advantages of high accuracy, high model learning convergence speed and good result stability.

The invention relates to a deep learning electricity price prediction method and device based on electric power Internet of Things multi-source data fusion, and the method comprises the steps: obtaining the predicted electricity price of each region in a target region in a preset time period, and enabling an electricity price prediction model disposed in each region in advance to carry out the prediction of a corresponding historical electricity price data set, thereby obtaining the predicted electricity price; then acquiring a historical power load mean value of each region; carrying out fusion weight calculation on the historical power load mean value to obtain a fusion weight corresponding to the historical load mean value of each region; and performing fusion calculation on the predicted electricity price based on the fusion weight to obtain a final predicted electricity price in the preset time period. The electricity price is predicted through fusion calculation, large-scale transmission of data is avoided, and the pressure on communication bandwidth and corresponding computer resources can be effectively relieved.

The invention provides a stored energy operation control method based on electricity spot price prediction and tracking and a system thereof, and belongs to the field of stored energy coupling electricity trade under electricity spot market. Assuming a real-time market is a balancing market before one hour and temporarily ignoring the duration of charge and discharge of energy storage device, the electricity price at time T is the real price tracked in real time and the electricity price after time T is the forecast price. Judgments are made respectively on the basis of the predicted price of electricity at different times. The stored energy operation control method based on electricity spot price prediction and tracking and the system thereof are mainly used for guiding the energy storage unit to participate in the electric energy trade of intraday market and the electric auxiliary service of real-time market. According to the result of the electricity price forecast, the system guides the energy storage device to choose an appropriate time node for the charge and discharge operation. The system always judges the present moment and outputs the action of the energy storage device after one hour and the matching adjustment of the power transaction plan, which provides the maximal economic benefits of energy storage device in electricity spot market for users and electric companies.

The invention provides a node electricity price prediction method and device, and the method comprises the steps of obtaining a historical data set of a target node, constructing a feature library, and dividing the historical data set into a training set and an integration set; training the training set according to a feature library to generate an overall prediction model, training the training set according to an energy component, a blocking component and a network loss component of a historical electricity price according to the feature library to generate three groups of component prediction models, and adding the three groups of component prediction models to obtain a component-based prediction model; constructing a linear programming problem by taking minimization of regression lossas a target according to the integrated set and taking the integrated weight of the overall prediction model and the integrated weight based on the component prediction model as decision variables, solving the optimal integrated weight, and adding the overall prediction model and the integrated weight based on the component prediction model to obtain an electricity price prediction model; and obtaining input features of a target moment and inputting the input features into the electricity price prediction model to obtain predicted electricity price. Therefore, accurate node electricity price prediction can be realized, the economic benefits of participants in the electricity market are improved, and the method has relatively high application value.

An economical charging and discharging method for a storage battery of an electric vehicle belongs to the technical field of battery control, and comprises the following steps: 1, setting T as a total planning time scale, and issuing a day-ahead electricity price prediction curve Ya by a provincial power grid by using a power grid information end; 2, establishing a machine learning knowledge database AT, learning by using XGBoost to obtain a learning electricity price Yb; 3 forming a final electricity price Yc(t) at the moment t; and 4 generating a day-ahead charging and discharging arrangement curve graph S of the electric vehicle according to Ya, correcting and updating the curve graph S and the battery state SOCt of the electric vehicle, forming and returning to a power grid information end to feed back actual charging and discharging behaviors, and the like. By means of the method, the charging and discharging economical efficiency of the electric vehicle can be guaranteed.

The invention discloses an electricity price prediction method based on similar days. The electricity price prediction method comprises the steps: obtaining a load prediction curve of a prediction region on a prediction day from a load prediction system; determining a first similar day set comprising M similar days based on a least square method according to the load prediction curve and a pre-established historical database, and determining a second similar day set comprising N similar days based on a correlation coefficient statistical algorithm; calculating a first weighting coefficient anda second weighting coefficient according to the effective historical load curve and the load prediction curve of the similar days in the first similar day set and the second similar day set; calculating a first electricity price prediction value and a second electricity price prediction value according to the first weighting coefficient, the second weighting coefficient and the effective historical electricity price curve; calculating an electricity price correction amount according to the first electricity price prediction value and the second electricity price prediction value; and according to the electricity price correction amount and the second electricity price prediction value, calculating to obtain a standard electricity price prediction value of the prediction day. The electricity price prediction method can improve the accuracy of the electricity price prediction value.