CN117788082A - Power market quotation decision method and system based on electricity price prediction - Google Patents
Power market quotation decision method and system based on electricity price prediction Download PDFInfo
- Publication number
- CN117788082A CN117788082A CN202410205139.3A CN202410205139A CN117788082A CN 117788082 A CN117788082 A CN 117788082A CN 202410205139 A CN202410205139 A CN 202410205139A CN 117788082 A CN117788082 A CN 117788082A
- Authority
- CN
- China
- Prior art keywords
- data
- power generation
- market
- price
- quotation
- 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
- 230000005611 electricity Effects 0.000 title claims abstract description 89
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000010248 power generation Methods 0.000 claims abstract description 129
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 238000004458 analytical method Methods 0.000 claims abstract description 25
- 230000006399 behavior Effects 0.000 claims abstract description 14
- 238000013523 data management Methods 0.000 claims abstract description 14
- 238000010801 machine learning Methods 0.000 claims abstract description 11
- 230000010354 integration Effects 0.000 claims abstract description 10
- 238000004140 cleaning Methods 0.000 claims abstract description 9
- 238000007418 data mining Methods 0.000 claims abstract description 9
- 230000008859 change Effects 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 230000014509 gene expression Effects 0.000 claims abstract description 6
- 238000012800 visualization Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 28
- 239000000446 fuel Substances 0.000 claims description 13
- 230000000007 visual effect Effects 0.000 claims description 13
- 238000006243 chemical reaction Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 238000012706 support-vector machine Methods 0.000 claims description 11
- 238000013527 convolutional neural network Methods 0.000 claims description 9
- 239000003245 coal Substances 0.000 claims description 8
- 238000007726 management method Methods 0.000 claims description 8
- 238000007637 random forest analysis Methods 0.000 claims description 8
- 230000008901 benefit Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000009827 uniform distribution Methods 0.000 claims description 7
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 7
- 230000007774 longterm Effects 0.000 claims description 6
- 230000003068 static effect Effects 0.000 claims description 5
- 230000003542 behavioural effect Effects 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 230000007787 long-term memory Effects 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000003062 neural network model Methods 0.000 claims description 3
- 238000012946 outsourcing Methods 0.000 claims description 3
- 230000008439 repair process Effects 0.000 claims description 3
- 238000013500 data storage Methods 0.000 claims description 2
- 238000013079 data visualisation Methods 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 claims description 2
- 208000001613 Gambling Diseases 0.000 claims 2
- 230000008569 process Effects 0.000 description 9
- 238000003066 decision tree Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 238000013507 mapping Methods 0.000 description 5
- 238000007405 data analysis Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000009977 dual effect Effects 0.000 description 2
- 238000009776 industrial production Methods 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005755 formation reaction Methods 0.000 description 1
- 239000002737 fuel gas Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000003345 natural gas Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a power market quotation decision method and a system based on electricity price prediction, wherein the method comprises the steps of firstly collecting data from a data source, cleaning the data and establishing a data management system; then, carrying out feature selection, feature extraction and data preprocessing on the collected data, selecting a model algorithm according to the requirements of the feature prediction task of the data, and establishing an electricity price prediction model; then, using machine learning and data mining to predict and analyze the electricity price based on the electricity price prediction model; and finally, establishing a quotation decision model, carrying out behavior analysis on the power generation enterprises and game parties, obtaining a quotation expression of the optimal bidding strategy of the power generation enterprises, providing a power market quotation decision suggestion, and establishing a visualization and decision support interface to realize integration and deployment of a power market quotation decision system. The method can better identify the risk of the market and reduce the risk of quotation decision; meanwhile, the change trend of the market is reflected, and the reliability and flexibility of quotation decision making are improved.
Description
Technical Field
The invention relates to the technical field of electricity price prediction of thermal power enterprises, in particular to an electricity market quotation decision-making method and system based on electricity price prediction.
Background
The power market quotation decision-making system needs to collect and analyze large amounts of market data, and therefore, data analysis and mining techniques need to be used to extract useful information to help businesses make more accurate decisions.
In the prior art, machine learning algorithms: the power market quotation decision system needs to build a predictive model to predict future power market price trends and supply-demand balances. Machine learning algorithms can help the system automatically learn and model from historical data. Automated decision making techniques: the power market quotation decision system needs to provide automated quotation decision suggestions to help businesses make optimal quotation decisions quickly. The automated decision making technique may be implemented by a computer program, reducing human intervention. Cloud computing technology: the power market quotation decision-making system needs to handle a large amount of data and computing tasks, and thus requires the use of cloud computing technology to provide computing power and storage space. Big data analysis technology: with the continuous development of the power market, the power enterprises need to process a large amount of data, so that the large data analysis technology is needed to process massive data, and the data analysis efficiency and accuracy are improved.
The invention relates to a Chinese patent application with the application number 202310611931.4, which discloses a method and a device for processing a unit transaction strategy in an electric power spot market, wherein the method comprises the following steps: acquiring initial output and termination output of a unit to be processed, and acquiring quotation information corresponding to the initial output and the termination output; generating a target transaction strategy corresponding to the unit to be processed according to the initial output, the final output and the quotation information, wherein the target transaction strategy consists of multiple sections of transaction data; and submitting the target transaction strategy to a server in response to the target transaction strategy meeting the verification condition, wherein the method can submit the transaction strategies of a plurality of units at one time when the units are more in the power spot market. But it has the following problems: a large amount of market data needs to be collected and analyzed, but these data may be missing, erroneous, or inaccurate, thus affecting the accuracy and reliability of the system; the price of the spot market is not predicted, the spot declaration cannot be guided, and the medium-long term contract is signed without practical guiding significance; while the target transaction policy may not be in line with the reality of the enterprise or applicable in a particular market environment.
Disclosure of Invention
The invention aims to solve the problems that: the utility model provides a power market quotation decision-making method and system based on electricity price prediction, which improves accuracy of quotation decision-making and better reflects the change trend of the market through the prediction and analysis of electricity price.
The invention adopts the following technical scheme: a power market quotation decision method and system based on electricity price prediction comprises the following steps:
step 1, collecting data from a data source, cleaning the data, and establishing a data management system;
the collected data includes: historical electricity price data, power supply and demand data, fuel price data, meteorological data, holiday and special event information, market rules, policy factors, and macro economic data.
Step 2, performing feature selection, feature extraction and data preprocessing on the collected data, selecting a model algorithm according to the requirements of a feature prediction task of the data, and establishing an electricity price prediction model;
the characteristics of the data include: time series characteristics, load related characteristics, power generation capability characteristics, power generation cost, meteorological characteristics, holiday and special event characteristics, market rules and policy characteristics, and macro economic characteristics;
predicting a demand for a task, comprising: ultra-short-term electricity price prediction, and current market discharge clear electricity price prediction in real time in the past; short-term electricity price prediction, namely off-the-shelf market electricity price prediction in front of spot goods within more than seven days; mid-long term electricity price prediction, and spot day-ahead market electricity price prediction for more than 7 days;
the model algorithm comprises: long-term memory neural network model, convolutional neural network, support vector machine and random forest; based on the collected data, four model algorithms are used for respectively carrying out prediction to obtain preliminary electricity price prediction results of the four model algorithms.
Step 3, utilizing machine learning and data mining to predict and analyze the electricity price based on the electricity price prediction model;
s3.1, taking the preliminary prediction results of the four model algorithms in the step 2 as input, comparing the prediction results of the current time node and actual price data in 60 days before, and evaluating the preliminary prediction accuracy by using a deviation model;
s3.2, combining the preliminary prediction accuracy of the four model algorithms to obtain a combined prediction result.
Step 4, establishing a quotation decision model, performing behavior analysis on the power generation enterprises and game parties to obtain quotation expressions of the optimal bidding strategies of the power generation enterprises, and providing quotation decision suggestions of the power market, wherein the steps comprise;
s4.1, acquiring the marginal cost of the power generation enterprise, and comparing the prediction result in the step 3 to construct a game model based on the real-time marginal cost;
s4.2, performing game party behavior analysis, and analyzing market profits of all power generation enterprises to obtain a reaction quotation formula of the power generation enterprises;
and S4.3, providing suggestions for bidding decisions of the power generation enterprises in the power spot market by using Bayesian equalization.
And 5, establishing a visual and decision support interface to realize the integration and deployment of the power market quotation decision system, wherein the specific steps are as follows:
s5.1, interface design: designing a decision support interface, wherein the decision support interface comprises input data, output results and interaction controls; the visual mode comprises a line graph, a column graph, a pie chart and a scatter graph;
s5.2, interface test and optimization: testing the decision support interface to ensure the correctness and stability of the interface; optimizing the decision support interface to improve user experience and decision efficiency;
s5.3, system integration and test: integrating all modules of the system together to form a complete system, and testing the system;
s5.4, preparing a deployment environment: preparing a deployment environment comprising a hardware environment and a software environment;
s5.5, system deployment: deploying the system into a designated environment;
s5.6, system maintenance: monitoring the system and knowing the running condition of the system; maintaining the system, including defect repair and function optimization; the system is upgraded to support new requirements.
The technical scheme of the invention also provides an electricity market quotation decision system based on electricity price prediction, which is used for realizing any one of the above electricity market quotation decision methods, and comprises the following steps:
and a data acquisition module: the system can integrate transaction center data in an automatic/manual mode, comprises meteorological data, thermal power overhaul information, cost data, contract data, new energy power prediction data, unit load data, fuel price information, market supply and demand data, power grid constraint and the like disclosed in the whole network, and combines a market transaction policy rule file to develop data unified management for long-term, spot and auxiliary service market conditions in a cross-province cross-district, long-term, spot and auxiliary service market conditions in a cross-province;
and a data processing module: the system performs basic information analysis of the electric power market, including analysis of supply and demand, network conditions, electricity price, power generation capacity, contract execution and operation conditions, compares predicted and actual data and operation date data, performs trend deviation analysis and visual display, establishes a data management system, selects a model algorithm, and establishes an electricity price prediction model;
the quotation decision module: using machine learning and data mining to predict and analyze electricity price, establishing a quotation decision model, analyzing game side behaviors, providing decision suggestions for medium-long term market and spot market transactions, and guiding spot market declaration and medium-long term bin adjustment;
and a visualization supporting module: and establishing a visual and decision support interface, providing types such as line graph display of electricity price trend analysis, histogram display of winning capacity, pie chart display of bin comparison and the like, and providing a large-screen cockpit interface display key index.
Compared with the prior art, the technical scheme provided by the invention has the following technical effects:
aiming at various uncertainties and risks of an electric power market trading body in market trading, the invention provides an electric power market quotation decision-making system based on electricity price prediction, and the change trend of the market can be better reflected through the prediction of the electricity price, so that the accuracy of quotation decision-making is improved; the automatic quotation decision system can complete the quotation decision process more quickly, so that the quotation decision efficiency is improved; thereby better identifying the risk of the market and reducing the risk of quotation decision; meanwhile, the change trend of the market is better reflected, and the reliability and flexibility of quotation decision making are improved.
Drawings
FIG. 1 is a flow chart of a power market quotation decision-making method based on price prediction of electricity according to the invention;
FIG. 2 is a basic step diagram of data processing of the power market quotation decision-making method of the invention;
FIG. 3 is a flow chart of the method for power market quotation decision-making of the present invention for establishing a quotation decision-making model.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the application will be further elaborated in conjunction with the accompanying drawings, and the described embodiments are only a part of the embodiments to which the present invention relates. All non-innovative embodiments in this example by others skilled in the art are intended to be within the scope of the invention. Meanwhile, the step numbers in the embodiments of the present invention are set for convenience of illustration, the order between the steps is not limited, and the execution order of the steps in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.
The invention relates to a power market quotation decision-making method and a system based on electricity price prediction, as shown in figure 1, comprising the following steps:
step 1: collecting data from a data source, cleaning the data, and establishing a data management system;
step 2: performing feature selection, feature extraction and data preprocessing on the collected data, selecting a model algorithm according to the requirements of a feature prediction task of the data, and establishing an electricity price prediction model;
step 3: using machine learning and data mining to predict and analyze electricity prices based on an electricity price prediction model;
step 4: establishing a quotation decision model, carrying out behavior analysis on a power generation enterprise and a game party, obtaining a quotation expression of an optimal bidding strategy of the power generation enterprise, and providing a power market quotation decision suggestion, wherein the method comprises the following substeps;
step 5: a visual and decision support interface is established to realize the integration and deployment of the power market quotation decision system,
in one embodiment of the invention, the power market quotation decision-making method is specifically as follows:
step 1: collecting and cleaning related data, and establishing a data management system; the collected data includes: historical electricity price data, power supply and demand data, fuel price data, meteorological data, holiday and special event information, market rules, policy factors, and macro economic data.
Historical electricity price data: real-time electricity prices or day-ahead market clear electricity prices for various periods of time (e.g., every 15 minutes, half hours, or hours) have elapsed;
power supply and demand data: grid load data, including historical and forecasted load curves; the power generation side supply capacity comprises output conditions of various generator sets, predicted output power of renewable energy sources (such as wind energy and solar energy) and the like;
fuel price data: the energy market price of coal, natural gas, petroleum and the like, because the fuel cost is an important component of the electricity price of thermal power;
weather data: information such as air temperature, humidity, wind speed, sunlight intensity and the like;
holiday and special event information: special date and event information that may significantly affect power demand, such as holidays, seasonal changes, large-scale activities, etc.;
market rules: market mechanism design parameters such as market bidding rules, marginal cost pricing principles, etc.;
policy factors: policy and regulation changes, such as patch policy adjustment, specific implementation details of the time-of-use price policy;
macroscopic economic data: the national economic running conditions, such as GDP growth rate, industrial production index and the like, indirectly reflect the power demand trend.
The data sources include: related internal management systems such as production operation, fuel management, financial settlement, marketing management, electricity selling management and the like of the power generation enterprises; power prediction related equipment installed by new energy enterprises; the transaction center information disclosure websites of each province; national power grid and southern power grid transaction center information disclosure websites; related files and macro economic data issued by national institutions such as national energy authorities, issuing and modifying committee and the like.
The data management system is established, as shown in fig. 2, and comprises the following sub-steps:
s1.1, clear data collection requirement: it is determined which data needs to be collected and the source of the data. And (3) data collection: data is collected from data sources, such as databases, files, sensors, and the like.
S1.2, data cleaning: the collected data is cleaned to ensure accuracy and quality of the data. The data cleaning process comprises the steps of detecting and deleting abnormal values, missing values, repeated values and the like.
S1.3, data integration: the cleaned data are integrated together to form a complete data set.
S1.4, data management: and establishing a data management system for managing the data, wherein the data comprises data storage, backup, security, access control and the like.
S1.5, data visualization: the data is visualized for better understanding and analysis of the data.
Step 2: selecting a proper model algorithm, and establishing an electricity price prediction model, wherein the specific steps are as follows:
s2.1, feature engineering: the data are subjected to feature selection and feature extraction so as to better reflect the change trend of electricity price;
s2.2, data preprocessing: carrying out preprocessing operations such as standardization, normalization and the like on the data so as to improve the accuracy and stability of the model;
s2.3, model selection: and selecting a model algorithm according to the characteristics of the data and the requirements of the prediction task.
Wherein, the characteristics of data include:
time series characteristics, load related characteristics, power generation capability characteristics, power generation cost, meteorological characteristics, holiday and special event characteristics, market rules and policy characteristics, and macro economic characteristics;
time series characteristics: trend of electricity price change in the past period of time;
load-related characteristics: the association relation between the data collected in the step 1 and the power consumption load comprises load growth rate, fluctuation and the like;
the power generation capacity is characterized in that: actual output and available capacity of various generator sets, including thermal power start-stop, climbing, hydropower storage capacity and incoming water, and predicted output power of renewable energy sources (wind energy and solar energy);
the power generation cost is as follows: the association relation between the electricity generation cost information and electricity price of various energy types such as fire coal, fuel gas, water electricity, nuclear power and the like;
weather characteristics: the influence of meteorological indexes such as temperature, humidity, wind speed, sunlight intensity and the like on the power generation efficiency of renewable energy sources and on the power demand;
holiday and special event features: whether the day is a weekday, a weekend, a holiday or whether a large-scale activity is held or not, and the influence on the load is caused;
market rules and policy features: bidding mode, marginal cost calculation rule and policy guide in market mechanism to influence electricity price;
macroscopic economic features: GDP, industrial production index, etc. are associated with electrical loads.
The demands of the predicted task include:
ultra-short-term electricity price prediction: predicting the current price of the spot market in real time;
short-term electricity price prediction: predicting the current market price of electricity in the spot day-ahead within seven days of three days;
medium-long term electricity price prediction: and forecasting the current market price of electricity of more than 7 days.
In this embodiment, the model algorithm selects four prediction models, i.e., LSTM (long and short term memory neural network model), CNN (convolutional neural network), SVM (support vector machine), and RF (random forest).
(1) LSTM: the control Gate comprises three parts, namely a Forget Gate, an Input Gate and an Output Gate.
The input gate is used for processing information input by the network, the information of the neurons can be updated by using the output product of two activation functions, namely Sigmoid and tanh, and the output gate is used for controlling the output of the neural network.
(2) CNN: the output result of the CNN is a specific feature space of each image, when the image classification task is processed, the feature space output by the CNN is used as the input of a fully-connected layer or a fully-connected neural network, and the fully-connected layer is used for completing the mapping from the input image to the label set, namely the classification; the most important task in the whole process is how to iteratively adjust the network weights through training data, namely a backward propagation algorithm.
The CNN model adopts a local connection and weight sharing mode to carry out high-dimensional mapping processing on the original data, so that the data characteristics are effectively extracted. In the convolutional layer operation, the CNN network greatly reduces the parameter quantity in the training process by a local connection and convolutional kernel weight sharing mode of the neurons, improves the model training speed, and enables the model to more effectively extract the characteristic information in the original data. In the pooling layer, the original data is subjected to abstract understanding, so that the feature dimension is reduced, the number of training parameters is effectively reduced, the degree of model overfitting is reduced, and the extraction efficiency of the feature data is improved.
Let the spatial coordinates of the single channel input image be%x,y) The convolution kernel size isp*qThe kernel weight is omega, and the image brightness value isvThe convolution process is the sum of kernel ownership and its corresponding element brightness on the input image, expressed as follows:
；
(3) And (3) SVM: searching an optimal hyperplane capable of meeting the requirements of classification conditions, ensuring satisfactory classification accuracy, and maximizing blank areas on two sides of the hyperplane.
And converting the data classification problem into a quadratic programming problem, solving a dual problem solution to obtain an original problem solution, and finally obtaining a decision function through the original problem solution. For the linearity problem, the support vector machine can realize the optimal classification, and for the nonlinearity problem, the sample vector is mapped to the high-dimensional characteristic space by a nonlinear mapping method, and the kernel function of the low-dimensional space is used for replacing the complex operation of the high-dimensional space, so that the dimension disaster is successfully avoided.
1) Linear support vector machine:
when the sample is linearly available, the process of machine learning is the process of determining the optimal hyperplane and classifying the sample.
Because the optimal classification hyperplane can maximize the classification interval, the problem can be converted into an optimal problem, and only one optimal solution is available, the problem can be solved by introducing a Lagrange method, and the solution of the optimal problem is converted into a dual problem, so that an optimal classification function is obtained.
2) Nonlinear support vector machine
In practice, most sample sets are linearly inseparable, and the SVM solves the problem of mapping samples in a low-dimensional space to a high-dimensional space, so that sample data can be linearly separable.
Therefore, the solution idea for a nonlinear support vector machine is to map the original samples to a feature space using a nonlinear mapping function, where the problem is solved using the principle of a linear support vector machine.
(4) Random forest RF
The decision tree algorithm is a supervised learning algorithm based on if-then-else rules, and the random forest is composed of a plurality of decision trees, and different decision trees are not associated with each other. When a classification task is carried out, a new input sample enters, each decision tree in the forest is respectively judged and classified, each decision tree can obtain a classification result of the decision tree, which classification result of the decision tree is the most, and then the random forest can take the result as a final result.
Step 3: and (3) predicting and analyzing electricity prices by using machine learning and data mining technologies:
the basic principle is as follows: and (3) determining the weight of the combined prediction model by adopting a BP neural network, so as to realize the combined prediction method, taking the prediction results of the single models of the four prediction models in the step (2) as input variables respectively, performing data preprocessing by adopting methods of clustering and removing abnormal values of data, normalizing the processed data, reducing noise of a data sequence and the like, and outputting clear price prediction results.
S3.1, taking the preliminary prediction results of the four model algorithms in the step 2 as input, comparing the prediction results of the current time node and actual price data in 60 days before, and evaluating the preliminary prediction accuracy by using a deviation model;
the bias model evaluation method is as follows:
；
wherein,price pre in order to predict the price to be cleared,price true for the actual price, the preliminary prediction accuracy of four model algorithms is calculated and expressed asacc 1 、acc 2 、acc 3 、acc 4 。
S3.2, combining the preliminary prediction accuracy of the four model algorithms, and expressing the combined prediction result as follows:
；
wherein,price pre i for a single forecast price to be clear,acc i for the prediction accuracy of the single model over the previous 60 days,i=1,2,3,4。
step 4: establishing a quotation decision model and providing decision suggestions:
based on the real-time cost measuring and analyzing model, an incomplete information static game model based on real-time marginal cost is constructed. By analyzing the benefit function of each power generation enterprise, an optimal reaction quotation formula of the power generation enterprise is deduced, as shown in fig. 3, and the method comprises the following sub-steps:
1) Market background description and gaming model assumptions
According to the electric power spot market transaction rule, the electric power spot market is assumed to existnIndividual power generation enterprises participate in bidding, each power generation enterprise hasn-1A gaming opponent; for power generation enterprisesIPutting it inton-1The game opponents are virtualized as an equivalent power generation enterpriseJAccording to the power generation enterprisesJThe bidding strategy estimation result is used for preparing the power generation enterpriseIIs a reaction strategy optimal for the reaction.
The clearing mechanism of the electric power spot market is based on market member declaration information and grid operation boundary price adjustment, and adopts a safety restraint unit combination (SCUC) and a safety restraint economic dispatch (SCED) program to perform optimization calculation, and clear to obtain spot market transaction results. In short, on the premise of ensuring the safety of the power grid, the unit with the cheapest quotation in the system is called in a limited way until the load requirement is met.
In the electric power spot market,nthe power generation enterprises are mutually independent, only limited information is held, only the power generation cost function of the enterprise is known, but the cost function of other power generation enterprises cannot be accurately obtained, and each power generation enterprise bids for simultaneous quotation, so that the power generation bidding of the power spot market belongs to the problem of incomplete information static game.
Therefore, an incomplete information static game model based on real-time marginal cost is constructed, bayesian equilibrium is solved, and advice is provided for bidding decisions of power generation enterprises in the power spot market.
In this embodiment, the real-time cost function of the power generation enterprise is:
；
wherein,Cfor the real-time cost of the power generation enterprises,C FU in order to be a cost of the fuel,C OP in order to purchase the power fee for outsourcing,C W for the water cost, the water cost is used,C VT in order to change the tax amount of the tax,C D in order to be a depreciated fee,C O other costs.
Among the cost components, the outsourcing power fee, the water fee, the variable material fee and the variable tax account for a relatively small margin cost, so the fuel cost is an important analysis element, and the power generation real-time margin cost function model is simplified as follows:
；
wherein,MCfor the sake of real-time marginal cost,b n x for the standard coal consumption of the power supply,P f is real-time comprehensive standard coal unit price data.
；
Wherein,HRis the heat consumption rate of the steam turbine,Q n is the low calorific value of the standard coal,for boiler efficiency>For the purpose of the efficiency of the pipe,is the plant power consumption.
In this embodiment, pricing is based on the actual marginal cost of the power generation enterprise, assuming that the bidding of the power generation enterprise is a unitary linear function of the marginal real-time costIThe bid formula for (a) is as follows:
；
wherein,P i power generation enterprisesIIs a function of the quotation of (1),α i 、β i respectively is a unitary linear functionA constant of numbers.
2) Game party behavior analysis
Assuming that the power generation enterprises participating in the bidding of the power spot market are rational and independent of each other, the bidding of each power generation enterprise aims at the maximization of the income.
Representing the electric power spot market bidding problem as a standard incomplete information static game, the game behavior space needs to be givenA i Type spaceT i Judging participants to obtain power generation enterprisesIIs a function of the benefit of (2)U i 。
Behavioral spaceA i : the price of each section of quotation of the power generation enterprises can not exceed the upper limit and the lower limit of the declared price specified by the transaction center. The upper limit and the lower limit of the declaration price are defined by comprehensively considering the factors such as the operation of power generation enterprises, the electricity price bearing capacity of market users and the like, and the upper limit of the declaration price is as followsP h The lower limit of declaration isP l Behavioral spaceA i =[P l ，P h ]。
Type spaceT i : in the incomplete information game of the power spot market bidding, each power generation enterprise can only grasp the cost information and the quotation strategy of the enterprise, and the quotation information of other power generation enterprises can not be determined. Real-time marginal cost compliance of hypothetical power generation enterprisesMC l ，MC h ]Uniform distribution on, type spaceT i =[MC l ，MC h ]Wherein, the method comprises the steps of, wherein,MC h 、MC l and respectively reporting real-time marginal cost corresponding to the upper limit and the lower limit of the price.
Judging the participants: each gaming party knows that the type of opponent is subject to [MC l ，MC h ]And the uniform distribution is the judgment of each power generation enterprise on the type of the game opponent.
Behavior space according to the gameA i Type spaceT i Judgment of participants, useP i AndP k respectively represent power generation enterprisesIPrice of quotation function and market margin to obtain power generation enterpriseIIs a function of the benefit of (2)U i The method comprises the following steps:
(1) When (when)P i >P k When ui=0.
When power generation enterprisesIWhen the quotation is higher than the price result predicted by the combination, the power generation enterpriseIThe network-surfing electric quantity of (1) is zero, namely the power generation enterpriseIThe power generation can not be performed on the internet, and the income is zero.
(2) When (when)P i ≤P k In the time-course of which the first and second contact surfaces,
；
wherein,E i in order to participate in the bidding electric quantity of spot market, the bidding electric quantity isQ i ，P i Is a power generation enterpriseIIs a function of the quotation of (1),P j is a power generation enterpriseJQuotation function, MC of i Is a power generation enterpriseIIs the marginal cost curve of MC j Is a power generation enterpriseJIs set in the upper limit of the marginal cost curve,nindicating the same number of power generation enterprises that are bidding.
When power generation enterprisesIThe quotation of the price is smaller than or equal to the marginal price of the market, namely the power generation enterpriseIWhen a certain internet power can be obtained, power generation enterprisesIThe benefit function of (c) is divided into three cases:
a. power generation enterprisesIThe price of the electricity generation system is lower than the price of the virtual electricity generation enterprise J, the power plant does not belong to a marginal unit, the corresponding declaration output is all winning, the declaration electric quantity is equal to the internet electric quantity,E i =Q i ;
b. in real-world situations, power generation enterprisesIThe quotation is equal to that of all virtual power generation enterprisesJThe situation of the quotation of (2) is not established, and is therefore not considered;
c. power generation enterprisesIThe quotation of (2) is larger than that of the virtual power generation enterpriseIndustry is provided withJWhen quoting, the power generation enterprise I is a marginal unit at the moment; when the total market demand is greater than or equal to the sum of all the generator declarations, the generator enterprise I declares that the output is all winning,E i =Q i the method comprises the steps of carrying out a first treatment on the surface of the When the total market demand is smaller than the sum of all the generator declarations, the power generation enterprise I declares the winning bid in the output part,
。
3) Bayesian equilibrium solution
And analyzing that the incomplete information game of the power generation enterprises in the power spot market needs to find Bayesian Nash equilibrium, and constructing a strategy space of a game party. In the game, the power generation enterprisesIIs in accordance with the functional relation of the strategyP i （MC i ) The collection of all functional relationships forms a power generation enterpriseIIs a policy space of (1); policy combinationIs Bayesian Nash equilibrium of the game;
policies representative of power generation enterprise IP i （MC i ) Strategy for power generation enterprise JP j （MC j ) The optimal reaction of each other is that the maximization of the income of each power generation enterprise meets the following conditions:
；
wherein,P i =P i （MC i ），P j =P j （MC j ），i,j=1,2,...,n,θE i indicating the bid in the declaration output section,P{P i <P j ' represent power generation enterprisesIThe price of the system is lower than that of a virtual power generation enterpriseJIn the case of a bid for (a) the (b),P{P i >P j ' represent power generation enterprisesIThe quotation is higher than that of a virtual power generation enterpriseJIs the case for the quotes of (a).
Due toMC i Obeys uniform distribution and thusAlso obeys uniform distribution, the above variants are:
；
the first-order guide condition is as follows:
；
let it be 0 to obtain:
；
from the following componentsThe method can obtain:
；
and (3) the same principle:
；
and (3) solving a simultaneous equation group:
；
；
wherein,θindicating the proportion of the winning power to the declared power,MC h 、MC l real-time marginal formations corresponding to upper and lower limits of declaration price respectivelyThe method is a power generation enterpriseITo power generation enterprisesJStrategyP j Is a reaction strategy optimal for the reaction.
Therefore, power generation enterprisesIThe bid expression of the optimal bidding strategy of (2) is:
；
step five: a visual and decision support interface is established to realize the integration and deployment of the system, and the specific steps are as follows:
(1) Interface design: designing a decision support interface, wherein the decision support interface comprises input data, output results and interaction controls; the visual mode comprises a line graph, a column graph, a pie chart and a scatter graph;
(2) Interface test and optimization: testing the decision support interface to ensure the correctness and stability of the interface; optimizing the decision support interface to improve user experience and decision efficiency;
(3) System integration and testing: integrating all modules of the system together to form a complete system; testing the system to ensure the correctness and stability of the system;
(4) Deployment environment preparation: preparing a deployment environment comprising a hardware environment, a software environment and the like;
(5) And (3) system deployment: deploying the system into a designated environment;
(6) And (3) system maintenance: monitoring the system so as to better know the running condition of the system; maintaining the system, including defect repair, function optimization and the like; the system is upgraded to better support the new needs.
In summary, in an embodiment of the present invention, an electricity market quotation decision system based on electricity price prediction includes:
and a data acquisition module: the system can integrate transaction center data in an automatic/manual mode, comprises meteorological data, thermal power overhaul information, cost data, contract data, new energy power prediction data, unit load data, fuel price information, market supply and demand data, power grid constraint and the like disclosed in the whole network, and combines a market transaction policy rule file to develop data unified management for long-term, spot and auxiliary service market conditions in a cross-province cross-district, long-term, spot and auxiliary service market conditions in a cross-province;
and a data processing module: the system performs basic information analysis of the electric power market, including analysis of supply and demand, network conditions, electricity price, power generation capacity, contract execution and operation conditions, compares predicted and actual data and operation date data, performs trend deviation analysis and visual display, establishes a data management system, selects a model algorithm, and establishes an electricity price prediction model;
the quotation decision module: using machine learning and data mining to predict and analyze electricity price, establishing a quotation decision model, analyzing game side behaviors, providing decision suggestions for medium-long term market and spot market transactions, and guiding spot market declaration and medium-long term bin adjustment;
and a visualization supporting module: and establishing a visual and decision support interface, providing types such as line graph display of electricity price trend analysis, histogram display of winning capacity, pie chart display of bin comparison and the like, and providing a large-screen cockpit interface for displaying key indexes.
The foregoing is only a preferred embodiment of the invention, it being noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the invention.
Claims (10)
1. The electricity market quotation decision-making method based on electricity price prediction is characterized by comprising the following steps of:
step 1, collecting data from a data source, cleaning the data, and establishing a data management system;
step 2, performing feature selection, feature extraction and data preprocessing on the collected data, selecting a model algorithm according to the requirements of a feature prediction task of the data, and establishing an electricity price prediction model;
step 3, utilizing machine learning and data mining to predict and analyze the electricity price based on the electricity price prediction model;
and 4, establishing a quotation decision model, performing behavior analysis on the power generation enterprises and the game parties, obtaining quotation expressions of the optimal bidding strategies of the power generation enterprises, and providing quotation decision suggestions of the power market.
2. The power market quotation decision-making method according to claim 1, wherein in step 1, the collected data comprises: historical electricity price data, power supply and demand data, fuel price data, meteorological data, holiday and special event information, market rules, policy factors and macro economic data; the data management system is established, which comprises the following sub-steps:
s1.1, data collection: determining data to be collected and a source of the data, and collecting the data from the data source;
s1.2, data cleaning: cleaning the collected data, including: detecting and deleting abnormal values, missing values and repeated values;
s1.3, data integration: integrating the cleaned data together to form a complete data set;
s1.4, data management: establishing an electric power market data management system, and managing data, wherein the management system comprises data storage, backup, security and access control;
s1.5, data visualization: the data is visualized for understanding and analyzing the data.
3. The power market quotation decision-making method according to claim 1, wherein in step 2, the characteristics of the data comprise: time series characteristics, load related characteristics, power generation capability characteristics, power generation cost, meteorological characteristics, holiday and special event characteristics, market rules and policy characteristics, and macro economic characteristics;
predicting a demand for a task, comprising: ultra-short-term electricity price prediction, and current market discharge clear electricity price prediction in real time in the past; short-term electricity price prediction, namely off-the-shelf market electricity price prediction in front of spot goods within more than seven days; mid-long term electricity price prediction, and spot day-ahead market electricity price prediction for more than 7 days;
the model algorithm comprises: long-term memory neural network model, convolutional neural network, support vector machine and random forest; based on the collected data, four model algorithms are used for respectively carrying out prediction to obtain preliminary electricity price prediction results of the four model algorithms.
4. A method of making a decision on a market quote for electricity according to claim 3, characterized in that said electricity price prediction and analysis of step 3 comprises the sub-steps of:
s3.1, taking the preliminary prediction results of the four model algorithms in the step 2 as input, comparing the prediction results of the current time node and actual price data in 60 days before, and evaluating the preliminary prediction accuracy by using a deviation model;
the bias model evaluation method is as follows:
（1）
wherein,price pre in order to predict the price to be cleared,price true for the actual price, the preliminary prediction accuracy of four model algorithms is calculated and expressed asacc 1 、acc 2 、acc 3 、acc 4 ；
S3.2, combining the preliminary prediction accuracy of the four model algorithms, and expressing the combined prediction result as follows:
（2）
wherein,price pre i for a single forecast price to be clear,acc i for the prediction accuracy of the single model over the previous 60 days,i=1,2,3,4。
5. the power market quotation decision-making method according to claim 4, wherein step 4 provides power market quotation decision-making advice comprising the sub-steps of:
s4.1, acquiring the marginal cost of the power generation enterprise, and comparing the prediction result in the step 3 to construct a game model based on the real-time marginal cost;
s4.2, performing game party behavior analysis, and analyzing market profits of all power generation enterprises to obtain a reaction quotation formula of the power generation enterprises;
and S4.3, providing suggestions for bidding decisions of the power generation enterprises in the power spot market by using Bayesian equalization.
6. The method of claim 5, wherein S4.1 establishes a game model based on real-time marginal cost, comprising the steps of:
s4.1.1, according to the electric power spot market transaction rule, the electric power spot market hasnIndividual power generation enterprises participate in bidding, each power generation enterprise hasn-1A gaming opponent; for power generation enterprisesIPutting it inton-1The game opponents are virtualized as an equivalent power generation enterpriseJAccording to the power generation enterprisesJThe bidding strategy estimation result is used for preparing the power generation enterpriseIIs the best reaction strategy of (a):
s4.1.2, constructing a game model based on real-time marginal cost, wherein the real-time cost function of a power generation enterprise is as follows:
（3）
wherein,Cfor the real-time cost of the power generation enterprises,C FU in order to be a cost of the fuel,C OP in order to purchase the power fee for outsourcing,C W for the water cost, the water cost is used,C VT in order to change the tax amount of the tax,C D in order to be a depreciated fee,C O other costs;
among the cost components, the fuel cost is an important analysis component, and the power generation real-time marginal cost function model is simplified as follows:
（4）
wherein,MCfor the sake of real-time marginal cost,b n x for the standard coal consumption of the power supply,P f real-time comprehensive coal marking unit price data;
（5）
wherein,HRis the heat consumption rate of the steam turbine,Q n is the low calorific value of the standard coal,for boiler efficiency>For the pipe efficiency>The power consumption is the factory power consumption;
s4.1.3 pricing is based on the actual marginal cost of the power generation enterprise, assuming that the bidding of the power generation enterprise is a unitary linear function of the marginal real-time costIThe bid formula for (a) is as follows:
（6）
wherein,P i is a power generation enterpriseIIs a function of the quotation of (1),α i 、β i respectively a constant of a unitary linear function.
7. The power market quotation decision-making method according to claim 6, wherein the S4.2 gambling party behavior analysis, representing the power spot market bidding problem as a standard incomplete information static gambling, comprises:
behavioral spaceA i : generating electricityThe price of each section of quotation of the enterprise cannot exceed the upper limit and the lower limit of the declared price specified by the transaction center, and the declared price upper limit is thatP h The lower limit of declaration isP l Behavioral space A i =[P l ，P h ]；
Type spaceT i : in the incomplete information game of the electric power spot market bidding, each power generation enterprise grasps the cost information and quotation strategy of the enterprise, and the real-time marginal cost of the power generation enterprise obeys [MC l ，MC h ]Uniform distribution over the class space, the class space is expressed as:T i = [MC l ，MC h ]，MC h 、MC l real-time marginal cost corresponding to the upper limit and the lower limit of the declaration price respectively;
judging the participants: each gaming party knows the type compliance of the gaming opponentMC l ，MC h ]Uniformly distributed on the upper part;
behavior space according to the gameA i Type spaceT i Judgment of participants, useP i AndP k respectively represent power generation enterprisesIPrice of quotation function and market margin to obtain power generation enterpriseIIs a function of the benefit of (2)U i The method comprises the following steps:
s4.2.1 whenP i >P k In the time-course of which the first and second contact surfaces,U i =0；
when power generation enterprisesIWhen the price of the price is higher than the marginal price of the market, the power generation enterpriseIThe network-surfing electric quantity is zero, and the power generation enterpriseIThe internet cannot be used for generating electricity, and the income is zero;
s4.2.2 whenP i ≤P k At the time, power generation enterprisesIPrice of market marginal price is smaller than or equal to that of price of market marginal price, and power generation enterpriseICan obtain a certain amount of electricity to get on the internet, and is a power generation enterpriseIThe profit function of (c) is divided into three cases:
（7）
wherein,E i in order to participate in the bidding electric quantity of spot market, the bidding electric quantity isQ i ，P j Is a power generation enterpriseJIs a function of the quotation of (1),MC i is a power generation enterpriseIIs set in the upper limit of the marginal cost curve,MC j is a power generation enterpriseJIs set in the upper limit of the marginal cost curve,nrepresenting the same number of power generation enterprises with quotations;
power generation enterprisesIThe three cases of the benefit function of (a) are as follows:
(1) Power generation enterprisesIThe price of the system is lower than that of a virtual power generation enterpriseJThe power plant does not belong to the marginal machine set, the corresponding declaration output is all winning, the declaration electric quantity is equal to the internet electric quantity,E i =Q i ；
(2) Power generation enterprisesIThe quotation is equal to that of all virtual power generation enterprisesJIn reality, this is not true;
(3) Power generation enterprisesIThe quotation is larger than that of a virtual power generation enterpriseJDuring quotation of (a) power generation enterprisesIAt this time, a marginal unit;
when the total market demand is greater than or equal to the sum of all the generator claims, generating enterprisesIThe declaration output is all the winning bid,E i =Q i the method comprises the steps of carrying out a first treatment on the surface of the When the total market demand is smaller than the sum of all the generator declarations, generating enterprisesIThe bid-winning in the declaration output section,Q i =θE i ，θindicating the proportion of the winning electric quantity to the declared electric quantity, 0<θ<1。
8. The method for determining the price quote of the electric power market according to claim 7, wherein the method for solving the bayesian equilibrium of the S4.3 is as follows:
constructing a strategy space of a game party: in the game, the power generation enterprisesIIs in accordance with the policy of (a)Functional relationshipP i （MC i ) Policy combinationIs Bayesian Nash equilibrium of the game;
representative of power generation enterprisesIPolicy of (2)P i （MC i ) And power generation enterprisesJPolicy of (2)P j （MC j ) The optimal reaction of each other is that the maximization of the income of each power generation enterprise meets the following conditions:
（8）
wherein,i,j=1,2,...,n，θE i indicating the bid in the declaration output section,P{P i < P j ' represent power generation enterprisesIThe price of the system is lower than that of a virtual power generation enterpriseJIn the case of a bid for (a) the (b),P{P i >P j ' represent power generation enterprisesIThe quotation is higher than that of a virtual power generation enterpriseJIs the case of offers of (a);
due toMC i Obeys uniform distribution and thusAlso obeys uniform distribution, the above variants are:
（9）
the first-order guide condition is as follows:
（10）
let it be 0 to obtain:
（11）
from the following componentsThe method can obtain:
（12）
and (3) the same principle:
（13）
and (3) solving a simultaneous equation group:
（14）
（15）
power generation enterprisesIThe bid expression of the optimal bidding strategy of (2) is:
（16）
thereby obtaining the power generation enterpriseITo power generation enterprisesJStrategyP j Is a reaction strategy optimal for the reaction.
9. The power market quotation decision-making method according to claim 1, further comprising step 5: the method comprises the following specific steps of establishing a visual and decision support interface to realize the integration and deployment of an electric power market quotation decision system:
s5.1, interface design: designing a decision support interface, wherein the decision support interface comprises input data, output results and interaction controls; the visual mode comprises a line graph, a column graph, a pie chart and a scatter graph;
s5.2, interface test and optimization: testing the decision support interface to ensure the correctness and stability of the interface; optimizing the decision support interface to improve user experience and decision efficiency;
s5.3, system integration and test: integrating all modules of the system together to form a complete system, and testing the system;
s5.4, preparing a deployment environment: preparing a deployment environment comprising a hardware environment and a software environment;
s5.5, system deployment: deploying the system into a designated environment;
s5.6, system maintenance: monitoring the system and knowing the running condition of the system; maintaining the system, including defect repair and function optimization; the system is upgraded to support new requirements.
10. An electricity market quotation decision-making system based on electricity price prediction for implementing the electricity market quotation decision-making method according to any one of the preceding claims 1-9, comprising:
and a data acquisition module: the system can integrate transaction center data in an automatic/manual mode, comprises meteorological data, thermal power overhaul information, cost data, contract data, new energy power prediction data, unit load data, fuel price information, market supply and demand data, power grid constraint and the like disclosed in the whole network, and combines a market transaction policy rule file to develop data unified management for long-term, spot and auxiliary service market conditions in a cross-province cross-district, long-term, spot and auxiliary service market conditions in a cross-province;
and a data processing module: the system performs basic information analysis of the electric power market, including analysis of supply and demand, network conditions, electricity price, power generation capacity, contract execution and operation conditions, compares predicted and actual data and operation date data, performs trend deviation analysis and visual display, establishes a data management system, selects a model algorithm, and establishes an electricity price prediction model;
the quotation decision module: using machine learning and data mining to predict and analyze electricity price, establishing a quotation decision model, analyzing game side behaviors, providing decision suggestions for medium-long term market and spot market transactions, and guiding spot market declaration and medium-long term bin adjustment;
and a visualization supporting module: and establishing a visual and decision support interface, providing types such as line graph display of electricity price trend analysis, histogram display of winning capacity, pie chart display of bin comparison and the like, and providing a large-screen cockpit interface for displaying key indexes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410205139.3A CN117788082A (en) | 2024-02-26 | 2024-02-26 | Power market quotation decision method and system based on electricity price prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410205139.3A CN117788082A (en) | 2024-02-26 | 2024-02-26 | Power market quotation decision method and system based on electricity price prediction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117788082A true CN117788082A (en) | 2024-03-29 |
Family
ID=90392852
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410205139.3A Pending CN117788082A (en) | 2024-02-26 | 2024-02-26 | Power market quotation decision method and system based on electricity price prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117788082A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112001744A (en) * | 2020-07-28 | 2020-11-27 | 安徽电力交易中心有限公司 | Power generator auxiliary quotation system and method based on prospect theory in electric power spot market |
| CN116777616A (en) * | 2023-06-13 | 2023-09-19 | 黑龙江龙源新能源发展有限公司 | Probability density distribution-based future market new energy daily transaction decision method |
| CN117114449A (en) * | 2023-09-13 | 2023-11-24 | 贵州电网有限责任公司 | Visual analysis system and method for big electric power data |
| CN117151345A (en) * | 2023-10-30 | 2023-12-01 | 智唐科技(北京)股份有限公司 | Enterprise management intelligent decision platform based on AI technology |
| CN117237054A (en) * | 2023-09-27 | 2023-12-15 | 特变电工新疆新能源股份有限公司 | Recommendation method, system and equipment for long-term centralized competitive price transaction strategy in electric power market |
-
2024
- 2024-02-26 CN CN202410205139.3A patent/CN117788082A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112001744A (en) * | 2020-07-28 | 2020-11-27 | 安徽电力交易中心有限公司 | Power generator auxiliary quotation system and method based on prospect theory in electric power spot market |
| CN116777616A (en) * | 2023-06-13 | 2023-09-19 | 黑龙江龙源新能源发展有限公司 | Probability density distribution-based future market new energy daily transaction decision method |
| CN117114449A (en) * | 2023-09-13 | 2023-11-24 | 贵州电网有限责任公司 | Visual analysis system and method for big electric power data |
| CN117237054A (en) * | 2023-09-27 | 2023-12-15 | 特变电工新疆新能源股份有限公司 | Recommendation method, system and equipment for long-term centralized competitive price transaction strategy in electric power market |
| CN117151345A (en) * | 2023-10-30 | 2023-12-01 | 智唐科技(北京)股份有限公司 | Enterprise management intelligent decision platform based on AI technology |
Non-Patent Citations (4)
| Title |
|---|
| 朱国荣等: "电力现货市场环境下的火电厂定价策略研究—基于短期竞价博弈模型的分析", 价格理论与实践, no. 6, 25 June 2020 (2020-06-25), pages 92 - 96 * |
| 杨周行: "电力现货市场下发电企业的竞价策略研究", 中国优秀硕士学位论文全文数据库工程科技Ⅱ辑, no. 1, 15 January 2023 (2023-01-15), pages 042 - 139 * |
| 杨尚宝;: "电力竞价上网智能决策支持系统", 电力自动化设备, no. 11, 25 November 2006 (2006-11-25), pages 105 - 107 * |
| 蔡鸿明等: "互联网时代的软件工程", vol. 1, 30 November 2021, 上海交通大学出版社, pages: 24 - 27 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11699198B2 (en) | Methods and systems for machine-learning for prediction of grid carbon emissions | |
| Hong et al. | Probabilistic electric load forecasting: A tutorial review | |
| Tsao et al. | Toward blockchain-based renewable energy microgrid design considering default risk and demand uncertainty | |
| US10873209B2 (en) | System and method for dynamic energy storage system control | |
| CN107845022B (en) | Electric power market aid decision-making system | |
| CN111582911A (en) | Multi-user and power grid friendly interactive power utilization system | |
| Li et al. | Monthly Henry Hub natural gas spot prices forecasting using variational mode decomposition and deep belief network | |
| Danish et al. | Shaping the future of sustainable energy through AI-enabled circular economy policies | |
| Jin et al. | Impact of cost–benefit analysis on financial benefit evaluation of investment projects under back propagation neural network | |
| Huang et al. | A control strategy based on deep reinforcement learning under the combined wind-solar storage system | |
| Li et al. | Machine learning-driven virtual bidding with electricity market efficiency analysis | |
| CN114943565A (en) | Electric power spot price prediction method and device based on intelligent algorithm | |
| Jeyaraj et al. | Computer‐assisted demand‐side energy management in residential smart grid employing novel pooling deep learning algorithm | |
| Lai et al. | Risk hedging for gas power generation considering power-to-gas energy storage in three different electricity markets | |
| CN117010946A (en) | Thermal power plant production and operation cost accounting system and application method thereof | |
| Li et al. | Multi-objective unit commitment optimization with ultra-low emissions under stochastic and fuzzy uncertainties | |
| Beichter et al. | Net load forecasting using different aggregation levels | |
| Dong et al. | Trading performance evaluation for traditional power generation group based on an integrated matter-element extension cloud model | |
| Al-Ani et al. | A study of load demand forecasting models in electricity using artificial neural networks and fuzzy logic model | |
| Liu et al. | A novel electricity load forecasting based on probabilistic least absolute shrinkage and selection operator-Quantile regression neural network | |
| Đukanović et al. | A machine learning approach for time series forecasting with application to debt risk of the Montenegrin electricity industry | |
| Chen et al. | Toward future information market: An information valuation paradigm | |
| CN116777616A (en) | Probability density distribution-based future market new energy daily transaction decision method | |
| CN117788082A (en) | Power market quotation decision method and system based on electricity price prediction | |
| Alikhani et al. | Optimal demand response programs selection using CNN‐LSTM algorithm with big data analysis of load curves |
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 |