WO2021077977A1 - 风电竞价场内交易数据分析方法、装置、设备及介质 - Google Patents

风电竞价场内交易数据分析方法、装置、设备及介质 Download PDF

Info

Publication number
WO2021077977A1
WO2021077977A1 PCT/CN2020/117295 CN2020117295W WO2021077977A1 WO 2021077977 A1 WO2021077977 A1 WO 2021077977A1 CN 2020117295 W CN2020117295 W CN 2020117295W WO 2021077977 A1 WO2021077977 A1 WO 2021077977A1
Authority
WO
WIPO (PCT)
Prior art keywords
vector
historical
wind power
market
data
Prior art date
Application number
PCT/CN2020/117295
Other languages
English (en)
French (fr)
Inventor
郭映军
Original Assignee
华能大理风力发电有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 华能大理风力发电有限公司 filed Critical 华能大理风力发电有限公司
Publication of WO2021077977A1 publication Critical patent/WO2021077977A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • G06Q30/0206Price or cost determination based on market factors
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06Electricity, gas or water supply
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04SSYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04SSYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S50/00Market activities related to the operation of systems integrating technologies related to power network operation or related to communication or information technologies
    • Y04S50/14Marketing, i.e. market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards

Definitions

  • the present invention relates to the technical field of wind power bidding, in particular to a method, device, equipment and medium for analyzing data in a wind power bidding field.
  • the bidding decision refers to the process in which the seller or the buyer who organizes the transaction through the market operation agency (or power trading center) participates in the market bidding, and determines the transaction volume and its price in a competitive manner.
  • the present invention proposes a method, device, equipment and medium for analyzing data in the field of wind power bidding, which aims to solve the problem that the prior art cannot analyze historical data, predict volume and price, and assist in bidding decision-making.
  • Technical issues are mainly responsible for solving the problem that the prior art cannot analyze historical data, predict volume and price, and assist in bidding decision-making.
  • the present invention provides a method for analyzing data in the field of wind power bidding.
  • the method for analyzing data of the field of wind power bidding includes the following steps:
  • S1 obtain the historical raw data of the market, and establish a vector based on the historical raw data of the market;
  • the historical raw data of the market includes historical quotations, historical transaction prices, and historical settlement prices of both bidders.
  • step S1 the historical raw data of the market is obtained, and the vector is established based on the historical raw data of the market, and the following steps are further included: the vector is established based on the historical raw data of the market:
  • Y represents the vector established based on the historical raw data of the market
  • v max represents the maximum value of historical quotations
  • v min represents the minimum value of historical quotations
  • p aver represents the average of historical transaction prices
  • h aver represents the average of historical settlement prices value.
  • step S2 a maximum-minimum difference method is established, and the historical original data is calculated by the maximum-minimum difference method to obtain the calculation vector.
  • the method also includes the following steps: In the maximum-minimum difference method, the vector is normalized according to the maximum-minimum difference method to obtain the normalized vector.
  • x i (j) [y i (j)-m(j)]/[M(j)-m(j)];
  • x i (j) is the j-th component after normalization
  • y i (j) is the component before normalization
  • m(j) is the minimum value of the j-th component in the sample
  • M(j) is the sample The maximum value of the j-th component in.
  • step S3 a neural network algorithm is established, the predicted value is calculated through the neural network algorithm, and the calculated result is obtained as the predicted electricity price, and the following steps are further included.
  • the neural network algorithm is:
  • f'(x) is the predicted electricity price
  • Opj is the actual output value of the mode p node j
  • f(x) is the threshold function
  • the threshold function f(x) is:
  • e- x is the correction value set by the system.
  • the device for analyzing data on the field of wind power bidding includes:
  • the acquisition module is used to acquire the historical raw data of the market and establish a vector based on the historical raw data of the market;
  • the first calculation module is used to establish the maximum-minimum difference method, calculate the historical original data through the maximum-minimum difference method, obtain the calculation vector, obtain the market forecast vector, and perform the calculation between the forecast vector and the calculation vector Compare, when the prediction vector meets the range of the calculation vector, use the prediction vector as the prediction value;
  • the second calculation module is used to establish a neural network algorithm, calculate the predicted value through the neural network algorithm, and obtain the calculated result as the predicted electricity price.
  • the method for analyzing data in the wind power bidding field further includes a device, the device comprising: a memory, a processor, and a wind power bidding field stored on the memory and running on the processor
  • a transaction data analysis method program which is configured to implement the steps of the above-mentioned method for analyzing wind power bidding on-site transaction data.
  • the method for analyzing data in the wind power bidding field further includes a medium, the medium is a computer medium, and a method program for analyzing data of the wind power bidding in the wind power bidding field is stored on the computer medium.
  • the transaction data analysis method program is executed by the processor, the steps of the above-mentioned wind power bidding field transaction data analysis method are realized.
  • the method for analyzing data in the field of wind power bidding of the present invention has the following beneficial effects:
  • the predicted value is calculated through the neural network, and the calculated result is obtained as the predicted electricity price. In this way, the electricity price can be quickly predicted, which improves work efficiency and saves resources.
  • FIG. 1 is a schematic diagram of a structure of a device in a hardware operating environment involved in a solution of an embodiment of the present invention
  • FIG. 2 is a schematic flow chart of the first embodiment of the method for analyzing data in the field of wind power bidding according to the present invention
  • FIG. 3 is a schematic diagram of functional modules of the first embodiment of the method for analyzing the transaction data in the wind power bidding field according to the present invention.
  • the device may include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (for example, a wireless fidelity (WI-FI) interface).
  • WI-FI wireless fidelity
  • the memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory.
  • RAM Random Access Memory
  • NVM Non-Volatile Memory
  • the memory 1005 may also be a storage device independent of the aforementioned processor 1001.
  • FIG. 1 does not constitute a limitation on the device.
  • the device may include more or fewer components than those shown in the figure, or combine certain components, or different components. Layout.
  • the memory 1005 as a medium may include an operating system, a network communication module, a user interface module, and a method and program for analyzing transaction data in a wind power bidding field.
  • the network interface 1004 is mainly used to establish a communication connection between the device and a server that stores all the data required in the system for analyzing method data of on-site transactions of wind power bidding;
  • the user interface 1003 is mainly used to communicate with users.
  • the processor 1001 and the memory 1005 in the wind power bidding field transaction data analysis method device of the present invention can be set in the wind power bidding field transaction data analysis method device, and the wind power bidding field transaction data analysis method device passes the processor 1001
  • the program of the method for analyzing data on the field of wind power bidding stored in the memory 1005 is called, and the method for analyzing the method for data on the field of wind power bidding provided by the implementation of the invention is executed.
  • Fig. 2 is a schematic flow chart of the first embodiment of the method for analyzing the data in the field of wind power bidding according to the present invention.
  • the method for analyzing on-site transaction data of wind power bidding includes the following steps:
  • S10 Obtain the historical raw data of the market, and establish a vector based on the historical raw data of the market.
  • the system will automatically obtain some historical raw data of the market. These historical raw data include the historical quotations, historical transaction prices, and historical settlement prices of both bidders. Then, the corresponding vector will be established based on these historical raw data. The vector will cover All the original data, in this way, after the original data is built into a vector, it is conducive to subsequent calculations.
  • Y represents the vector established based on the historical raw data of the market
  • v max represents the maximum value of historical quotations
  • v min represents the minimum value of historical quotations
  • p aver represents the average of historical transaction prices
  • h aver represents the average of historical settlement prices value.
  • S20 Establish a maximum-minimum difference method, calculate the historical raw data through the maximum-minimum difference method, obtain the calculation vector, obtain the prediction vector of the market, and compare the prediction vector with the calculation vector. When the prediction vector When the range of the calculation vector is satisfied, the predicted vector is used as the predicted value.
  • normalization is a way to simplify calculations, that is, a dimensional expression is transformed into a non-dimensional expression and becomes a scalar.
  • the normalization formula is:
  • x i (j) [y i (j)-m(j)]/[M(j)-m(j)];
  • x i (j) is the j-th component after normalization
  • y i (j) is the component before normalization
  • m(j) is the minimum value of the j-th component in the sample
  • M(j) is the sample The maximum value of the j-th component in.
  • x j [x j (1), x j (2), x j (3), x j (4)].
  • the system will obtain forecast data for the market. These forecast data include the predicted quotations of the bidders, the predicted transaction price, and the predicted settlement price. After obtaining the data, the system will organize these data and establish a forecast vector. , And use the similarity discrimination formula to determine the similarity between the predicted vector and the vector obtained by historical data technology. The higher the similarity, the more accurate the predicted data.
  • t represents the forecast date
  • i represents the i-th sample
  • j represents the j-th component in the vector.
  • neural network is a parallel and distributed information processing network structure, it is generally composed of many neurons, each neuron has only one output, it can be connected to many other neurons , Each neuron input has many connection channels, and each connection channel corresponds to a connection weight coefficient.
  • a multilayer neural network model is divided into three layers: input layer, input layer and middle layer.
  • N be the number of input layer units
  • L be the number of middle layer units
  • M be the number of output layer units.
  • the middle layer does not want to connect with the actual input and output, so it is also called the hidden layer.
  • the model is a three-layer network structure, the number of input layer units is N; the number of hidden layer units is L>2N; the number of output layer units is M.
  • the error function E p is in full network mode p, defined error function E p for each node and the squared difference between the desired value and the actual output value output:
  • t pj is the desired output value of node j in mode p
  • Opj is the actual output value of node j in mode p.
  • e- x is the correction value set by the system.
  • Derivation of the above function can get the function of forecasting electricity price.
  • the forecasted electricity price can be obtained quickly.
  • Opj is the actual output value of mode p node j.
  • this embodiment obtains historical raw data of the market and establishes a vector based on the historical raw data of the market; establishes the maximum-minimum difference method, and performs the historical raw data through the maximum-minimum difference method.
  • Calculate obtain the calculation vector, obtain the prediction vector of the market, compare the prediction vector with the calculation vector, when the prediction vector meets the range of the calculation vector, use the prediction vector as the prediction value; establish a neural network algorithm, and use the neural network algorithm to The predicted value is calculated, and the calculated result is obtained as the predicted electricity price.
  • the present invention predicts the required data by obtaining the original market historical data and establishing the maximum-minimum difference algorithm in the similar day method to predict the required data. After the required data is obtained, the predicted data is calculated by building a neural network algorithm to obtain Forecast electricity prices to assist in bidding decisions.
  • the embodiment of the present invention also provides a device for analyzing data in the field of wind power bidding.
  • the device for analyzing transaction data in the wind power bidding field includes: an acquisition module 10, a first calculation module 20, and a second calculation module 30.
  • the obtaining module 10 is used to obtain historical raw data of the market, and establish a vector based on the historical raw data of the market;
  • the first calculation module 20 is used to establish a maximum-minimum difference method, calculate the historical original data through the maximum-minimum difference method, obtain a calculation vector, obtain a market forecast vector, and combine the forecast vector with the calculation vector For comparison, when the prediction vector meets the range of the calculation vector, the prediction vector is used as the prediction value;
  • the second calculation module 30 is used to establish a neural network algorithm, calculate the predicted value through the neural network algorithm, and obtain the calculated result as the predicted electricity price.
  • the embodiment of the present invention also provides a medium, the medium is a computer medium, the computer medium stores a wind power bidding on-site transaction data analysis method program, and the wind power bidding on-site transaction data analysis method program is processed by a processor.
  • the following operations are implemented during execution:
  • S1 obtain the historical raw data of the market, and establish a vector based on the historical raw data of the market;
  • the historical raw data of the market includes historical quotations, historical transaction prices, and historical settlement prices of both parties.
  • Y represents the vector established based on the historical raw data of the market
  • v max represents the maximum value of historical quotations
  • v min represents the minimum value of historical quotations
  • p aver represents the average of historical transaction prices
  • h aver represents the average of historical settlement prices value.
  • the maximum-minimum difference method is established, and the vector is normalized according to the maximum-minimum difference method to obtain the normalized vector.
  • the normalization formula is:
  • x i (j) [y i (j)-m(j)]/[M(j)-m(j)];
  • x i (j) is the j-th component after normalization
  • y i (j) is the component before normalization
  • m(j) is the minimum value of the j-th component in the sample
  • M(j) is the sample The maximum value of the j-th component in.
  • the neural network algorithm is:
  • f'(x) is the predicted electricity price
  • Opj is the actual output value of the mode p node j
  • f(x) is the threshold function
  • the threshold function f(x) is:
  • e- x is the correction value set by the system.

Abstract

一种风电竞价场内交易数据分析方法、装置、设备及介质。方法包括:获取市场的历史原始数据,根据市场的历史原始数据建立向量(S10);建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值(S20);建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价(S30)。上述方法通过获取市场历史原始数据,通过建立相似日法中的极大极小差值算法来预测需要的数据,在得到需要的数据之后,通过搭建神经网络算法,对预测的数据进行计算,得到预测电价,为竞价决策做辅助。

Description

风电竞价场内交易数据分析方法、装置、设备及介质 技术领域
本发明涉及风电竞价技术领域,尤其涉及一种风电竞价场内交易数据分析方法、装置、设备及介质。
背景技术
近年来,全球环境问题与能源问题日益突出,可再生能源在国内外受到广泛的关注并迅速发展,随着风电技术的成熟和成本的下降,为了实现市场优化资源配置,风电商需要参与到市场竞价中,竞价决策因此而来,竞价决策是指通过市场运营机构(或电力交易中心)组织交易的卖方或买方参与市场投标,以竞争方式确定交易量以及其价格的过程。
但是,现有的竞价场内交易数据的展示方式过于简单,仅仅只通过图表、曲线以及柱状图来展示当月的数据,同时,也没有对历史数据进行分析和预测,所以,如何对历史数据进行分析,对量价进行预测,为竞价决策做出辅助成为了一个亟待解决的问题。
上述内容仅用于辅助理解本发明的技术方案,并不代表承认上述内容是现有技术。
发明内容
有鉴于此,本发明提出了一种风电竞价场内交易数据分析方法、装置、设备及介质,旨在解决现有技术无法对历史数据进行分析,对量价进行预测,为竞价决策做出辅助的技术问题。
本发明的技术方案是这样实现的:
一方面,本发明提供了一种风电竞价场内交易数据分析方法,所述风电竞价场内交易数据分析方法包括以下步骤:
S1,获取市场的历史原始数据,根据市场的历史原始数据建立向量;
S2,建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
S3,建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
在以上技术方案的基础上,优选的,步骤S1中,市场的历史原始数据包括竞价双方的历史报价、历史成交价格、历史结算价格。
在以上技术方案的基础上,优选的,步骤S1中,获取市场的历史原始数据,根据市场的历史原始数据建立向量,还包括以下步骤,根据市场的历史原始数据建立向量:
Y=[v max,v min,p aver,h aver];
其中,Y表示根据市场的历史原始数据建立的向量,v max代表历史报价的最大值,v min代表历史报价的最小值,p aver代表历史成交价格的平均值,h aver代表历史结算价格的平均值。
在以上技术方案的基础上,优选的,步骤S2中,建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,还包括以下步骤,建立极大极小差值法,根据极大极小差值法对该向量进行归一化,获取归化后的向量。
在以上技术方案的基础上,优选的,还包括以下步骤,归一化公式为:
x i(j)=[y i(j)-m(j)]/[M(j)-m(j)];
其中,x i(j)为归化后的第j个分量,y i(j)为归化前的分量,m(j)为样本中第j个分量的最小值,M(j)为样本中第j个分量的最大值。
在以上技术方案的基础上,优选的,步骤S3中,建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价,还包括以下步骤,神经网络算法为:
Figure PCTCN2020117295-appb-000001
其中,f'(x)为预测电价,O pj为模式p节点j的实际输出值,f(x)为阈函数。
在以上技术方案的基础上,优选的,还包括以下步骤,阈函数f(x)为:
f(x)=1/(1+e -x);
其中,e -x为系统设定的修正值。
更进一步优选的,所述风电竞价场内交易数据分析装置包括:
获取模块,用于获取市场的历史原始数据,根据市场的历史原始数据建立向量;
第一计算模块,用于建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
第二计算模块,用于建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
第二方面,所述风电竞价场内交易数据分析方法还包括一种设备,所述设备包括:存储器、处理器以及存储在所述存储器上并可在所述处理器上运行的风电竞价场内交易数据分析方法程序,所述风电竞价场内交易数据分析方法程序配置为实现如上文所述的风电竞价场内交易数据分析方法的步骤。
第三方面,所述风电竞价场内交易数据分析方法还包括一种介质,所述介质为计算机介质,所述计算机介质上存储有风电竞价场内交易数据分析方法程序,所述风电竞价场内交易数据分析方法程序被处理器执行时实现如上文所述的风电竞价场内交易数据分析方法的步骤。
本发明的一种风电竞价场内交易数据分析方法相对于现有技术具有以下有益效果:
(1)通过相似日法中的极大极小差值法对市场的历史原始数据进行计算,获取计算后的向量,将该向量作为对比标准,获取系统对市场数据的预测向量, 将该预测向量与计算后的向量进行比较,确定满足需求的预测值,通过这种方式,可以自动且准确对预测值进行判断,提高了竞价效率;
(2)通过建立神经网络算法,通过神经网络对预测值进行计算,获取计算后的结果作为预测电价,通过这种方式可以快速对电价进行预测,提高了工作效率,同时也节省了资源。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1是本发明实施例方案涉及的硬件运行环境的设备的结构示意图;
图2为本发明风电竞价场内交易数据分析方法第一实施例的流程示意图;
图3为本发明风电竞价场内交易数据分析方法第一实施例的功能模块示意图。
具体实施方式
下面将结合本发明实施方式,对本发明实施方式中的技术方案进行清楚、完整地描述,显然,所描述的实施方式仅仅是本发明一部分实施方式,而不是全部的实施方式。基于本发明中的实施方式,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施方式,都属于本发明保护的范围。
如图1所示,该设备可以包括:处理器1001,例如中央处理器(Central Processing Unit,CPU),通信总线1002、用户接口1003,网络接口1004,存储器1005。其中,通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard),可选用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口(如无线保真(WIreless-FIdelity,WI-FI)接口)。存储器1005可以是高速的随机存取存储器(Random Access Memory,RAM)存 储器,也可以是稳定的非易失性存储器(Non-Volatile Memory,NVM),例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解,图1中示出的结构并不构成对设备的限定,在实际应用中设备可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
如图1所示,作为一种介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及风电竞价场内交易数据分析方法程序。
在图1所示的设备中,网络接口1004主要用于建立设备与存储风电竞价场内交易数据分析方法系统中所需的所有数据的服务器的通信连接;用户接口1003主要用于与用户进行数据交互;本发明风电竞价场内交易数据分析方法设备中的处理器1001、存储器1005可以设置在风电竞价场内交易数据分析方法设备中,所述风电竞价场内交易数据分析方法设备通过处理器1001调用存储器1005中存储的风电竞价场内交易数据分析方法程序,并执行本发明实施提供的风电竞价场内交易数据分析方法。
结合图2,图2为本发明风电竞价场内交易数据分析方法第一实施例的流程示意图。
本实施例中,所述风电竞价场内交易数据分析方法包括以下步骤:
S10:获取市场的历史原始数据,根据市场的历史原始数据建立向量。
应当理解的是,系统会自动获取市场的一些历史原始数据,这些历史原始数据包括竞价双方的历史报价、历史成交价格、历史结算价格,然后会根据这些历史原始数据建立对应的向量,向量会涵盖所有的原始数据,通过这样的方式,将原始数据建立向量后,有利于后续的计算。
应当理解的是,建立的向量为:
Y=[v max,v min,p aver,h aver];
其中,Y表示根据市场的历史原始数据建立的向量,v max代表历史报价的最大值,v min代表历史报价的最小值,p aver代表历史成交价格的平均值,h aver代表历 史结算价格的平均值。
S20:建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值。
应当理解的是,建立极大极小差值法,通过极大极小差值法对原始数据建立的向量进行归一化,用于后续进行相似度的判断。
应当理解的是,归一化是一种简化计算的方式,即将有量纲的表达式,经过变换,化为无量纲的表达式,成为标量。归一化公式为:
x i(j)=[y i(j)-m(j)]/[M(j)-m(j)];
其中,x i(j)为归化后的第j个分量,y i(j)为归化前的分量,m(j)为样本中第j个分量的最小值,M(j)为样本中第j个分量的最大值。
应当理解的是,对于本实施例中的向量,一共有4个分量,所以本实施例中的向量其规划后的第j向量可表示为:
x j=[x j(1),x j(2),x j(3),x j(4)]。
应当理解的是,系统会获取对市场的预测数据,这些预测数据包括竞价双方的预测报价、预测成交价格以及预测结算价格,在获取数据之后,系统会将这些数据整理起来,同样建立一个预测向量,并通过相似度判别公式来判别预测向量与历史数据技术得到的向量的相似度,相似度越高,表示预测的数据越准确。
应当理解的是,相似度判别公式为:
Figure PCTCN2020117295-appb-000002
其中,t表示预测日,i表示第i个样本,j表示向量中的第j个分量。
S30:建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
应当理解的是,建立神经网络算法,神经网络是一个并行和分布式的信息处理网络结构,它一般由许多个神经元组成,每个神经元只有一个输出,它可以连接到很多其他的神经元,每个神经元输入有很多个连接通道,每个连接通道对应于一个连接权系数。
应当理解的是,一个多层神经网络模型分为三层:输入层、输入层和中间层。设N为输入层单元数,L为中间层单元个数,M为输出层单元个数。中间层不与实际的输入输出想连接,故又称隐含层。
应当理解的是,本实施例中,通过构建神经网络算法对预测值进行计算的步骤为:
设模型为三层网络结构,输入层单元数为N个;隐含层单元数为L>2N个;输出层单元数为M个。设E p为模式p下全网络的误差函数,定义误差函数E p为各节点希望输出值与实际输出值之差的平方和:
Figure PCTCN2020117295-appb-000003
其中,t pj为模式p下节点j的希望输出值;O pj为模式p下节点j的实际输出值。
取S型函数作为阈函数(S的意义为修正系数),则有:
f(x)=1/(1+e -x);
其中,e -x为系统设定的修正值。
对上述函数求导即可以得到预测电价的函数,采用三层神经网络建立每个报价时段的预测模型,既可以快速获取预测电价。
Figure PCTCN2020117295-appb-000004
其中,O pj为模式p节点j的实际输出值。
需要说明的是,以上仅为举例说明,并不对本申请的技术方案构成任何限定。
通过上述描述不难发现,本实施例通过获取市场的历史原始数据,根据市场的历史原始数据建立向量;建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。本发明通过获取市场历史原始数据,通过建立相似日法中的极大极小差值算法来预测需要的数据,在得到需要的数据之后,通过搭建神经网络算法,对预测的数据进行计算,得到预测电价,为竞价决策做辅助。
此外,本发明实施例还提出一种风电竞价场内交易数据分析装置。如图3所示,该风电竞价场内交易数据分析装置包括:获取模块10、第一计算模块20、第二计算模块30。
获取模块10,用于获取市场的历史原始数据,根据市场的历史原始数据建立向量;
第一计算模块20,用于建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
第二计算模块30,用于建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
此外,需要说明的是,以上所描述的装置实施例仅仅是示意性的,并不对本发明的保护范围构成限定,在实际应用中,本领域的技术人员可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的,此处不做限制。
另外,未在本实施例中详尽描述的技术细节,可参见本发明任意实施例所提供的风电竞价场内交易数据分析方法,此处不再赘述。
此外,本发明实施例还提出一种介质,所述介质为计算机介质,所述计算 机介质上存储有风电竞价场内交易数据分析方法程序,所述风电竞价场内交易数据分析方法程序被处理器执行时实现如下操作:
S1,获取市场的历史原始数据,根据市场的历史原始数据建立向量;
S2,建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
S3,建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
市场的历史原始数据包括竞价双方的历史报价、历史成交价格、历史结算价格。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
根据市场的历史原始数据建立向量:
Y=[v max,v min,p aver,h aver];
其中,Y表示根据市场的历史原始数据建立的向量,v max代表历史报价的最大值,v min代表历史报价的最小值,p aver代表历史成交价格的平均值,h aver代表历史结算价格的平均值。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
建立极大极小差值法,根据极大极小差值法对该向量进行归一化,获取归化后的向量。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
归一化公式为:
x i(j)=[y i(j)-m(j)]/[M(j)-m(j)];
其中,x i(j)为归化后的第j个分量,y i(j)为归化前的分量,m(j)为样本中第j个分量的最小值,M(j)为样本中第j个分量的最大值。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
神经网络算法为:
Figure PCTCN2020117295-appb-000005
其中,f'(x)为预测电价,O pj为模式p节点j的实际输出值,f(x)为阈函数。
进一步地,所述风电竞价场内交易数据分析方法程序被处理器执行时还实现如下操作:
阈函数f(x)为:
f(x)=1/(1+e -x);
其中,e -x为系统设定的修正值。
以上所述仅为本发明的较佳实施方式而已,并不用以限制本发明,凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。

Claims (10)

  1. 一种风电竞价场内交易数据分析方法,其特征在于:包括以下步骤;
    S1,获取市场的历史原始数据,根据市场的历史原始数据建立向量;
    S2,建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
    S3,建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
  2. 如权利要求1所述的风电竞价场内交易数据分析方法,其特征在于:步骤S1中,市场的历史原始数据包括竞价双方的历史报价、历史成交价格、历史结算价格。
  3. 如权利要求2所述的风电竞价场内交易数据分析方法,其特征在于:步骤S1中,获取市场的历史原始数据,根据市场的历史原始数据建立向量,还包括以下步骤,根据市场的历史原始数据建立向量:
    Y=[v max,v min,p aver,h aver];
    其中,Y表示根据市场的历史原始数据建立的向量,v max代表历史报价的最大值,v min代表历史报价的最小值,p aver代表历史成交价格的平均值,h aver代表历史结算价格的平均值。
  4. 如权利要求3所述的风电竞价场内交易数据分析方法,其特征在于:步骤S2中,建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,还包括以下步骤,建立极大极小差值法,根据极大极小差值法对该向量进行归一化,获取归化后的向量。
  5. 如权利要求4所述的风电竞价场内交易数据分析方法,其特征在于:还包括以下步骤,归一化公式为:
    x i(j)=[y i(j)-m(j)]/[M(j)-m(j)];
    其中,x i(j)为归化后的第j个分量,y i(j)为归化前的分量,m(j)为样本中第j个分量的最小值,M(j)为样本中第j个分量的最大值。
  6. 如权利要求1所述的风电竞价场内交易数据分析方法,其特征在于:步骤S3中,建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价,还包括以下步骤,神经网络算法为:
    Figure PCTCN2020117295-appb-100001
    其中,f'(x)为预测电价,O pj为模式p节点j的实际输出值,f(x)为阈函数。
  7. 如权利要求6所述的风电竞价场内交易数据分析方法,其特征在于:还包括以下步骤,阈函数f(x)为:
    f(x)=1/(1+e -x);
    其中,e -x为系统设定的修正值。
  8. 一种风电竞价场内交易数据分析装置,其特征在于,所述风电竞价场内交易数据分析装置包括:
    获取模块,用于获取市场的历史原始数据,根据市场的历史原始数据建立向量;
    第一计算模块,用于建立极大极小差值法,通过极大极小差值法对该历史原始数据进行计算,获取计算向量,获取市场的预测向量,将该预测向量与计算向量进行比较,当预测向量满足计算向量的范围时,将该预测向量作为预测值;
    第二计算模块,用于建立神经网络算法,通过神经网络算法对预测值进行计算,获取计算的结果作为预测电价。
  9. 一种设备,其特征在于,所述设备包括:存储器、处理器以及存储在所述存储器上并可在所述处理器上运行的风电竞价场内交易数据分析方法程序,所述风电竞价场内交易数据分析方法程序配置为实现如权利要求1至7任一项所述的风电竞价场内交易数据分析方法的步骤。
  10. 一种介质,其特征在于,所述介质为计算机介质,所述计算机介质上存储有风电竞价场内交易数据分析方法程序,所述风电竞价场内交易数据分析方 法程序被处理器执行时实现如权利要求1至7任一项所述的风电竞价场内交易数据分析方法的步骤。
PCT/CN2020/117295 2019-10-23 2020-09-24 风电竞价场内交易数据分析方法、装置、设备及介质 WO2021077977A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911014039.8A CN110826777A (zh) 2019-10-23 2019-10-23 风电竞价场内交易数据分析方法、装置、设备及介质
CN201911014039.8 2019-10-23

Publications (1)

Publication Number Publication Date
WO2021077977A1 true WO2021077977A1 (zh) 2021-04-29

Family

ID=69550320

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/117295 WO2021077977A1 (zh) 2019-10-23 2020-09-24 风电竞价场内交易数据分析方法、装置、设备及介质

Country Status (2)

Country Link
CN (1) CN110826777A (zh)
WO (1) WO2021077977A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110826777A (zh) * 2019-10-23 2020-02-21 华能大理风力发电有限公司 风电竞价场内交易数据分析方法、装置、设备及介质
CN113129109B (zh) * 2021-04-30 2024-01-26 深圳市点石成金科技有限公司 一种基于大数据的人工智能竞价算法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105117810A (zh) * 2015-09-24 2015-12-02 国网福建省电力有限公司泉州供电公司 一种阶梯电价机制下居民用电中期负载预测方法
EP3067840A1 (en) * 2015-03-13 2016-09-14 Hitachi, Ltd. Energy control unit, energy control system, and energy management method
CN106682934A (zh) * 2016-11-18 2017-05-17 云南电网有限责任公司电力科学研究院 一种购电竞价策略
CN109242532A (zh) * 2018-08-03 2019-01-18 广东工业大学 基于局域均值分解及优化rbf神经网络的短期电价预测方法
CN110135643A (zh) * 2019-05-17 2019-08-16 国网山东省电力公司莱芜供电公司 考虑钢铁期货价格和实时电价因素的短期日负荷预测方法
CN110826777A (zh) * 2019-10-23 2020-02-21 华能大理风力发电有限公司 风电竞价场内交易数据分析方法、装置、设备及介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3067840A1 (en) * 2015-03-13 2016-09-14 Hitachi, Ltd. Energy control unit, energy control system, and energy management method
CN105117810A (zh) * 2015-09-24 2015-12-02 国网福建省电力有限公司泉州供电公司 一种阶梯电价机制下居民用电中期负载预测方法
CN106682934A (zh) * 2016-11-18 2017-05-17 云南电网有限责任公司电力科学研究院 一种购电竞价策略
CN109242532A (zh) * 2018-08-03 2019-01-18 广东工业大学 基于局域均值分解及优化rbf神经网络的短期电价预测方法
CN110135643A (zh) * 2019-05-17 2019-08-16 国网山东省电力公司莱芜供电公司 考虑钢铁期货价格和实时电价因素的短期日负荷预测方法
CN110826777A (zh) * 2019-10-23 2020-02-21 华能大理风力发电有限公司 风电竞价场内交易数据分析方法、装置、设备及介质

Also Published As

Publication number Publication date
CN110826777A (zh) 2020-02-21

Similar Documents

Publication Publication Date Title
WO2021077977A1 (zh) 风电竞价场内交易数据分析方法、装置、设备及介质
CN110751326B (zh) 一种光伏日前功率预测方法、装置以及存储介质
CN106845777A (zh) 案件工作量评价方法、法官业绩评价方法、建模方法、计算装置和计算机可读存储介质
WO2021035975A1 (zh) 基于多评价维度的热点学科预测方法、装置、终端、及介质
WO2021068513A1 (zh) 异常对象识别方法、装置、介质及电子设备
US20220130495A1 (en) Method and Device for Determining Correlation Between Drug and Target, and Electronic Device
CN110796513A (zh) 多任务学习方法、装置、电子设备及存储介质
WO2023217127A1 (zh) 一种因果关系确定方法及相关设备
CN111144721A (zh) 电网项目需求评价模型的构建方法、装置和计算设备
CN111179055B (zh) 授信额度调整方法、装置和电子设备
CN115330435A (zh) 一种碳排放权价格指标体系建立方法、装置、设备及介质
Li et al. What can be learned from the historical trend of crude oil prices? An ensemble approach for crude oil price forecasting
CN111506876A (zh) 一种数据预测分析方法、系统、设备及可读存储介质
Li et al. Resource usage prediction based on BiLSTM-GRU combination model
CN114118570A (zh) 业务数据预测方法及装置、电子设备和存储介质
CN113361959A (zh) 一种银行业务集中运营成熟度的计算方法和装置
CN117114776A (zh) 一种面向省间日前现货交易的量价申报方法
CN111126714A (zh) 基于长租公寓租房场景下的退租预测系统及方法
Zhang et al. RSVRs based on feature extraction: a novel method for prediction of construction projects’ costs
CN115879975A (zh) 基于用电企业建设分布式光伏的电力零售套餐制定方法
CN114970357A (zh) 节能效果评价方法、系统、装置及存储介质
WO2022105137A1 (zh) 案件处理方法、装置、计算机设备和计算机可读存储介质
CN112632846B (zh) 一种电力系统输电断面极限概率评估方法及电子设备
CN111046321A (zh) 光伏电站运维策略优化方法和装置
CN113723712B (zh) 风电功率预测方法、系统、设备及介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20880122

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20880122

Country of ref document: EP

Kind code of ref document: A1