WO2014086226A1 - 电力负荷预测精度的检测方法及装置 - Google Patents

电力负荷预测精度的检测方法及装置 Download PDF

Info

Publication number
WO2014086226A1
WO2014086226A1 PCT/CN2013/086800 CN2013086800W WO2014086226A1 WO 2014086226 A1 WO2014086226 A1 WO 2014086226A1 CN 2013086800 W CN2013086800 W CN 2013086800W WO 2014086226 A1 WO2014086226 A1 WO 2014086226A1
Authority
WO
WIPO (PCT)
Prior art keywords
time series
historical load
load
frequency component
historical
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.)
Ceased
Application number
PCT/CN2013/086800
Other languages
English (en)
French (fr)
Inventor
王海云
周作春
袁清芳
孙健
高明伟
杨楠
杜晨红
刘慧珍
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Beijing Electric Power Co Ltd
State Grid Corp of China SGCC
Original Assignee
State Grid Beijing Electric Power Co Ltd
State Grid Corp of China SGCC
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 State Grid Beijing Electric Power Co Ltd, State Grid Corp of China SGCC filed Critical State Grid Beijing Electric Power Co Ltd
Publication of WO2014086226A1 publication Critical patent/WO2014086226A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; 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 OR CALCULATING; 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
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06Energy or water supply

Definitions

  • the present invention relates to the field of electric power, and in particular to a method and an apparatus for detecting electric power load prediction accuracy.
  • BACKGROUND OF THE INVENTION In the short-term load forecasting of power systems, the accuracy of load forecasting plays a decisive role in realizing the real-time balance of power systems, saving energy, and improving economic efficiency.
  • the load itself has a certain randomness.
  • the distribution of power grids and climatic conditions in various regions will directly affect the load changes. Therefore, the difficulty level of the load regularity is different in each region.
  • the load forecasting can be The precision assessment requirements achieved are based on a “one size fits all” approach.
  • the present invention provides a method and apparatus for detecting power load prediction accuracy, so as to at least solve the problem in the prior art that it is impossible to determine whether the power load prediction accuracy satisfies a preset condition according to local load regularity conditions.
  • a method of detecting an electric power load prediction accuracy is provided.
  • the method for detecting power load prediction accuracy includes: acquiring a historical load time series, wherein the time series is load data of a continuous time series; performing time-frequency transform on the data in the time series to obtain a frequency domain decomposition result; The domain decomposition results are combined periodically to obtain low-frequency components and high-frequency components.
  • the low-frequency components are related to meteorological factors, and the high-frequency components are caused by the random fluctuation of the load.
  • the historical load time series, low-frequency components, and high-frequency components are determined. Historical load stability; and whether the accuracy of the electric load prediction meets the preset condition based on the historical load stability.
  • acquiring the historical load time series includes acquiring a plurality of historical load time series, wherein the plurality of historical load time series are historical load time series corresponding to the plurality of different regions, and time-frequency transforming the data in the time series to obtain
  • the frequency domain decomposition results include: Performing a historical load time series in the preset modeling time domain
  • determining the historical load stability according to the historical load time series, the low frequency component, and the high frequency component comprises: determining an upper limit of the historical load stability according to the historical load time series and the high frequency component determining the historical load stability; and according to the historical load time series The low frequency component and the high frequency component determine the historical load stability to determine the lower limit of the historical load stability, wherein the historical load stability is a numerical range between the upper limit and the lower limit.
  • One step, determining the upper limit of historical load stability based on historical load time series and high frequency components includes:
  • determining whether the power load prediction accuracy satisfies the preset condition according to the historical load stability includes: calculating the power load prediction accuracy; determining whether the power load prediction accuracy is between the upper limit and the lower limit, wherein when the power load prediction accuracy is at the upper limit and the lower limit When the power load prediction accuracy is determined to meet the preset condition, when the power load prediction accuracy is below the lower limit, it is determined that the power load prediction accuracy does not satisfy the preset condition.
  • a device for detecting an accuracy of electric power load which is for performing a method for detecting an electric power load prediction accuracy provided by the present invention.
  • a detecting apparatus for electric power load prediction accuracy is provided.
  • the device for detecting the accuracy of the power load prediction includes: a first acquiring unit, configured to acquire a historical load time series, wherein the time series is load data of a continuous time series; and the second acquiring unit is configured to perform data in the time series Time-frequency transform to obtain frequency domain decomposition results; third acquisition unit, configured to perform frequency domain decomposition results according to periodicity And obtaining a low frequency component and a high frequency component, wherein the low frequency component is related to meteorological factors, the high frequency component is caused by random fluctuation of the load; the determining unit is configured to determine the historical load according to the historical load time series, the low frequency component and the high frequency component a stability unit; and a judging unit configured to judge whether the electric power load prediction accuracy satisfies a preset condition according to the historical load stability.
  • the obtaining the historical load time series includes acquiring a plurality of historical load time series, wherein the plurality of historical load time series are historical load time series corresponding to the plurality of different regions, and the second obtaining unit includes: a decomposition subunit, configured to: Perform the following Fourier decomposition on the historical load time series in the preset modeling time domain:
  • the determining unit includes: a first determining subunit, configured to determine an upper limit of the historical load stability according to the historical load time series and the high frequency component; and a second determining subunit configured to use the historical load time series, the low frequency component, and The high frequency component determines the lower limit of the historical load stability, wherein the historical load stability is a numerical range between the upper and lower limits.
  • the first determining subunit is further used to determine the upper limit of the historical load stability according to the following formula: ⁇ . Determine the calendar according to the following formula
  • the determining unit includes: a calculating subunit for calculating a power load prediction accuracy; and a determining subunit, configured to determine whether the power load prediction accuracy is between an upper limit and a lower limit, wherein when the electric load prediction accuracy is at an upper limit and a lower limit During the interval, it is determined that the power load prediction accuracy satisfies the preset condition, and when the power load prediction accuracy is below the lower limit, it is determined that the power load prediction accuracy does not satisfy the preset condition.
  • the time-frequency transform of the historical data is performed, and the low-frequency component and the high-frequency component are obtained, and the historical load stability is further obtained.
  • FIG. 1 is a block diagram showing a configuration of a power load prediction accuracy detecting apparatus according to an embodiment of the present invention
  • FIG. 2 is a flowchart of a power load prediction accuracy detecting method according to an embodiment of the present invention.
  • An embodiment of the present invention provides a device for detecting power load prediction accuracy.
  • the following describes an apparatus for detecting power load prediction accuracy according to an embodiment of the present invention.
  • 1 is a block diagram showing the structure of a device for detecting power load prediction accuracy according to an embodiment of the present invention. As shown in FIG. 1, the apparatus for detecting power load prediction accuracy includes a first acquisition unit 11, a second acquisition unit 12, a third acquisition unit 13, a determination unit 14, and a determination unit 15.
  • the first obtaining unit 11 is configured to acquire a historical load time series, wherein the time series is load data of a continuous time series.
  • the second obtaining unit 12 is configured to perform time-frequency transform on the data in the time series to obtain a frequency domain decomposition result.
  • the third obtaining unit 13 is configured to combine frequency domain decomposition results according to periodicity to obtain low frequency components and high frequency components, wherein the low frequency components are related to meteorological factors, that is, weather factors such as temperature and humidity generate images for low frequency components.
  • the high frequency components are caused by the random volatility of the load.
  • the determining unit 14 is configured to determine the historical load stability based on the historical load time series, the low frequency component, and the high frequency component.
  • the judging unit 15 is configured to judge whether the electric power load prediction accuracy satisfies the preset condition according to the historical load stability.
  • the time-frequency transform is performed on the historical data, and the high-frequency component and the low-frequency component are obtained, and the historical load stability is further obtained. Since the historical load stability has a relatively reliable regularity, it can be used as the power load prediction accuracy. Therefore, the problem of judging whether the accuracy of the electric power load prediction meets the preset condition according to the regional load regularity condition is solved in the prior art, and the corresponding electric power load prediction accuracy can be determined according to the situation of different regions. Since the requirements for the accuracy of the power load prediction in different regions are different, different regions are separately detected.
  • the acquired historical load time series includes acquiring a plurality of historical load time series, wherein the plurality of historical load time series correspond to a historical load time series of a plurality of different regions.
  • the second obtaining unit includes a decomposing subunit and an obtaining subunit, wherein the decomposing subunit is configured to perform a historical load time series in the preset modeling time domain as follows:
  • P(t) represents the historical load time series
  • t is time
  • a0, ai and bi are different in different regions
  • D(t;) represents the daily cycle load component that varies with the daily cycle
  • W ⁇ t) represents the weekly cycle load component that varies with the cycle of the week
  • L(t) represents the low-frequency component
  • H(t) represents the high-frequency component.
  • the historical load stability is determined by the high frequency component and the low frequency component, and the power load accuracy is determined according to the stability.
  • the determining unit includes the first determining subunit and the first And determining a subunit, wherein the first determining subunit is configured to determine an upper limit of the historical load stability according to the historical load time series and the high frequency component.
  • the second determining subunit is configured to determine a lower limit of the historical load stability according to the historical load time series, the low frequency component, and the high frequency component, wherein the historical load stability is a numerical range between the upper limit and the lower limit. Determining historical load stability through high frequency components and low frequency components is not only convenient, but also scientific and reasonable.
  • the first determining subunit is further configured to determine an upper limit of the historical load stability according to the following formula:
  • the second determining subunit is further configured to determine a lower limit of the historical load stability according to the following formula:
  • L uppCT is the upper limit of historical load stability
  • L 1()WCT is the lower limit of historical load stability.
  • the determining subunit is configured to determine whether the electric power load prediction accuracy is between an upper limit and a lower limit, wherein when the electric load prediction accuracy is between the upper limit and the lower limit, determining that the electric load prediction accuracy satisfies a preset condition may also be said to be qualified, When the power load prediction accuracy is below the lower limit, it is determined that the power load prediction accuracy does not satisfy the preset condition, that is, fails.
  • the calculation of the power load prediction accuracy is performed in various ways in the prior art, and is not described in detail in this embodiment.
  • Embodiments of the present invention also provide a method for detecting power load prediction accuracy, which may be performed based on the foregoing apparatus.
  • the method for detecting the accuracy of the power load prediction includes the following steps S202 to S21 (h step S202, acquiring a historical load time series, wherein the time series is a continuous time series of load data, the time series is continuous
  • the time series for example, the load data calculation is collected in 15 minutes, and the historical load time series of one day may be a continuous historical load time series arranged in chronological order by 96 load data.
  • Step S204 data in the time series Performing a time-frequency transform to obtain a frequency domain decomposition result, When calculating the accuracy of the electric load forecasting, the different regions may be separately calculated.
  • the acquired historical load time series includes acquiring a plurality of historical load time series, wherein the plurality of historical load time series are historical loads corresponding to the plurality of different regions.
  • the historical load time series in the preset modeling time domain can be decomposed, and the preset modeling time domain is historical load data for selecting a length of time, such as two weeks and three weeks. Or four weeks of historical load data, etc., that is, to determine N, the time-frequency transform is performed for a region, a region performs a time-frequency transform, including frequency domain decomposition and reconstruction, and then calculates a stability, time series
  • the following Fourier decomposition can be performed:
  • Step S206 combining frequency domain decomposition results according to periodicity to obtain low frequency components and high frequency components, wherein the low frequency components are related to meteorological factors, and the high frequency components are caused by random fluctuation of the load.
  • the load component, (t) represents a weekly cycle load component that varies in cycles, L(t) represents a low frequency component, and H(t) represents a high frequency component.
  • the daily periodic component and the weekly periodic component (t) are load components that vary by a fixed period, and are relatively easy to predict.
  • the key issue is how to build a predictive model for the residual components.
  • the period of each component of the low-frequency component (t) is greater than 24 hours.
  • Practice shows that the modeling of (t;» will improve the accuracy of load prediction.
  • the high-frequency component H(t) reflects the random fluctuation of the electrical load, and the model cannot be established.
  • the prediction is an unpredictable component.
  • the load time series is decomposed and reconstructed, it has periodic periodic components, periodic components, and low-frequency components related to meteorological factors and high-frequency components with high randomness.
  • a discriminant index of historical load regularity is introduced, that is, stability, which is a ratio of a portion of the historical load that is easy to grasp in the total load, and is generally expressed by a percentage value.
  • Step S208 determining historical load stability based on the historical load time series, the low frequency component, and the high frequency component.
  • the upper limit of the historical load stability may be determined according to the historical load time series and the high-frequency component to determine the historical load stability
  • the historical load stability is determined according to the historical load time series, the low-frequency component, and the high-frequency component.
  • the lower limit, wherein the historical load stability is a numerical range between the upper limit and the lower limit.
  • the upper limit of the historical load stability can be determined according to the following formula:
  • the corpse (for the historical load sequence, H (t) is the separated high-frequency load sequence.
  • the stability of the high-frequency component in the electric load determines the stability. Therefore, we can separate the high-frequency components in the time series.
  • L uppCT is the upper limit of historical load stability
  • L 1()WCT is the lower limit of historical load stability.
  • the following steps may be included: First, Calculating the power load prediction accuracy; then, determining whether the power load prediction accuracy is between the upper limit and the lower limit, wherein when the electric load prediction accuracy is between the upper limit and the lower limit, determining that the electric load prediction accuracy satisfies a preset condition, when the electric load forecasting When the accuracy is lower than the lower limit, it is determined that the power load prediction accuracy does not satisfy the preset condition.
  • the accuracy level that the short-term load forecasting can reach is pre-evaluated.
  • the embodiment of the present invention establishes a scientific and reasonable short-term load forecasting accuracy evaluation standard system by calculating the stability of the pre-assessed short-term load forecasting, and is a fair and reasonable dispatching and evaluating work management system.
  • a theoretical basis It should be noted that the steps shown in the flowchart of the accompanying drawings may be performed in a computer system such as a set of computer executable instructions, and, although the logical order is shown in the flowchart, in some cases, The steps shown or described may be performed in an order different than that herein.
  • modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device, or they may be separately fabricated into individual integrated circuit modules, or they may be Multiple modules or steps are made into a single integrated circuit module.
  • the invention is not limited to any specific combination of hardware and software.
  • the above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Water Supply & Treatment (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

公开了一种电力负荷预测精度的检测方法及装置,该方法包括:获取历史负荷时间序列;对时间序列中的数据进行时频变换以获取频域分解结果;对频域分解结果按照周期性进行合并以获得低频分量和高频分量;根据历史负荷时间序列、低频分量和高频分量确定历史负荷稳定度;以及根据历史负荷稳定度判断电力负荷预测精度是否满足预设条件。由于历史负荷稳定度一定程度上反映了该地区负荷变化的规律性,从而可以作为电力负荷预测精度的评估标准,进而可以根据不同地区的情况确定对应的电力负荷预测精度。

Description

电力负荷预测精度的检测方法及装置 技术领域 本发明涉及电力领域, 具体而言,涉及一种电力负荷预测精度的检测方法及装置。 背景技术 在电力系统的短期负荷预测的工作中, 负荷预测的准确率的高低, 对于电网企业 实现电力系统的实时平衡性、 节约能源以及提高经济效益等方面具有决定性的作用。 但是负荷本身具有一定的随机性, 各地区的电网分布情况以及气候状况等因素都会直 接影响负荷变化, 这样各个地区对于负荷规律性的把握的难易程度是不同的, 目前, 对负荷预测所能达到的精率考核要求就采用"一刀切 "的方式, 例如, 对北京市的海淀 地区和房山地区要求达到同样的预测精度, 但是, 往往对海淀地区的负荷预测更容易 达到较高的精度。 针对现有技术中无法根据当地负荷规律性条件来判断电力负荷预测精度是否满足 预设条件的问题, 目前尚未提出有效的解决方案。 发明内容 本发明提供了一种电力负荷预测精度的检测方法及装置, 以至少解决现有技术中 无法根据当地负荷规律性条件来判断电力负荷预测精度是否满足预设条件的问题。 为了实现上述目的, 根据本发明的一个方面, 提供了一种电力负荷预测精度的检 测方法。 根据本发明的电力负荷预测精度的检测方法包括: 获取历史负荷时间序列,其中, 时间序列为连续时间序列的负荷数据; 对时间序列中的数据进行时频变换以获取频域 分解结果; 对频域分解结果按照周期性进行合并以获得低频分量和高频分量, 其中, 低频分量与气象因素相关, 高频分量由负荷的随机波动性造成; 根据历史负荷时间序 列、 低频分量和高频分量确定历史负荷稳定度; 以及根据历史负荷稳定度判断电力负 荷预测精度是否满足预设条件。 进一步地, 获取历史负荷时间序列包括获取多个历史负荷时间序列, 其中, 多个 历史负荷时间序列为对应于多个不同地区的历史负荷时间序列, 对时间序列中的数据 进行时频变换以获取频域分解结果包括: 对预设建模时域中的历史负荷时间序列进行 如下傅里叶分解: p(i) = aQ + j(ai C0Si + 6i Sini¾t;); 以及对频域分解结果按照周期性 进行合并以获得低频分量和高频分量的步骤包括: 对 P(t)进行组合以获取下式: P(t) = a,+D(t) + W(t) + L(t) + H(t), 其中, P(t)表示历史负荷时间序列, t 为时间, 不同地区的 a0、 ai和 bi不同, ωί =~^χ2π = 1,2,''、Ν_· , ί¾ + Z)(t)表示以日为周 期变化的日周期负荷分量, (t)表示以周为周期变化的周周期负荷分量, L(t)表示低 频分量, H(t)表示高频分量。 进一步地, 根据历史负荷时间序列、 低频分量和高频分量确定历史负荷稳定度包 括: 根据历史负荷时间序列和高频分量确定历史负荷稳定度确定历史负荷稳定度的上 限; 以及根据历史负荷时间序列、 低频分量和高频分量确定历史负荷稳定度确定历史 负荷稳定度的下限, 其中, 历史负荷稳定度为上限和下限之间的数值范围。 一步地, 根据历史负荷时间序列和高频分量确定历史负荷稳定度的上限包括:
L.. 量和高频分
t确 χ100%, 其
Figure imgf000004_0001
中, Lupper为历史负荷稳定度的上限, Llower为历史负荷稳定度的下限。 进一步地, 根据历史负荷稳定度判断电力负荷预测精度是否满足预设条件包括: 计算电力负荷预测精度; 判断电力负荷预测精度是否在上限和下限之间, 其中, 当电 力负荷预测精度在上限和下限之间时, 确定电力负荷预测精度满足预设条件, 当电力 负荷预测精度在下限以下时, 确定电力负荷预测精度不满足预设条件。 为了实现上述目的, 根据本发明的另一个方面, 提供了一种电力负荷预测精度的 检测装置, 该装置用于执行本发明提供的任意一种电力负荷预测精度的检测方法。 根据本发明的另一方面, 提供了一种电力负荷预测精度的检测装置。 该电力负荷 预测精度的检测装置包括: 第一获取单元, 用于获取历史负荷时间序列, 其中, 时间 序列中为连续时间序列的负荷数据; 第二获取单元, 用于对时间序列中的数据进行时 频变换以获取频域分解结果; 第三获取单元, 用于对频域分解结果按照周期性进行合 并以获得低频分量和高频分量, 其中, 低频分量与气象因素相关, 高频分量由负荷的 随机波动性造成; 确定单元, 用于根据历史负荷时间序列、 低频分量和高频分量确定 历史负荷稳定度; 以及判断单元, 用于根据历史负荷稳定度判断电力负荷预测精度是 否满足预设条件。 进一步地, 获取历史负荷时间序列包括获取多个历史负荷时间序列, 其中, 多个 历史负荷时间序列为对应于多个不同地区的历史负荷时间序列, 第二获取单元包括: 分解子单元, 用于对预设建模时域中的历史负荷时间序列进行如下傅里叶分解:
P(t) = a0 (at cos ω,ί + bt sin ω,ί) 以及第三获取单元包括: 获取子单元, 用于对 P(t) 进行组合以获取下式: Ρ(^ = α。+ Ζ)(^ + ^(^ + (^ + (;), 其中, P(t)表示历史负荷 时间序列, t为时间, 不同地区的 a0、 ai和 bi不同, i¾ = "^x 2;r, ( = l, 2, ... , N - l),
' N , fl。+ D(t)表示以日为周期变化的日周期负荷分量, (t)表示以周为周期变化的周周期 负荷分量, L(t)表示低频分量, H(t)表示高频分量。 进一步地, 确定单元包括: 第一确定子单元, 用于根据历史负荷时间序列和高频 分量确定历史负荷稳定度的上限; 以及第二确定子单元,用于根据历史负荷时间序列、 低频分量和高频分量确定历史负荷稳定度的下限, 其中, 历史负荷稳定度为上限和下 限之间的数值范围。 一步地, 第一确定子单元还用于根据以下公式确定历史负荷稳定度的上限: 厶. 根据以下公式确定历
史负 χ 100%,其中, Lupper
Figure imgf000005_0001
为历史负荷稳定度的上限, Llower为历史负荷稳定度的下限。 进一步地, 判断单元包括: 计算子单元, 用于计算电力负荷预测精度; 判断子单 元, 用于判断电力负荷预测精度是否在上限和下限之间, 其中, 当电力负荷预测精度 在上限和下限之间时, 确定电力负荷预测精度满足预设条件, 当电力负荷预测精度在 下限之下时, 确定电力负荷预测精度不满足预设条件。 通过本发明, 对历史数据进行时频变换, 得到了低频分量和高频分量, 并进一步 得到历史负荷稳定度, 由于历史负荷稳定度一定程度上反映了该地区负荷变化的规律 性, 从而可以作为电力负荷预测精度的评估标准, 因此解决了现有技术中无法根据当 地负荷规律性条件来判断电力负荷预测精度是否满足预设条件的问题, 进而可以根据 不同地区的情况确定对应的电力负荷预测精度。 附图说明 构成本申请的一部分的附图用来提供对本发明的进一步理解, 本发明的示意性实 施例及其说明用于解释本发明, 并不构成对本发明的不当限定。 在附图中: 图 1是根据本发明实施例的电力负荷预测精度的检测装置的结构框图; 以及 图 2是根据本发明实施例的电力负荷预测精度的检测方法的流程图。 具体实施方式 需要说明的是, 在不冲突的情况下, 本申请中的实施例及实施例中的特征可以相 互组合。 下面将参考附图并结合实施例来详细说明本发明。 本发明实施例提供了一种电力负荷预测精度的检测装置, 以下对本发明实施例所 提供的电力负荷预测精度的检测装置进行介绍。 图 1是根据本发明实施例的电力负荷预测精度的检测装置的结构框图。 如图 1所示, 该电力负荷预测精度的检测装置包括第一获取单元 11、 第二获取单 元 12、 第三获取单元 13、 确定单元 14和判断单元 15。 第一获取单元 11用于获取历史负荷时间序列,其中, 时间序列中为连续时间序列 的负荷数据。 第二获取单元 12用于对时间序列中的数据进行时频变换以获取频域分解结果。 第三获取单元 13 用于对频域分解结果按照周期性进行合并以获得低频分量和高 频分量, 其中, 低频分量与气象因素相关, 即温度和湿度等气象因素均会对低频分量 产生影像, 高频分量由负荷的随机波动性造成。 确定单元 14用于根据历史负荷时间序列、低频分量和高频分量确定历史负荷稳定 度。 判断单元 15用于根据历史负荷稳定度判断电力负荷预测精度是否满足预设条件。 在本实施例中, 对历史数据进行时频变换, 得到了高频分量和低频分量, 并进一 步得到历史负荷稳定度, 由于历史负荷稳定度具有比较可靠的规律性, 从而可以作为 电力负荷预测精度的评估标准, 因此解决了现有技术中无法根据地区负荷规律性条件 来判断电力负荷预测精度是否满足预设条件的问题, 进而可以根据不同地区的情况确 定对应的电力负荷预测精度。 由于对不同地区的电力负荷预测精度的要求不同,因此对不同区域分别进行检测, 因此, 获取到的历史负荷时间序列包括获取多个历史负荷时间序列, 其中, 多个历史 负荷时间序列为对应于多个不同地区的历史负荷时间序列, 优选地, 第二获取单元包 括分解子单元和获取子单元, 其中, 分解子单元用于对预设建模时域中的历史负荷时 间序列进行如下傅里叶分解:
N-1
= 0 ( . sin ^.i)。 第三获取单元包括: 获取子单元用于对 P(t)进行组合以获取下式: P{t) = a0 + D(t) + W{t) + L(t) + H(t) o 其中, P(t)表示历史负荷时间序列, t 为时间, 不同地区的 a0、 ai和 bi 不同, i¾ = x2 , ( = l,2,— , N - l), a。+ D(t;)表示以日为周期变化的日周期负荷分量, W{t) 表示以周为周期变化的周周期负荷分量, L(t)表示低频分量, H(t)表示高频分量。 作为一种优选实施例, 本实施例中通过高频分量和低频分量来确定历史负荷稳定 度, 并根据该稳定度判断电力负荷精度是否合格, 优选地, 确定单元包括第一确定子 单元和第二确定子单元, 其中, 第一确定子单元用于根据历史负荷时间序列和高频分 量确定历史负荷稳定度的上限。 第二确定子单元用于根据历史负荷时间序列、 低频分 量和高频分量确定历史负荷稳定度的下限, 其中, 历史负荷稳定度为上限和下限之间 的数值范围。 通过高频分量和低频分量来确定历史负荷稳定度不仅方便快捷, 而且科 学合理。 具体地, 在本实施例中, 第一确定子单元还用于根据以下公式确定历史负荷稳定 度的上限:
Figure imgf000008_0001
第二确定子单元还用于根据以下公式确定历史负荷稳定度的下限包括:
Figure imgf000008_0002
其中, LuppCT为历史负荷稳定度的上限, L1()WCT为历史负荷稳定度的下限。 在获取计算出的电力负荷预测精度后, 通过判断该精度是否处于稳定度区间内, 即可确定该精度的计算是否合格, 优选地, 判断单元包括计算子单元和判断子单元, 其中, 计算子单元用于计算电力负荷预测精度。 判断子单元用于判断电力负荷预测精 度是否在上限和下限之间, 其中, 当电力负荷预测精度在上限和下限之间时, 确定电 力负荷预测精度满足预设条件, 也可以说是合格的, 当电力负荷预测精度在下限以下 时, 确定电力负荷预测精度不满足预设条件, 即不合格。 现有技术中包括多种方式进 行电力负荷预测精度的计算, 本实施例不再赘述。 本发明实施例还提供了一种电力负荷预测精度的检测方法, 该方法可以基于上述 的装置来执行。 图 2是根据本发明实施例的电力负荷预测精度的检测方法的流程图。 如图 2所示,该电力负荷预测精度的检测方法包括如下的步骤 S202至步骤 S21(h 步骤 S202, 获取历史负荷时间序列, 其中, 时间序列中为连续时间序列的负荷数 据 该时间序列为连续时间序列, 例如, 按 15分钟采集一下负荷数据计算, 一天的历 史负荷时间序列可以是由 96 个负荷数据按照时间先后排列成的连续历史负荷时间序 列。 步骤 S204, 对所述时间序列中的数据进行时频变换以获取频域分解结果, 在计算电力负荷预测精度时, 可以对不同地区分别计算, 因此, 获取的历史负荷 时间序列包括获取多个历史负荷时间序列, 其中, 多个历史负荷时间序列为对应于多 个不同地区的历史负荷时间序列, 因此, 在本步骤中, 可以对预设建模时域中的历史 负荷时间序列进行分解, 预设建模时域是选择一段时间长度的历史负荷数据, 比如两 周的、 三周的或四周的历史负荷数据等等, 即, 确定 N, 时频变换是针对一个地区进 行的, 一个地区进行一次时频变换, 包括频域分解和重构, 然后计算一个稳定度, 对 时间序列可以进行如下傅里叶分解:
N-1
P{t) = 0 ( . ύ ω^),
其中, N为历史负荷时间序列的长度, ω{ =^χ2π人 i = 1,2, '·、Ν_1、, P(t)表示历 史负荷时间序列, t为时间, 不同地区的 a0、 ai和 bi不同。 步骤 S206, 对频域分解结果按照周期性进行合并以获得低频分量和高频分量, 其 中, 低频分量与气象因素相关, 高频分量由负荷的随机波动性造成。 上述步骤包括: 对 P(t)进行组合以获取下式: P{t) = a0+D(t) + W{t) + L(t) + H(t), 其中, P(t)表示历史负荷时间序列, t为时间, 依据负荷周期性变化的特点, 将明 显的周期分量进行组合后, 实现对该 P(t)时间序列的重构, + 表示以日为周期 变化的日周期负荷分量, (t)表示以周为周期变化的周周期负荷分量, L(t)表示低频 分量, H(t)表示高频分量。 日周期分量 和周周期分量 (t)是按固定周期变化的负荷分量,比较容易 进行预测。 因此, 关键问题是如何对剩余分量建立预测模型。 低频分量 (t)的各分量 的周期大于 24小时,实践表明对 (t;»的建模会改善负荷预测的精度。高频分量 H(t)反 映了电力负荷的随机波动, 无法建立模型进行预测, 属于不可预测分量。 在将负荷时间序列通过分解和重构得到其具有周期性的日周期分量、周周期分量, 以及与气象等因素相关的低频分量和随机性很强的高频分量后, 本实施例引入了历史 负荷规律性的判别指标, 即, 稳定度, 稳定度是历史负荷中变化规律易于把握的部分 占总负荷的比例, 一般用一个百分比数值来表示。 步骤 S208, 根据历史负荷时间序列、 低频分量和高频分量确定历史负荷稳定度。 具体地, 可以根据历史负荷时间序列和高频分量确定历史负荷稳定度确定历史负 荷稳定度的上限, 并根据历史负荷时间序列、 低频分量和高频分量确定历史负荷稳定 度确定历史负荷稳定度的下限,其中, 历史负荷稳定度为上限和下限之间的数值范围。 进一步地, 可以根据以下公式确定历史负荷稳定度的上限:
Figure imgf000010_0001
尸( 为历史负荷序列, H (t)为分离出来的高频负荷序列, 由于电力负荷中高频 分量比重的大小就决定了其稳定度, 因此, 我们可以分离时间序列中的高频分量, 来 估计历史负荷稳定度的上限。 根据以下公式确定历史负荷稳定度的下限:
Figure imgf000010_0002
其中, LuppCT为历史负荷稳定度的上限, L1()WCT为历史负荷稳定度的下限。 步骤 S210, 根据历史负荷稳定度判断电力负荷预测精度是否满足预设条件。 历史负荷稳定度体现了历史负荷规律性部分所占比例的波动范围, 根据对历史负 荷规律性的分析可以大致判断出当前预测的电力负荷预测精度是否合格, 具体地, 可 以包括以下步骤: 首先, 计算电力负荷预测精度; 然后, 判断电力负荷预测精度是否 在上限和下限之间, 其中, 当电力负荷预测精度在上限和下限之间时, 确定电力负荷 预测精度满足预设条件, 当电力负荷预测精度低于下限时, 确定电力负荷预测精度不 满足预设条件。 通过分析计算历史负荷的规律性成分所占比例, 即稳定度上 /下限的计算, 预评估 短期负荷预测可能达到的精度水平。 从以上的描述中, 可以看出, 本发明实施例通过计算预评估短期负荷预测的稳定 度, 以此为基础建立科学合理的短期负荷预测精度考核标准体系, 为公平合理的调度 考核工作管理体系提供理论基础。 需要说明的是, 在附图的流程图示出的步骤可以在诸如一组计算机可执行指令的 计算机系统中执行, 并且, 虽然在流程图中示出了逻辑顺序, 但是在某些情况下, 可 以以不同于此处的顺序执行所示出或描述的步骤。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可以用通用 的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布在多个计算装置所 组成的网络上, 可选地, 它们可以用计算装置可执行的程序代码来实现, 从而, 可以 将它们存储在存储装置中由计算装置来执行, 或者将它们分别制作成各个集成电路模 块, 或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。 这样, 本发明 不限制于任何特定的硬件和软件结合。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本领域的技 术人员来说, 本发明可以有各种更改和变化。 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

权 利 要 求 书
1. 一种电力负荷预测精度的检测方法, 其特征在于, 包括:
获取历史负荷时间序列,其中,所述时间序列为连续时间序列的负荷数据; 对所述时间序列中的数据进行时频变换以获取频域分解结果; 对所述频域分解结果按照周期性进行合并以获得低频分量和高频分量, 其 中, 所述低频分量与气象因素相关, 所述高频分量由负荷的随机波动性造成; 根据所述历史负荷时间序列、 所述低频分量和所述高频分量确定历史负荷 稳定度; 以及
根据所述历史负荷稳定度判断电力负荷预测精度是否满足预设条件。
2. 根据权利要求 1所述的方法, 其特征在于,
获取历史负荷时间序列包括获取多个历史负荷时间序列, 其中, 所述多个 历史负荷时间序列为对应于多个不同地区的历史负荷时间序列,
对所述时间序列中的数据进行时频变换以获取频域分解结果包括: 对预设建模时域中的所述历史负荷时间序列进行如下傅里叶分解:
N-1
P{t) = 0 + ^( . cos + έ. ύ ω^); 以及 对所述频域分解结果按照周期性进行合并以获得低频分量和高频分量的步 骤包括: 对 P(t)进行组合以获取下式:
P(t) = a,+D(t) + W(t) + L(t) + H(t), 其中, 所述 P(t)表示所述历史负荷时间序列, t为时间, 不同地区的 a0、 ai和 bi 不同, ωί =^χ2π人 i = l,2,"、N_l、, α。 表示以日为周期变化的日周期负 荷分量, 所述 (t)表示以周为周期变化的周周期负荷分量, 所述 L(t)表示所述 低频分量, 所述 H(t)表示所述高频分量。
3. 根据权利要求 1所述的方法, 其特征在于, 根据所述历史负荷时间序列、 所述 低频分量和所述高频分量确定历史负荷稳定度包括: 根据所述历史负荷时间序列和所述高频分量确定历史负荷稳定度确定历史 负荷稳定度的上限; 以及
根据所述历史负荷时间序列、 所述低频分量和所述高频分量确定历史负荷 稳定度确定历史负荷稳定度的下限, 其中, 所述历史负荷稳定度为所述上限和 所述下限之间的数值范围。
4. 根据权利要求 3所述的方法, 其特征在于,
根据所述历史负荷时间序列和所述高频分量确定历史负荷稳定度的上限包 括-
Figure imgf000013_0001
根据所述历史负荷时间序列、 所述低频分量和所述高频分量确定历史负荷 稳定度的下限包括:
Figure imgf000013_0002
其中, 所述 LuppCT为所述历史负荷稳定度的上限, 所述 L1()WCT为所述历史负 荷稳定度的下限。
5. 根据权利要求 4所述的方法, 其特征在于, 根据所述历史负荷稳定度判断电力 负荷预测精度是否满足预设条件包括:
计算所述电力负荷预测精度;
判断所述电力负荷预测精度是否在所述上限和所述下限之间, 其中, 当所述电力负荷预测精度在所述上限和所述下限之间时, 确定所述 电力负荷预测精度满足所述预设条件, 当所述电力负荷预测精度在所述下限以 下时, 确定所述电力负荷预测精度不满足所述预设条件。
6. 一种电力负荷预测精度的检测装置, 其特征在于, 包括:
第一获取单元, 用于获取历史负荷时间序列, 其中, 所述时间序列中为连 续时间序列的负荷数据; 第二获取单元, 用于对所述时间序列中的数据进行时频变换以获取频域分 解结果;
第三获取单元, 用于对所述频域分解结果按照周期性进行合并以获得低频 分量和高频分量, 其中, 所述低频分量与气象因素相关, 所述高频分量由负荷 的随机波动性造成;
确定单元, 用于根据所述历史负荷时间序列、 所述低频分量和所述高频分 量确定历史负荷稳定度; 以及
判断单元, 用于根据所述历史负荷稳定度判断电力负荷预测精度是否满足 预设条件。
7. 根据权利要求 6所述的装置, 其特征在于,
获取历史负荷时间序列包括获取多个历史负荷时间序列, 其中, 所述多个 历史负荷时间序列为对应于多个不同地区的历史负荷时间序列,
所述第二获取单元包括:
分解子单元, 用于对预设建模时域中的所述历史负荷时间序列进行如下傅 里叶分解:
N-1
P(t) = 0 + cos ω{ί + bt sin^ ); 以及 所述第三获取单元包括: 获取子单元, 用于对 P(t)进行组合以获取下式: P{t) = a0+D(t) + W{t) + L(t) + H(t), 其中, 所 P(t)表示所述历史负荷时间序列, t为时间, 不同地区的 a0、 ai 和 bi不同, ί¾ = 1,2,·.·, _1), a。+D(t)表示以日为周期变化的日
Figure imgf000014_0001
周期负荷分量, 所述 (t)表示以周为周期变化的周周期负荷分量, 所述 L(t) 表示所述低频分量, 所述 H(t)表示所述高频分量。
8. 根据权利要求 6所述的装置, 其特征在于, 所述确定单元包括:
第一确定子单元, 用于根据所述历史负荷时间序列和所述高频分量确定历 史负荷稳定度的上限; 以及 第二确定子单元, 用于根据所述历史负荷时间序列、 所述低频分量和所述 高频分量确定历史负荷稳定度的下限, 其中, 所述历史负荷稳定度为所述上限 和所述下限之间的数值范围。
9. 根据权利要求 8所述的装置, 其特征在于,
所述第一确定子单元还用于根据以下公式确定所述历史负荷稳定度的上 限:
Figure imgf000015_0001
所述第二确定子单元还用于根据以下公式确定所述历史负荷稳定度的下限 包括:
Figure imgf000015_0002
其中, 所述 LuppCT为所述历史负荷稳定度的上限, 所述 L1()WCT为所述历史负 荷稳定度的下限。
10. 根据权利要求 9所述的装置, 其特征在于, 所述判断单元包括:
计算子单元, 用于计算所述电力负荷预测精度;
判断子单元, 用于判断所述电力负荷预测精度是否在所述上限和所述下限 之间,
其中, 当所述电力负荷预测精度在所述上限和所述下限之间时, 确定所述 电力负荷预测精度满足所述预设条件, 当所述电力负荷预测精度在所述下限之 下时, 确定所述电力负荷预测精度不满足所述预设条件。
PCT/CN2013/086800 2012-12-06 2013-11-08 电力负荷预测精度的检测方法及装置 Ceased WO2014086226A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210520888.2 2012-12-06
CN201210520888.2A CN103854067B (zh) 2012-12-06 2012-12-06 电力负荷预测精度的检测方法及装置

Publications (1)

Publication Number Publication Date
WO2014086226A1 true WO2014086226A1 (zh) 2014-06-12

Family

ID=50861697

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/086800 Ceased WO2014086226A1 (zh) 2012-12-06 2013-11-08 电力负荷预测精度的检测方法及装置

Country Status (2)

Country Link
CN (1) CN103854067B (zh)
WO (1) WO2014086226A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016109946A1 (en) * 2015-01-06 2016-07-14 Accenture Global Services Limited Power distribution transformer load prediction analysis system
CN112580187A (zh) * 2020-11-10 2021-03-30 广东电网有限责任公司广州供电局 干式变压器过热预警方法、装置、计算机设备、存储介质
CN117851414A (zh) * 2024-03-07 2024-04-09 杭州永德电气有限公司 一种避雷器老化试验数据存储方法及系统
CN119448273A (zh) * 2024-11-15 2025-02-14 国网四川省电力公司营销服务中心 一种基于时间序列核心融合的电力负荷预测方法
CN119903306A (zh) * 2025-03-27 2025-04-29 水利部交通运输部国家能源局南京水利科学研究院 基于物理方法与深度学习耦合的水库发电流量计算方法
CN120197742A (zh) * 2025-02-10 2025-06-24 浙江工商大学 基于频域自适应滤波器的电力负荷预测方法及系统

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335539B (zh) * 2014-08-08 2018-11-06 国家电网公司 用于电力负荷的数据处理方法和装置
CN106655223B (zh) * 2016-12-12 2019-04-12 国网北京市电力公司 分析电网独立控制区运行方法以及装置
CN109978717A (zh) * 2019-03-15 2019-07-05 国网冀北电力有限公司 分布式光伏最大渗透率确定方法及装置
CN111639011B (zh) * 2020-06-11 2023-05-12 支付宝(杭州)信息技术有限公司 一种数据监控方法、装置及设备

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
CHAI, YI ET AL.: "Research on method of short load term forecasting under smart grid client environment", LOW VOLTAGE APPARATUS, no. 17, September 2012 (2012-09-01), pages 8 - 12 *
CHEN, GUIYUAN ET AL.: "A new method for short-term forecast evaluation based on stability degree.", SOUTHERN POWER SYSTEM TECHNOLOGY, vol. 4, no. 6, December 2010 (2010-12-01), pages 94 - 97 *
JIANG, YUSHAN ET AL.: "Short-term load forecasting based on frequency analysis of regression", POWER DEMAND SIDE MANAGEMENT, vol. 13, no. 5, September 2011 (2011-09-01), pages 12 - 16 *
MU, GANG ET AL.: "Load forecasting method based on intrinsic error evaluation", AUTOMATION OF ELECTRIC POWER SYSTEMS, 25 November 2001 (2001-11-25), pages 37 - 40 *
MU, GANG ET AL.: "Studies on load regularity evaluating method for load forecasting, PROCEEDINGS OF THE CSEE", vol. 21, no. 10, October 2001 (2001-10-01), pages 96 - 101 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016109946A1 (en) * 2015-01-06 2016-07-14 Accenture Global Services Limited Power distribution transformer load prediction analysis system
US10289954B2 (en) 2015-01-06 2019-05-14 Accenture Global Services Limited Power distribution transformer load prediction analysis system
CN112580187A (zh) * 2020-11-10 2021-03-30 广东电网有限责任公司广州供电局 干式变压器过热预警方法、装置、计算机设备、存储介质
CN117851414A (zh) * 2024-03-07 2024-04-09 杭州永德电气有限公司 一种避雷器老化试验数据存储方法及系统
CN117851414B (zh) * 2024-03-07 2024-05-17 杭州永德电气有限公司 一种避雷器老化试验数据存储方法及系统
CN119448273A (zh) * 2024-11-15 2025-02-14 国网四川省电力公司营销服务中心 一种基于时间序列核心融合的电力负荷预测方法
CN120197742A (zh) * 2025-02-10 2025-06-24 浙江工商大学 基于频域自适应滤波器的电力负荷预测方法及系统
CN119903306A (zh) * 2025-03-27 2025-04-29 水利部交通运输部国家能源局南京水利科学研究院 基于物理方法与深度学习耦合的水库发电流量计算方法

Also Published As

Publication number Publication date
CN103854067B (zh) 2017-03-15
CN103854067A (zh) 2014-06-11

Similar Documents

Publication Publication Date Title
WO2014086226A1 (zh) 电力负荷预测精度的检测方法及装置
CN104616079B (zh) 基于气温变化的电网日用电量预测方法
Akay et al. Grey prediction with rolling mechanism for electricity demand forecasting of Turkey
US20160012340A1 (en) Temperature-based estimation of building occupancy states
JP6806333B2 (ja) 消費電力予測装置および消費電力予測方法
CN104636874B (zh) 检测业务异常的方法及设备
JP2012254007A5 (zh)
CN112736902B (zh) 一种基于stl分解的时间序列短期电力负荷预测方法
CN115015474A (zh) 一种电力用户的碳排放量检测方法及设备
CN105120487A (zh) 一种业务数据预测方法及装置
CN115130788B (zh) 基于大数据分析的用电预测系统
KR20130044765A (ko) 예측 속도 향상을 위한 신경망과 보정계수를 이용한 주간 부하 예측 장치 및 방법
CN111507607A (zh) 日电量预测方法、装置、计算机设备和介质
CN116632838B (zh) 应用于发电企业的电能供给分析的方法及装置
CN114254864A (zh) 用电数据处理方法、装置、计算机设备、存储介质
Iqbal et al. Power and performance analysis of network traffic prediction techniques
CN106897940B (zh) 一种电网调峰能力限值的风电功率预测指标评价方法及系统
CN104134096A (zh) 基于省地一体化智能负荷预测方法
CN105590140A (zh) 电力系统短期负荷的预测方法和装置
CN116664365B (zh) 一种基于物联网的智慧社区数字化管理方法及装置
CN110009161A (zh) 供水预测方法及装置
Qiao et al. Hour-ahead power load demand time series forecasting using four methods in three cases
CN104252647A (zh) 基于反距离权重插值法的用电负荷预测方法
WO2019119319A1 (zh) 停车场泊位占用率预测方法、装置、设备及存储介质
Seid et al. A novel technique for prediction of time points for scheduling of multipurpose batch plants

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: 13860498

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: 13860498

Country of ref document: EP

Kind code of ref document: A1