WO2020118586A1 - 一种能源消耗预测方法及装置 - Google Patents

一种能源消耗预测方法及装置 Download PDF

Info

Publication number
WO2020118586A1
WO2020118586A1 PCT/CN2018/120746 CN2018120746W WO2020118586A1 WO 2020118586 A1 WO2020118586 A1 WO 2020118586A1 CN 2018120746 W CN2018120746 W CN 2018120746W WO 2020118586 A1 WO2020118586 A1 WO 2020118586A1
Authority
WO
WIPO (PCT)
Prior art keywords
energy consumption
value
data
prediction
time length
Prior art date
Application number
PCT/CN2018/120746
Other languages
English (en)
French (fr)
Inventor
刘鹏
马亚泽
Original Assignee
华北电力大学扬中智能电气研究中心
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 华北电力大学扬中智能电气研究中心 filed Critical 华北电力大学扬中智能电气研究中心
Priority to PCT/CN2018/120746 priority Critical patent/WO2020118586A1/zh
Priority to JP2021531541A priority patent/JP2022510667A/ja
Publication of WO2020118586A1 publication Critical patent/WO2020118586A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION 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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising

Definitions

  • the invention relates to the technical field of energy consumption prediction, in particular to an energy consumption prediction method and device.
  • the current building energy consumption prediction technology is generally based on the SVM model for energy consumption prediction, or based on the improved SVM model for energy consumption prediction.
  • the time series problem is not considered, so the accuracy of the energy consumption prediction in the prior art is low.
  • Embodiments of the present invention provide an energy consumption prediction method and device to solve the problem of inaccurate energy consumption prediction in the prior art.
  • An embodiment of the present invention provides an energy consumption prediction method.
  • the method includes:
  • the predicted predicted energy consumption value at the first predicted time is determined.
  • the determining the target energy consumption prediction value at the first prediction time based on the first energy consumption prediction value and the second energy consumption prediction value includes:
  • the first prediction error and the second prediction error determine a first weight corresponding to the first energy consumption prediction value and a second weight corresponding to the second energy consumption prediction value
  • weighted average processing is performed on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the training process of the ARIMA model includes:
  • the third data within the second time length in the first training set For each third data within the second time length in the first training set, the third data within the second time length and the true value of energy consumption at the second predicted time corresponding to the second time length are input into the ARIMA model, Train the ARIMA model.
  • the training process of the SVM model includes:
  • the fourth data in each second time length in the second training set the fourth data in the second time length and the true value of energy consumption in the second predicted time corresponding to the second time length are input into the SVM model, Train the SVM model.
  • the first data and the third data include energy consumption values.
  • the second data and the fourth data include energy consumption values and weather data.
  • an energy consumption prediction device which includes:
  • the input module is used to input the acquired first data within the first time length into the pre-trained autoregressive integral moving average ARIMA model, and input the acquired second data within the first time length into the pre-training support Vector machine SVM model;
  • a first determining module configured to determine a first predicted value of energy consumption based on the ARIMA model, and determine a second predicted value of energy consumption based on the SVM model;
  • the second determination module is configured to determine the target energy consumption prediction value at the first prediction time based on the first energy consumption prediction value and the second energy consumption prediction value.
  • the second determination module is specifically configured to determine the latest acquired real value of energy consumption; determine the first prediction error according to the first predicted energy consumption value and the real energy consumption value; according to the second energy consumption The predicted value and the real value of energy consumption determine the second prediction error; based on the first prediction error and the second prediction error, determine the first weight corresponding to the first energy consumption prediction value and the second corresponding to the second energy consumption prediction value Weight; according to the first weight and the second weight, perform weighted average processing on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the device further includes:
  • a first training module for the third data in each second time length in the first training set, the third data in the second time length, and the energy of the second prediction time corresponding to the second time length
  • the real value is consumed to input the ARIMA model, and the ARIMA model is trained.
  • the device further includes:
  • a second training module for the fourth data in each second time length in the second training set, the fourth data in the second time length, and the energy of the second prediction time corresponding to the second time length
  • the real value is consumed to input the SVM model, and the SVM model is trained.
  • An embodiment of the present invention provides an energy consumption prediction method and device.
  • the method includes: inputting the acquired first data within a first time length into a pre-trained autoregressive integral moving average ARIMA model, and then acquiring the acquired The second data within the first time length is input into the pre-trained support vector machine SVM model; based on the ARIMA model, the first energy consumption prediction value is determined, and based on the SVM model, the second energy consumption prediction value is determined; Describe the first predicted value of energy consumption and the second predicted value of energy consumption to determine the predicted value of target energy consumption at the first predicted time.
  • the first energy consumption prediction value is determined based on the ARIMA model
  • the second energy consumption prediction value is determined based on the SVM model.
  • the ARIMA model can fit the time series well, and the SVM model can be well applied to the problem of small sample, nonlinear and high-dimensional pattern recognition, combined with the first energy consumption prediction value determined by the ARIMA model and the second energy source determined by the SVM model
  • the predicted value of consumption determines the predicted value of target energy consumption. Make full use of the advantages of ARIMA model and SVM model to make the determined target energy consumption prediction value more accurate.
  • Embodiment 1 is a schematic diagram of an energy consumption prediction process provided by Embodiment 1 of the present invention.
  • Embodiment 2 is a schematic diagram of determining a target energy consumption prediction value provided by Embodiment 2 of the present invention
  • Embodiment 3 is a schematic diagram of an energy consumption prediction process provided by Embodiment 2 of the present invention.
  • FIG. 4 is a schematic structural diagram of an energy consumption prediction device according to an embodiment of the present invention.
  • FIG. 1 is a schematic diagram of an energy consumption prediction process provided by an embodiment of the present invention. The process includes the following steps:
  • S101 Input the acquired first data in the first time length into the pre-trained autoregressive integral moving average ARIMA model, and input the acquired second data in the first time length into the pre-trained support vector machine SVM model.
  • S102 Determine a first predicted value of energy consumption based on the ARIMA model, and determine a second predicted value of energy consumption based on the SVM model.
  • S103 Determine the target energy consumption prediction value at the first prediction time according to the first energy consumption prediction value and the second energy consumption prediction value.
  • the energy consumption prediction method provided by the embodiment of the present invention is applied to an electronic device, and the electronic device may be a device such as a PC or a tablet computer.
  • the energy consumption prediction method provided by the embodiment of the present invention includes, but is not limited to, energy consumption prediction for large public buildings.
  • the electronic device may acquire the first data and the second data within the first time length.
  • the first time length may be three days, four days, etc., and the first data may be energy consumption within the first time length.
  • the second data may be the same as or different from the first data. If the second data is different from the first data, the second data may be energy consumption within the first time length and weather data within the first time length.
  • the weather data may be data such as average temperature and average humidity per day.
  • the pre-trained autoregressive integral moving average ARIMA model and support vector machine SVM model are stored in the electronic device. After acquiring the first data and the second data within the first time length, the electronic device inputs the first data into the ARIMA model and the second data into the SVM model. Because the ARIMA model can fit the time series well, the SVM model can be well adapted to the problem of small sample, non-linear and high-dimensional pattern recognition. Therefore, based on the ARIMA model, the determined first energy consumption prediction value fits well Based on the time series, based on the SVM model, the determined second energy consumption prediction value is suitable for small sample, nonlinear and high-dimensional pattern recognition problems.
  • the electronic device After determining the first energy consumption predicted value and the second energy consumption predicted value, the electronic device determines the target energy consumption predicted value at the first predicted time according to the first energy consumption predicted value and the second energy consumption predicted value.
  • the first prediction time may be the first day after the first time length, or the second day after the first time length, and so on.
  • the first time period is from October 1, 2018 to October 3, 2018, and the first predicted time may be October 4, 2018, or October 5, 2018.
  • the correspondence between the first time length and the first prediction time is determined by the trained ARIMA model and SVM model.
  • the first prediction time is also the first day after the first time length; if when training the ARIMA model and the SVM model, The prediction is the second day after the first time length, then the first prediction time is also the second day after the first time length.
  • the electronic device determines the target energy consumption prediction value at the first prediction time based on the first energy consumption prediction value and the second energy consumption prediction value, and may take the average value of the first energy consumption prediction value and the second energy consumption prediction value as the first
  • the target energy consumption forecast value of a forecast time can also determine the importance of the first energy consumption forecast value and the second energy consumption forecast value according to experience, and allocate the first energy consumption forecast value and the second energy consumption forecast value according to the different importance levels
  • the corresponding weight where the greater the importance, the greater the corresponding weight.
  • a weighted average calculation is performed to obtain the predicted target energy consumption value at the first predicted time.
  • the first energy consumption prediction value is determined based on the ARIMA model
  • the second energy consumption prediction value is determined based on the SVM model.
  • the ARIMA model can fit the time series well, and the SVM model can be well applied to the problem of small sample, nonlinear and high-dimensional pattern recognition, combined with the first energy consumption prediction value determined by the ARIMA model and the second energy source determined by the SVM model
  • the predicted value of consumption determines the predicted value of target energy consumption. Make full use of the advantages of ARIMA model and SVM model to make the determined target energy consumption prediction value more accurate.
  • the first prediction is determined according to the first energy consumption prediction value and the second energy consumption prediction value
  • the time target energy consumption forecast values include:
  • the first prediction error and the second prediction error determine a first weight corresponding to the first energy consumption prediction value and a second weight corresponding to the second energy consumption prediction value
  • weighted average processing is performed on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the electronic device after determining the first energy consumption predicted value and the second energy consumption predicted value, in order to be able to accurately determine the target energy consumption predicted value at the first predicted time, the electronic device first determines the latest acquired energy consumption actual value. For example, if you want to predict the energy consumption value on October 4, 2018, the latest real value of energy consumption may be the actual value of energy consumption on October 3, 2018.
  • the electronic device determines the first prediction error based on the first energy consumption prediction value and the real energy consumption value; and determines the second prediction error based on the second energy consumption prediction value and the real energy consumption value. Specifically, the electronic device calculates the difference between the real energy consumption value and the first energy consumption prediction value as the first prediction error, and calculates the difference between the actual energy consumption value and the second energy consumption prediction value as the second prediction error. If the determined first prediction error or second prediction error is a negative value, the absolute value of the first prediction error or the second prediction error is taken. Ensure that the first prediction error and the second prediction error are positive.
  • the electronic device determines the first weight corresponding to the first energy consumption prediction value and the second weight corresponding to the second energy consumption prediction value according to the first prediction error and the second prediction error. Specifically, the sum of the first prediction error and the second prediction error can be calculated, and then the ratio of the first prediction error and the sum value can be calculated as the first weight, and the ratio of the second prediction error and the sum value can be calculated as the second weight . After the first weight and the second weight are determined, the first weight and the second weight are used to perform weighted average processing on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the ratio of the sum value and the first prediction error can also be calculated as the first intermediate parameter, and the ratio of the sum value and the second prediction error can be calculated as The second intermediate parameter.
  • the sum value of the first intermediate parameter and the second intermediate parameter is calculated, and then the ratio of the first intermediate parameter and the sum value is calculated as the first weight, and the ratio of the second intermediate parameter and the sum value is calculated as the second weight.
  • the first weight and the second weight are determined, the first weight and the second weight are used to perform weighted average processing on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • FIG. 2 is a schematic diagram of determining a target energy consumption prediction value provided by an embodiment of the present invention.
  • , and calculate the second prediction error E2
  • . Then calculate the sum of the first prediction error and the second prediction error E1 + E2.
  • the electronic device determines the first prediction error based on the first energy consumption predicted value and the real energy consumption value; based on the second energy consumption predicted value and the real energy consumption value, the second prediction error is determined; then The first prediction error and the second prediction error determine the first weight corresponding to the first energy consumption prediction value and the second weight corresponding to the second energy consumption prediction value. Finally, according to the first weight and the second weight, the first energy consumption predicted value and the second energy consumption predicted value are subjected to weighted average processing to determine the target energy consumption predicted value. Therefore, the determined target energy consumption prediction value can be made more accurate.
  • FIG. 3 is a schematic diagram of an energy consumption prediction process provided by an embodiment of the present invention.
  • the interference data may affect the final determination The accuracy of the predicted target energy consumption value of the target, in order to make the determined target energy consumption prediction value more accurate, before entering the first data into the ARIMA model and the second data into the SVM model, the first data and the second Data is cleaned.
  • the electronic device may save a smaller first threshold and a larger second threshold, use the data in the first data and the second data that are smaller than the first threshold and larger than the second threshold as interference data, and use the interference data Filter out from the first data and the second data, and then input the remaining first data into the ARIMA model, and the remaining second data into the SVM model.
  • the first energy consumption prediction value is obtained based on the ARIMA model
  • the second energy consumption prediction value is obtained based on the SVM model
  • the first weight corresponding to the first energy consumption prediction value and the second weight corresponding to the second energy consumption prediction value are calculated according to
  • the first weight and the second weight perform weighted average processing on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the training process of the ARIMA model includes:
  • the third data within the second time length in the first training set For each third data within the second time length in the first training set, the third data within the second time length and the true value of energy consumption at the second predicted time corresponding to the second time length are input into the ARIMA model, Train the ARIMA model.
  • the training set used to train the ARIMA model is used as the first training set, and the data in the first training set is used as the third data.
  • the first data and the third data include energy consumption values.
  • the electronic device divides the third data in the first training set into third data within each second time length. For example, if the second time length is three days, the third data in the first training set may be divided into January 1, 2018 Third data from Sunday to January 3, 2018, Third data from January 2, 2018 to January 4, 2018, Third data from January 3, 2018 to January 4, 2018, etc. . Then for each second time length, the second predicted time corresponding to the second time length is determined. Specifically, the second predicted time corresponding to the second time length may be the first day after the second time length, or the second day after the second time length, etc., as long as it is set in advance.
  • the second predicted time corresponding to January 1, 2018 to January 3, 2018 can be January 4, 2018, and the second predicted time corresponding to January 2, 2018 to January 4, 2018 can be It is January 5, 2018, etc.
  • the ARIMA model When training the ARIMA model, for the third data in each second time length in the first training set, the third data in the second time length and the second prediction time corresponding to the second time length
  • the real value of energy consumption is input into the ARIMA model.
  • the parameters of the ARIMA model are adjusted according to the difference between the predicted energy consumption value of the second prediction time output by the ARIMA model and the true value of energy consumption until the ARIMA model training is completed.
  • the training process of the SVM model includes:
  • the fourth data in each second time length in the second training set the fourth data in the second time length and the true value of energy consumption in the second predicted time corresponding to the second time length are input into the SVM model, Train the SVM model.
  • the training set used to train the SVM model is used as the second training set, and the data in the second training set is used as the fourth data.
  • the second data and the fourth data include energy consumption values and weather data.
  • the electronic device divides the third data in the second training set into fourth data within each second time length. For example, if the second time length is three days, the fourth data in the second training set may be divided into January 1, 2018 Third data from Sunday to January 3, 2018, Third data from January 2, 2018 to January 4, 2018, Third data from January 3, 2018 to January 4, 2018, etc. . Then for each second time length, the second predicted time corresponding to the second time length is determined. Specifically, the second predicted time corresponding to the second time length may be the first day after the second time length, or the second day after the second time length, etc., as long as it is set in advance.
  • the second predicted time corresponding to January 1, 2018 to January 3, 2018 can be January 4, 2018, and the second predicted time corresponding to January 2, 2018 to January 4, 2018 can be It is January 5, 2018, etc.
  • the fourth data in the second time length in the second training set For the fourth data in each second time length in the second training set, the fourth data in the second time length and the second prediction time corresponding to the second time length
  • the real value of energy consumption is input into the SVM model.
  • the parameters of the SVM model are adjusted according to the difference between the predicted value of energy consumption and the true value of energy consumption at the second predicted time output by the SVM model until the training of the SVM model is completed.
  • the process of training the ARIMA model and the SVM model belongs to the prior art, and the process will not be repeated here.
  • FIG. 4 is a schematic structural diagram of an energy consumption prediction device according to an embodiment of the present invention.
  • the device includes:
  • the input module 41 is configured to input the acquired first data within the first time length into the pre-trained autoregressive integral moving average ARIMA model, and input the acquired second data within the first time length into the pre-trained completed Support vector machine SVM model;
  • the first determining module 42 is configured to determine a first predicted value of energy consumption based on the ARIMA model, and determine a second predicted value of energy consumption based on the SVM model;
  • the second determination module 43 is configured to determine the target energy consumption prediction value at the first prediction time based on the first energy consumption prediction value and the second energy consumption prediction value.
  • the second determination module 43 is specifically configured to determine the latest acquired energy consumption real value; determine the first prediction error according to the first energy consumption prediction value and the energy consumption real value; according to the second energy consumption prediction value Determine the second prediction error with the real value of energy consumption; determine the first weight corresponding to the first energy consumption prediction value and the second weight corresponding to the second energy consumption prediction value according to the first prediction error and the second prediction error; According to the first weight and the second weight, weighted average processing is performed on the first energy consumption predicted value and the second energy consumption predicted value to determine the target energy consumption predicted value.
  • the device also includes:
  • the first training module 44 is configured to, for each third data in the second time length in the first training set, the third data in the second time length and the second prediction time corresponding to the second time length
  • the real value of energy consumption is input into the ARIMA model, and the ARIMA model is trained.
  • the device also includes:
  • the second training module 45 is used for the fourth data in each second time length in the second training set, the fourth data in the second time length, and the second prediction time corresponding to the second time length
  • the real value of energy consumption is input into the SVM model, and the SVM model is trained.
  • An embodiment of the present invention provides an energy consumption prediction method and device.
  • the method includes: inputting the acquired first data within a first time length into a pre-trained autoregressive integral moving average ARIMA model, and then acquiring the acquired The second data within the first time length is input to the pre-trained support vector machine SVM model; based on the ARIMA model, the first energy consumption prediction value is determined, and based on the SVM model, the second energy consumption prediction value is determined; Describe the first predicted value of energy consumption and the second predicted value of energy consumption to determine the predicted value of target energy consumption at the first predicted time.
  • the first energy consumption prediction value is determined based on the ARIMA model
  • the second energy consumption prediction value is determined based on the SVM model.
  • the ARIMA model can fit the time series well, and the SVM model can be well applied to the problem of small sample, nonlinear and high-dimensional pattern recognition, combined with the first energy consumption prediction value determined by the ARIMA model and the second energy determined by the SVM model
  • the predicted value of consumption determines the predicted value of target energy consumption. Make full use of the advantages of ARIMA model and SVM model to make the determined target energy consumption prediction value more accurate.
  • each flow and/or block in the flowchart and/or block diagram and a combination of the flow and/or block in the flowchart and/or block diagram may be implemented by computer program instructions.
  • These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processing machine, or other programmable data processing device to produce a machine that enables the generation of instructions executed by the processor of the computer or other programmable data processing device
  • These computer program instructions may also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, the instructions
  • the device implements the functions specified in one block or multiple blocks of the flowchart one flow or multiple flows and/or block diagrams.
  • These computer program instructions can also be loaded on a computer or other programmable data processing device, so that a series of operating steps are performed on the computer or other programmable device to generate computer-implemented processing, which is executed on the computer or other programmable device
  • the instructions provide steps for implementing the functions specified in one block or multiple blocks of the flowchart one flow or multiple flows and/or block diagrams.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种能源消耗预测方法及装置，该方法包括：将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型（S101）；基于ARIMA模型，确定第一能源消耗预测值，基于SVM模型，确定第二能源消耗预测值（S102）；根据第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值（S103）。利用ARIMA模型和SVM模型的优点，使确定的目标能源消耗预测值更准确。

Description

一种能源消耗预测方法及装置 技术领域
本发明涉及能源消耗预测技术领域，尤其涉及一种能源消耗预测方法及装置。
背景技术
随着我国城镇化的快速推进，总建筑量必定持续增长，建筑能源消耗用能比例增加是发展的必然趋势。我国建筑每平方米能源消耗约是气候相近的发达国家耗能的3倍，随着经济的持续快速增长，我国的建筑能源消耗也日益攀升，我国已成为三大能源消耗国之一，能源紧张的问题日益明显。如何准确预测建筑能源消耗已成为我国急需解决的问题。
现在的建筑能源消耗预测技术，一般是基于SVM模型进行能源消耗预测，或者基于改进的SVM模型进行能源消耗预测。但是，现有技术中在进行能源消耗预测时，并未考虑时间序列问题，因此现有技术中能源消耗预测的准确性较低。
发明内容
本发明实施例提供了一种能源消耗预测方法及装置，用以解决现有技术中能源消耗预测不准确的问题。
本发明实施例提供了一种能源消耗预测方法，所述方法包括：
将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；
基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；
根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测 时间的目标能源消耗预测值。
进一步地，所述根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值包括：
确定最新获取的能源消耗真实值；
根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；
根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；
根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
进一步地，所述ARIMA模型的训练过程包括：
针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
进一步地，所述SVM模型的训练过程包括：
针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
进一步地，所述第一数据和第三数据包括能源消耗值。
进一步地，所述第二数据和第四数据包括能源消耗值和天气数据。
另一方面，本发明实施例提供了一种能源消耗预测装置，所述装置包括：
输入模块，用于将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；
第一确定模块，用于基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；
第二确定模块，用于根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
进一步地，所述第二确定模块，具体用于确定最新获取的能源消耗真实值；根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
进一步地，所述装置还包括：
第一训练模块，用于针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
进一步地，所述装置还包括：
第二训练模块，用于针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
本发明实施例提供了一种能源消耗预测方法及装置，所述方法包括：将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
由于在本发明实施例中，在进行能源消耗预测时，基于ARIMA模型，确定第一能源消耗预测值，基于SVM模型，确定第二能源消耗预测值。ARIMA模型能够很好的拟合时间序列，SVM模型能够很好地适用于小样 本、非线性及高维模式识别问题，结合ARIMA模型确定的第一能源消耗预测值和SVM模型确定的第二能源消耗预测值，确定出目标能源消耗预测值。充分利用ARIMA模型和SVM模型的优点，使得确定的目标能源消耗预测值更准确。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简要介绍，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，对于本领域的普通技术人员来讲，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为本发明实施例1提供的能源消耗预测过程示意图；
图2为本发明实施例2提供的确定目标能源消耗预测值的示意图；
图3为本发明实施例2提供的能源消耗预测流程示意图；
图4为本发明实施例提供的能源消耗预测装置结构示意图。
具体实施方式
下面将结合附图对本发明作进一步地详细描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，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它实施例，都属于本发明保护的范围。
实施例1：
图1为本发明实施例提供的能源消耗预测过程示意图，该过程包括以下步骤：
S101：将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型。
S102：基于所述ARIMA模型，确定第一能源消耗预测值，基于所述 SVM模型，确定第二能源消耗预测值。
S103：根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
本发明实施例提供的能源消耗预测方法应用于电子设备，该电子设备可以是PC、平板电脑等设备。本发明实施例提供的能源消耗预测方法包括但不限定是对大型公共建筑的能源消耗预测。
电子设备可以获取第一时间长度内的第一数据和第二数据，第一时间长度可以是三天、四天等，第一数据可以是第一时间长度内的能源消耗。第二数据可以和第一数据相同或不同。如果第二数据和第一数据不同，则第二数据可以是第一时间长度内的能源消耗以及第一时间长度内的天气数据。其中，天气数据可以是每天的平均温度、平均湿度等数据。
电子设备中保存有预先训练完成的自回归积分滑动平均ARIMA模型和支持向量机SVM模型。电子设备在获取到第一时间长度内的第一数据和第二数据后，将第一数据输入ARIMA模型，将第二数据输入SVM模型。由于ARIMA模型能够很好的拟合时间序列，SVM模型能够很好地适用于小样本、非线性及高维模式识别问题，因此基于ARIMA模型，确定的第一能源消耗预测值很好的拟合了时间序列，基于SVM模型，确定的第二能源消耗预测值适用于小样本、非线性及高维模式识别问题。
电子设备在确定第一能源消耗预测值和第二能源消耗预测值后，根据第一能源消耗预测值和第二能源消耗预测值确定第一预测时间的目标能源消耗预测值。其中，第一预测时间可以是第一时间长度之后的第一天，也可以是第一时间长度之后的第二天等。例如第一时间长度是2018年10月1日至2018年10月3日，第一预测时间可以是2018年10月4日，也可以是2018年10月5日。具体的，第一时间长度和第一预测时间的对应关系是由训练的ARIMA模型和SVM模型决定的。如果在训练ARIMA模型和SVM模型时，预测的是第一时间长度之后的第一天，则第一预测时间也就是第一时间长度之后的第一天；如果在训练ARIMA模型和SVM模型 时，预测的是第一时间长度之后的第二天，则第一预测时间也就是第一时间长度之后的第二天。
电子设备根据第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值，可以是取第一能源消耗预测值和第二能源消耗预测值的平均值作为第一预测时间的目标能源消耗预测值，也可以根据经验确定第一能源消耗预测值和第二能源消耗预测值重要程度，根据重要程度不同为第一能源消耗预测值和第二能源消耗预测值分配对应的权重，其中重要程度越大，对应的权重越大。然后根据第一能源消耗预测值、第二能源消耗预测值以及对应的权重，进行加权平均计算得到第一预测时间的目标能源消耗预测值。
由于在本发明实施例中，在进行能源消耗预测时，基于ARIMA模型，确定第一能源消耗预测值，基于SVM模型，确定第二能源消耗预测值。ARIMA模型能够很好的拟合时间序列，SVM模型能够很好地适用于小样本、非线性及高维模式识别问题，结合ARIMA模型确定的第一能源消耗预测值和SVM模型确定的第二能源消耗预测值，确定出目标能源消耗预测值。充分利用ARIMA模型和SVM模型的优点，使得确定的目标能源消耗预测值更准确。
实施例2：
为了使确定的目标能源消耗预测值更准确，在上述实施例的基础上，在本发明实施例中，所述根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值包括：
确定最新获取的能源消耗真实值；
根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；
根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；
根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能 源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
在本发明实施例中，电子设备在确定第一能源消耗预测值和第二能源消耗预测值之后，为了能够准确确定出第一预测时间的目标能源消耗预测值，首先确定出最新获取的能源消耗真实值。例如想要预测2018年10月4日的能源消耗值，最新获取的能源消耗真实值可以是2018年10月3日的能源消耗真实值。
电子设备根据第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据第二能源消耗预测值和能源消耗真实值，确定第二预测误差。具体的，电子设备计算能源消耗真实值与第一能源消耗预测值的差值作为第一预测误差，计算能源消耗真实值与第二能源消耗预测值的差值作为第二预测误差。如果确定的第一预测误差或第二预测误差为负值，则取第一预测误差或第二预测误差的绝对值。保证第一预测误差和第二预测误差为正值。
电子设备根据第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重。具体的，可以计算第一预测误差和第二预测误差的和值，然后计算第一预测误差与该和值的比值作为第一权重，计算第二预测误差与该和值的比值作为第二权重。在确定出第一权重和第二权重之后，采用第一权重和第二权重对第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
另外，在计算出第一预测误差和第二预测误差的和值之后，也可以计算该和值与第一预测误差的比值作为第一中间参数，计算该和值与第二预测误差的比值作为第二中间参数。计算第一中间参数和第二中间参数的和值，然后计算第一中间参数与该和值的比值作为第一权重，计算第二中间参数与该和值的比值作为第二权重。在确定出第一权重和第二权重之后，采用第一权重和第二权重对第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
图2为本发明实施例提供的确定目标能源消耗预测值的示意图。如图2所示，首先确定第一能源消耗预测值P1、第二能源消耗预测值P2以及最新获取的能源消耗真实值Act。然后计算第一预测误差E1＝|Act-P1|，计算第二预测误差E2＝|Act-P2|。再计算第一预测误差和第二预测误差的和值＝E1+E2。计算第一中间参数为α＝S/E1，计算第二中间参数为β＝S/E2。计算第一中间参数和第二中间参数的和值α+β，然后计算第一权重为W1＝α/(α+β)，计算第二权重为W2＝β/(α+β)。在确定出第一权重W1和第二权重W2之后，确定目标能源消耗预测值为Out＝W1×P1+W2×P2。
由于在本发明实施例中，电子设备根据第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据第二能源消耗预测值和能源消耗真实值，确定第二预测误差；然后根据第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重。最终根据第一权重和第二权重，对第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。因此可以使得确定的目标能源消耗预测值更准确。
图3为本发明实施例提供的能源消耗预测流程示意图，如图3所示，由于获取的第一时间长度内的第一数据和第二数据中有可能存在干扰数据，干扰数据会影响最终确定的目标能源消耗预测值的准确性，为了使确定的目标能源消耗预测值更准确，在将第一数据输入ARIMA模型，以及将第二数据输入SVM模型之前，还需要对第一数据和第二数据进行数据清洗处理。具体的，电子设备中可以保存较小的第一阈值和较大的第二阈值，将第一数据中和第二数据中小于第一阈值和大于第二阈值的数据作为干扰数据，将干扰数据从第一数据和第二数据中滤除，然后将剩余的第一数据输入ARIMA模型，将剩余的第二数据输入SVM模型。基于ARIMA模型得到第一能源消耗预测值，基于SVM模型，得到第二能源消耗预测值，然后计算第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应 的第二权重，根据第一权重和第二权重，对第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
实施例3：
在上述各实施例的基础上，在本发明实施例中，所述ARIMA模型的训练过程包括：
针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
在本发明实施例中，将用于训练ARIMA模型的训练集作为第一训练集中，将第一训练集中的数据作为第三数据。第一数据和第三数据包括能源消耗值。电子设备将第一训练集中第三数据划分为每个第二时间长度内的第三数据，第二时间长度例如是三天，则可以将第一训练集中第三数据划分为2018年1月1日至2018年1月3日的第三数据、2018年1月2日至2018年1月4日的第三数据、2018年1月3日至2018年1月4日的第三数据等等。然后针对每个第二时间长度，确定该第二时间长度对应的第二预测时间。具体的，该第二时间长度对应的第二预测时间可以是该第二时间长度之后的第一天，也可以是该第二时间长度之后的第二天等，只要预先设置好即可。
例如，2018年1月1日至2018年1月3日对应的第二预测时间可以是2018年1月4日，2018年1月2日至2018年1月4日对应的第二预测时间可以是2018年1月5日等。
在对ARIMA模型进行训练时，针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型。根据ARIMA模型输出的第二预测时间的能源消耗预测值与能源消耗真实值的差值对ARIMA模型的参数进行调整，直到ARIMA模型训练完成。
实施例4：
在上述各实施例的基础上，在本发明实施例中，所述SVM模型的训练过程包括：
针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
在本发明实施例中，将用于训练SVM模型的训练集作为第二训练集中，将第二训练集中的数据作为第四数据。第二数据和第四数据包括能源消耗值和天气数据。
电子设备将第二训练集中第三数据划分为每个第二时间长度内的第四数据，第二时间长度例如是三天，则可以将第二训练集中第四数据划分为2018年1月1日至2018年1月3日的第三数据、2018年1月2日至2018年1月4日的第三数据、2018年1月3日至2018年1月4日的第三数据等等。然后针对每个第二时间长度，确定该第二时间长度对应的第二预测时间。具体的，该第二时间长度对应的第二预测时间可以是该第二时间长度之后的第一天，也可以是该第二时间长度之后的第二天等，只要预先设置好即可。
例如，2018年1月1日至2018年1月3日对应的第二预测时间可以是2018年1月4日，2018年1月2日至2018年1月4日对应的第二预测时间可以是2018年1月5日等。
在对SVM模型进行训练时，针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型。根据SVM模型输出的第二预测时间的能源消耗预测值与能源消耗真实值的差值对SVM模型的参数进行调整，直到SVM模型训练完成。
其中，对ARIMA模型和SVM模型训练的过程属于现有技术，在此不再对该过程进行赘述。
图4为本发明实施例提供的能源消耗预测装置结构示意图，所述装置 包括：
输入模块41，用于将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；
第一确定模块42，用于基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；
第二确定模块43，用于根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
所述第二确定模块43，具体用于确定最新获取的能源消耗真实值；根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
所述装置还包括：
第一训练模块44，用于针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
所述装置还包括：
第二训练模块45，用于针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
本发明实施例提供了一种能源消耗预测方法及装置，所述方法包括：将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训 练完成的支持向量机SVM模型；基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
由于在本发明实施例中，在进行能源消耗预测时，基于ARIMA模型，确定第一能源消耗预测值，基于SVM模型，确定第二能源消耗预测值。ARIMA模型能够很好的拟合时间序列，SVM模型能够很好地适用于小样本、非线性及高维模式识别问题，结合ARIMA模型确定的第一能源消耗预测值和SVM模型确定的第二能源消耗预测值，确定出目标能源消耗预测值。充分利用ARIMA模型和SVM模型的优点，使得确定的目标能源消耗预测值更准确。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能 的步骤。
尽管已描述了本发明的优选实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例作出另外的变更和修改。所以，所附权利要求意欲解释为包括优选实施例以及落入本发明范围的所有变更和修改。
显然，本领域的技术人员可以对本发明进行各种改动和变型而不脱离本发明的精神和范围。这样，倘若本发明的这些修改和变型属于本发明权利要求及其等同技术的范围之内，则本发明也意图包含这些改动和变型在内。

Claims (10)

  1. 一种能源消耗预测方法，其特征在于，所述方法包括：
    将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；
    基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；
    根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
  2. 如权利要求1所述的方法，其特征在于，所述根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值包括：
    确定最新获取的能源消耗真实值；
    根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；
    根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；
    根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
  3. 如权利要求1所述的方法，其特征在于，所述ARIMA模型的训练过程包括：
    针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
  4. 如权利要求1所述的方法，其特征在于，所述SVM模型的训练过程包括：
    针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
  5. 如权利要求3所述的方法，其特征在于，所述第一数据和第三数据包括能源消耗值。
  6. 如权利要求4所述的方法，其特征在于，所述第二数据和第四数据包括能源消耗值和天气数据。
  7. 一种能源消耗预测装置，其特征在于，所述装置包括：
    输入模块，用于将获取的第一时间长度内的第一数据输入预先训练完成的自回归积分滑动平均ARIMA模型，将获取的所述第一时间长度内的第二数据输入预先训练完成的支持向量机SVM模型；
    第一确定模块，用于基于所述ARIMA模型，确定第一能源消耗预测值，基于所述SVM模型，确定第二能源消耗预测值；
    第二确定模块，用于根据所述第一能源消耗预测值和第二能源消耗预测值，确定第一预测时间的目标能源消耗预测值。
  8. 如权利要求7所述的装置，其特征在于，所述第二确定模块，具体用于确定最新获取的能源消耗真实值；根据所述第一能源消耗预测值和能源消耗真实值，确定第一预测误差；根据所述第二能源消耗预测值和能源消耗真实值，确定第二预测误差；根据所述第一预测误差和第二预测误差，确定第一能源消耗预测值对应的第一权重和第二能源消耗预测值对应的第二权重；根据所述第一权重和第二权重，对所述第一能源消耗预测值和第二能源消耗预测值进行加权平均处理，确定目标能源消耗预测值。
  9. 如权利要求7所述的装置，其特征在于，所述装置还包括：
    第一训练模块，用于针对第一训练集中每个第二时间长度内的第三数据，将该第二时间长度内的第三数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入ARIMA模型，对所述ARIMA模型进行训练。
  10. 如权利要求7所述的装置，其特征在于，所述装置还包括：
    第二训练模块，用于针对第二训练集中每个第二时间长度内的第四数据，将该第二时间长度内的第四数据，和该第二时间长度对应的第二预测时间的能源消耗真实值输入SVM模型，对所述SVM模型进行训练。
PCT/CN2018/120746 2018-12-12 2018-12-12 一种能源消耗预测方法及装置 WO2020118586A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2018/120746 WO2020118586A1 (zh) 2018-12-12 2018-12-12 一种能源消耗预测方法及装置
JP2021531541A JP2022510667A (ja) 2018-12-12 2018-12-12 エネルギー消費量を予測するための方法および装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/120746 WO2020118586A1 (zh) 2018-12-12 2018-12-12 一种能源消耗预测方法及装置

Publications (1)

Publication Number Publication Date
WO2020118586A1 true WO2020118586A1 (zh) 2020-06-18

Family

ID=71077056

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/120746 WO2020118586A1 (zh) 2018-12-12 2018-12-12 一种能源消耗预测方法及装置

Country Status (2)

Country Link
JP (1) JP2022510667A (zh)
WO (1) WO2020118586A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113920166A (zh) * 2021-10-29 2022-01-11 广州文远知行科技有限公司 一种选择物体运动模型方法、装置、交通工具及存储介质
CN116754828A (zh) * 2023-08-21 2023-09-15 济南瑞源智能城市开发有限公司 一种智慧隧道的能耗监测方法、设备及介质
WO2023218422A1 (en) * 2022-05-13 2023-11-16 Universita' Degli Studi Di Padova System and method for overvoltage protection of a load in a wireless power receiver
CN117709987A (zh) * 2024-02-06 2024-03-15 浙江省发展规划研究院 企业年度基准能耗的核定方法、装置、电子设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140095415A1 (en) * 2012-09-28 2014-04-03 Electronics And Telecommunications Research Institute Apparatus and method for forecasting energy consumption
CN104331635A (zh) * 2014-11-24 2015-02-04 国网吉林省电力有限公司信息通信公司 电力光纤通信光功率预测的方法
CN105205566A (zh) * 2015-10-08 2015-12-30 中国南方电网有限责任公司电网技术研究中心 能源消耗量预测方法和系统
CN105678414A (zh) * 2015-12-31 2016-06-15 远光软件股份有限公司 对资源消耗量进行预测的数据处理方法
CN108537581A (zh) * 2018-03-27 2018-09-14 四川大学 基于gmdh选择性组合的能源消费量时间序列预测方法及装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0784983A (ja) * 1993-09-17 1995-03-31 Nippon Telegr & Teleph Corp <Ntt> 時系列データの予測装置
JP5511698B2 (ja) * 2011-01-20 2014-06-04 日本電信電話株式会社 空調機連係制御システム、空調機連係制御方法および空調機連係制御プログラム
JP2013255390A (ja) * 2012-06-08 2013-12-19 Mitsubishi Electric Corp 電力需要予測装置及び電力需要予測方法
JP6831280B2 (ja) * 2017-03-24 2021-02-17 株式会社日立製作所 予測システム、及び、予測方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140095415A1 (en) * 2012-09-28 2014-04-03 Electronics And Telecommunications Research Institute Apparatus and method for forecasting energy consumption
CN104331635A (zh) * 2014-11-24 2015-02-04 国网吉林省电力有限公司信息通信公司 电力光纤通信光功率预测的方法
CN105205566A (zh) * 2015-10-08 2015-12-30 中国南方电网有限责任公司电网技术研究中心 能源消耗量预测方法和系统
CN105678414A (zh) * 2015-12-31 2016-06-15 远光软件股份有限公司 对资源消耗量进行预测的数据处理方法
CN108537581A (zh) * 2018-03-27 2018-09-14 四川大学 基于gmdh选择性组合的能源消费量时间序列预测方法及装置

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113920166A (zh) * 2021-10-29 2022-01-11 广州文远知行科技有限公司 一种选择物体运动模型方法、装置、交通工具及存储介质
CN113920166B (zh) * 2021-10-29 2024-05-28 广州文远知行科技有限公司 一种选择物体运动模型方法、装置、交通工具及存储介质
WO2023218422A1 (en) * 2022-05-13 2023-11-16 Universita' Degli Studi Di Padova System and method for overvoltage protection of a load in a wireless power receiver
CN116754828A (zh) * 2023-08-21 2023-09-15 济南瑞源智能城市开发有限公司 一种智慧隧道的能耗监测方法、设备及介质
CN116754828B (zh) * 2023-08-21 2023-12-01 济南瑞源智能城市开发有限公司 一种智慧隧道的能耗监测方法、设备及介质
CN117709987A (zh) * 2024-02-06 2024-03-15 浙江省发展规划研究院 企业年度基准能耗的核定方法、装置、电子设备及介质

Also Published As

Publication number Publication date
JP2022510667A (ja) 2022-01-27

Similar Documents

Publication Publication Date Title
WO2020118586A1 (zh) 一种能源消耗预测方法及装置
CN110298501B (zh) 基于长短时记忆神经网络的电负荷预测方法
CN111951128B (zh) 一种节能环保的建筑施工方法和装置
CN110705772B (zh) 区域电网风力发电功率预测优化方法和装置
CN105243502B (zh) 一种基于径流区间预测的水电站调度风险评估方法及系统
CN104217258B (zh) 一种电力负荷条件密度预测方法
CN103942461A (zh) 基于在线贯序极限学习机的水质参数预测方法
WO2022110582A1 (zh) 一种基于基流分割和人工神经网络模型的径流预报方法
CN105354620A (zh) 一种风机发电功率的预测方法
CN107705556A (zh) 一种基于支持向量机和bp神经网络结合的交通流预测方法
CN110631221A (zh) 一种空调的控制方法、装置、终端及存储介质
CN103218536A (zh) 跨越式发展地区需水预测方法
CN109657878A (zh) 一种空调负荷预测方法及装置
CN104598998A (zh) 基于经济增长指标的能源需求量预测方法
CN103279030B (zh) 基于贝叶斯框架的动态软测量建模方法及装置
CN110751312A (zh) 一种基于多因子的系统动力学生活需水预测方法和系统
CN113053536A (zh) 一种基于隐马尔科夫模型的传染病预测方法、系统和介质
CN112686693A (zh) 电力现货市场边际电价预测方法、系统、设备及存储介质
CN112596126A (zh) 风速预测模型训练方法、预测方法、装置及电子设备
CN111310960A (zh) 一种能源消耗预测方法及装置
CN106355264A (zh) 一种参考作物腾发量组合预测方法
CN115422826A (zh) 数据中心智能节能调控方法、装置、设备及存储介质
CN115034159A (zh) 一种海上风电场的功率预测方法、装置、存储介质及系统
CN111859783B (zh) 水压预测方法、系统、存储介质、设备、城市供水系统
CN114492923A (zh) 一种长时间尺度功率预测方法

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: 18942703

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021531541

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18942703

Country of ref document: EP

Kind code of ref document: A1