WO2021133253A1 - Method and apparatus for predicting energy consumption, and device and storage medium thereof - Google Patents

Method and apparatus for predicting energy consumption, and device and storage medium thereof Download PDF

Info

Publication number
WO2021133253A1
WO2021133253A1 PCT/SG2020/050775 SG2020050775W WO2021133253A1 WO 2021133253 A1 WO2021133253 A1 WO 2021133253A1 SG 2020050775 W SG2020050775 W SG 2020050775W WO 2021133253 A1 WO2021133253 A1 WO 2021133253A1
Authority
WO
WIPO (PCT)
Prior art keywords
energy consumption
sample
network
simulated
real
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/SG2020/050775
Other languages
French (fr)
Inventor
Qi Cheng
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.)
Shanghai Envision Digital Co Ltd
Univers Pte Ltd
Original Assignee
Shanghai Envision Digital Co Ltd
Envision Digital International Pte Ltd
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 Shanghai Envision Digital Co Ltd, Envision Digital International Pte Ltd filed Critical Shanghai Envision Digital Co Ltd
Publication of WO2021133253A1 publication Critical patent/WO2021133253A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0475—Generative networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/094—Adversarial learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/096—Transfer learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"

Definitions

  • step 102 includes the following sub-steps.
  • FIG. 7 shows a block diagram of an apparatus for predicting energy consumption according to an embodiment of the present disclosure
  • the apparatus has the function of practicing the above method examples.
  • the function may be implemented by hardware or implemented by executing corresponding software by hardware.
  • the apparatus may be a computer device or may be disposed in a computer device.
  • the apparatus 700 may include a first data acquiring module 701, a real sample acquiring module 702, a simulated sample acquiring module 703 and a first energy consumption predicting module 704.
  • the high -capacity storage device 907 is connected to the central processing unit 901 by a high -capacity storage controller (not shown) connected to the system bus 905.
  • the high -capacity storage device 907 and a computer-readable medium associated therewith provide non-volatile storage for the computer device 900. That is, the high-capacity storage device 907 may include the computer-readable medium (not shown), such as a hard disk or a compact disc read-only memory (CD-ROM) driver.
  • CD-ROM compact disc read-only memory

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Public Health (AREA)
  • Probability & Statistics with Applications (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Disclosed are a method and apparatus for predicting energy consumption. The method includes: acquiring first history data; acquiring a real energy consumption sample of a first object based on the first history data; acquiring a simulated energy consumption sample of the first object based on the real energy consumption sample; and calculating predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model.

Description

METHOD AND APPARATUS FOR PREDICTING ENERGY
CONSUMPTION, AND DEVICE AND STORAGE MEDIUM
THEREOF
TECHNICAU FIEUD
[0001] Embodiments of the present disclosure relate to the field of energy Internet technologies, and in particular, relate to a method and apparatus for predicting energy consumption, and a device and a storage medium thereof.
BACKGROUND
[0002] Currently, energy consumption prediction is an indispensable part of the energy Internet. [0003] Before prediction of energy consumption in a specific region, it is possible to acquire corresponding predicted energy consumption by predicting energy consumption conditions of this region in a future time period based on a large amount of history energy consumption data in this region, such as a history electricity consumption amount.
[0004] However, energy consumption prediction results of newly-built enterprises are inaccurate due to insufficient history energy consumption data of these enterprises.
SUMMARY
[0005] Embodiments of the present disclosure provide a method and apparatus for predicting energy consumption, and a device and a storage medium thereof, which may solve the technical problem that energy consumption prediction results of newly-built enterprises are inaccurate in the related art.
[0006] In one aspect, embodiments of the present disclosure provide a method for predicting energy consumption. The method is applicable to a computer device, and includes:
[0007] acquiring first history data, wherein the first history data includes history energy consumption data of a first object;
[0008] acquiring a real energy consumption sample of the first object based on the first history data;
[0009] acquiring a simulated energy consumption sample of the first object based on the real energy consumption sample, wherein the simulated energy consumption sample includes virtual energy consumption data acquired by expansion based on the real energy consumption sample; and
[0010] calculating predicted eneigy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model. [0011] In another aspect, embodiments of the present disclosure provide an apparatus for predicting energy consumption. The apparatus is applicable to a computer device, and includes: [0012] a first data acquiring module, configured to acquire first history data, wherein the first history data includes history eneigy consumption data of a first object;
[0013] a real sample acquiring module, configured to acquire a real energy consumption sample of the first object based on the first history data;
[0014] a simulated sample acquiring module, configured to acquire a simulated energy consumption sample of the first object based on the real energy consumption sample, wherein the simulated energy consumption sample includes virtual energy consumption data acquired by expansion by a generative adversarial network based on the real energy consumption sample; and [0015] a first energy consumption predicting module, configured to calculate predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model.
[0016] In still another aspect, embodiments of the present disclosure provide a computer device. The computer device includes a processor and a memory storing a computer program. The computer program, when loaded and run by the processor, causes the processor to perform the above method.
[0017] In yet still another aspect, embodiments of the present disclosure provide a non-transitory computer-readable storage medium. The storage medium stores a computer program. The computer program, when loaded and run by a processor, causes the processor to perform the above method.
[0018] In yet still another aspect, embodiments of the present disclosure provide a computer program product. The computer program product, when loaded and run by a computer device, causes the computer device to perform the above method.
[0019] The technical solutions according to the embodiments of the present disclosure achieve the following beneficial effects.
[0020] The simulated energy consumption sample is acquired by expansion of the real energy consumption sample, such that the technical problem that energy consumption prediction results of newly-built enterprises are inaccurate due to less history energy consumption data of these enterprises in related art is solved and the number of samples that can be used by the energy consumption prediction model is increased. Therefore, the accuracy of the energy consumption prediction result is improved.
BRIEF DESCRIPTION OF THE DRAWINGS [0021] For clearer descriptions of the technical solutions in the embodiments of the present disclosure, the following briefly introduces the accompanying drawings required for describing the embodiments. Apparently, the accompanying drawings in the following description show merely some embodiments of the present disclosure, and a person of ordinary skill in the art may still derive other drawings from these accompanying drawings without creative efforts.
[0022] FIG. 1 is a flowchart of a method for predicting energy consumption according to an embodiment of the present disclosure;
[0023] FIG. 2 exemplarily shows a diagram of a method for storing first history data;
[0024] FIG. 3 exemplarily shows a diagram of a method for generating a simulated energy consumption sample;
[0025] FIG. 4 exemplarily shows a diagram of an energy consumption prediction model;
[0026] FIG. 5 exemplarily shows a diagram of the training of a neural network;
[0027] FIG. 6 exemplarily shows a diagram of a method for training a generative adversarial network;
[0028] FIG. 7 is a block diagram of an apparatus for predicting energy consumption according to an embodiment of the present disclosure;
[0029] FIG. 8 is a block diagram of an apparatus for predicting energy consumption according to an embodiment of the present disclosure; and
[0030] FIG. 9 is a structural block diagram of a computer device according to an embodiment of the present disclosure.
DETAILED DESCRIPTION
[0031] For clearer descriptions of the objectives, technical solutions and advantages in the present disclosure, embodiments of the present disclosure are described in detail below in combination with the accompanying drawings.
[0032] In the technical solutions according to the embodiments of the present disclosure, a generative adversarial network is configured to perform data extension on a first object. In some embodiments, the generative adversarial network acquires a corresponding simulated energy consumption sample of the first object based on a real energy consumption sample of the first object so as to achieve the purpose of data extension. In some embodiments, the generative adversarial network includes a generator network and a discriminator network. The generator network is configured to generate a candidate simulated energy consumption sample and the discriminator network is configured to determine whether the authenticity of the candidate simulated energy consumption sample satisfies the condition of being a simulated energy consumption sample based on the real energy consumption sample.
[0033] An energy consumption prediction model involved in the embodiment of the present disclosure refers to a machine learning model acquired by machine learning training and configured to predict energy consumption of the first object. The energy consumption prediction model may be a deep-leaming energy consumption prediction model, i.e., an energy consumption prediction model constructed based on a deep neural network, such as an energy consumption prediction model constructed based on a deep convolutional neural network. This classification task model may be configured to predict energy consumption of any object, which is not limited in the embodiment of the present disclosure.
[0034] In the method according to the embodiment of the present disclosure, an execution subject of various steps may be a computer device. The computer device refers to an electronic device having data calculation, processing and storage capabilities, such as a personal computer (PC) or a server. It should be noted that the generative adversarial network and the energy consumption prediction model may be disposed on the same computer device or disposed on different computer devices, which is not limited in the embodiment of the present disclosure. [0035] Referring to FIG. 1, which shows a flowchart of a method for predicting energy consumption according to an embodiment of the present disclosure, the method may include the following steps (101-104).
[0036] In step 101, first history data is acquired.
[0037] The first history data refers to history energy consumption data of a first object. The history energy consumption data refers to history energy consumption amount of the first object. The energy may be electric energy, water energy, fuel, or the like, which is not limited in the embodiments of the present disclosure. The first object includes a region using the energy, which may be an enterprise, a plant, a residential area or the like. In the embodiments of the present disclosure, the computer device determines the first object according to a region in which energy consumption is predicted. Exemplarily, when energy consumption is predicted in a region A, if there is an enterprise a in the region A, the first object is the enterprise a. The computer device acquires predicted energy consumption of the region A by predicting energy consumption of the enterprise a. If both the enterprise a and a plant b exist in the region A, the first object is the enterprise a or the plant b. The computer device may predict energy consumption of the enterprise a and the plant b respectively and then the sum of predicted energy consumption of the enterprise a and the plant B is predicted energy consumption of the region A; or the first object is the enterprise a and the plant b, and the computer device may acquire predicted energy consumption of the region A by predicting energy consumption of the enterprise a and the plant b as a whole.
[0038] In some embodiments, the first history data is acquired by a load meter. For example, the history electricity consumption data of the first object is acquired by a load electric meter. For example, the history water consumption data of the first object is acquired by a load water meter. Optionally, after acquiring the first history data, the load meter sends the first history data to the computer device at certain time interval over a network. In some embodiments, the time interval is determined by the computer device according to actual conditions, and may be 10 ms, 1 hour, 1 day, 7 days, 14 days or the like, which is not limited in the embodiments of the present disclosure. For the accuracy of predicted energy consumption, the first history data may further include, but is not limited to, at least one of weather, information about holidays and festivals, temperature, and the like.
[0039] It should be noted that the computer device records a time period corresponding to the first history data while acquiring the first history data. In some embodiments, referring to FIG. 2, by taking the first history data being the history electricity consumption data as an example, the first history data may be stored in a coordinate system 20 constructed by the computer device. In the coordinate system 20, the abscissa represents time and the ordinate represents real history electricity consumption of the first object.
[0040] In step 102, a real energy consumption sample of the first object is acquired based on the first history data.
[0041] The real energy consumption sample includes real history energy consumption data of the first object. In some embodiments, after acquiring the first history data, the computer device acquires the corresponding real energy consumption sample by standardization and time period division on the first history data.
[0042] A method for acquiring the real energy consumption sample by the computer device based on the first history data is introduced hereinafter. In some embodiments, step 102 includes the following sub-steps.
[0043] 1. The first history data is standardized.
[0044] The standardizing is to map the history energy consumption of the first object into a preset value range. The preset value range is determined by the computer device according to actual conditions, and may be (0, 1), (0, 2), (1, 2), or the like, which is not limited in the embodiments of the present disclosure. It should be noted that the standardizing is only directed to history energy consumption data corresponding to the first history data.
[0045] In some embodiments, by taking the preset value range being (0, 1) as an example,
X represents any history energy consumption corresponding to the first history data, such as history energy consumption corresponding to a point 21 in FIG. 2. The any history energy consumption standardized Xscaled is
Figure imgf000007_0001
X
[0047] wherein max represents the maximum history energy consumption in the first history
X data and 111111 represents the minimum history energy consumption in the first history data. [0048] 2. A prediction time period corresponding to the predicted energy consumption is acquired.
[0049] The predicted energy consumption includes future energy consumption data acquired by predicting energy consumption of the first object. The prediction time period includes a time period in which the first object uses the predicted energy consumption. The prediction time period may be 1 hour, 1 day, 7 days, 14 days or the like, which is not limited in the embodiments of the present disclosure. In some embodiments, after standardizing the first history data, the computer device acquires the corresponding prediction time period based on the predicted energy consumption of the first object.
[0050] 3. The real energy consumption sample is acquired by segmenting the standardized first history data based on the prediction time period.
[0051] The real energy consumption sample includes a history energy consumption sample acquired by standardizing and segmenting the history energy consumption data corresponding to the first object. It should be noted that the real energy consumption sample includes at least one item of real energy consumption data. A duration corresponding to the real energy consumption data is the same as the duration of the prediction time period. In some embodiments, the computer device acquires the duration of the prediction time period based on the prediction time period and segments the standardized first history data.
[0052] In some embodiments, the computer device statically segments the standardized first history data based on the duration of the prediction time period, namely, segments the first history data based on a time sequence. For example, if the duration of the predication time period is 7 days and the duration corresponding to the first history data is 14 days, the real energy consumption sample acquired by segmenting the first history data includes two items of real energy consumption data, the first real energy consumption data is energy consumption data of the first 7 days, and the second real energy consumption data is energy consumption data of the 8th to 14th days. In another possible embodiment, the computer device dynamically segments the standardized first history data based on the duration of the prediction time period, namely, randomly segments the first history data based on a preset time period. For example, if the duration of the predication time period is 7 days and the duration corresponding to the first history data is 14 days, the real energy consumption sample acquired by segmenting the first history data includes eight items of real energy consumption data, the first real energy consumption data is energy consumption data of the first 7 days, the second real energy consumption data is energy consumption data of the 3 rd to 8th days, the third real energy consumption data is energy consumption data of the 3rd to 9th days, and by such analogy, the eighth real energy consumption data is energy consumption data of the 8 th to 14th days.
[0053] In step 103, a simulated energy consumption sample of the first object is acquired based on the real energy consumption sample.
[0054] The simulated energy consumption sample includes virtual energy consumption data acquired by expansion on the basis of the real energy consumption sample and includes at least one item of simulated energy consumption data. In some embodiments, the amount of the simulated energy consumption data in the simulated energy consumption sample may be the same or different from the amount of the real energy consumption data in the real energy consumption sample. In some embodiments, the computer device calculates the simulated energy consumption sample based on the real energy consumption sample by calling a generative adversarial network. [0055] A method for acquiring the simulated energy consumption sample is introduced below. In some embodiments, step 103 includes the following several sub-steps. [0056] 1. A first randomly-generated noise vector is acquired.
[0057] The first noise vector refers to a random vector for generating the simulated energy consumption sample. In some embodiments, the first noise vector may be a 16-dimensional vector or a 32-dimensional vector generated by a distribution function. The first noise vector may also be a 16 -dimensional vector or a 32 -dimensional vector generated randomly. [0058] 2. The simulated energy consumption sample is generated based on the first noise vector by a generator network in the generative adversarial network.
[0059] In some embodiments, the generative adversarial network includes the generator network and a discriminator network. In the embodiment of the present disclosure, after acquiring the first noise vector, the computer device generates a candidate simulated energy consumption sample based on the first noise vector by the generator network. The candidate simulated energy consumption sample includes at least one item of candidate simulated energy consumption data, the amount of which may be the same or different from the amount of the real energy consumption data in the real energy consumption sample.
[0060] It should be noted that before the generative adversarial network generates the simulated energy consumption sample, the generative adversarial network needs to be trained. A method for training the generative adversarial network is introduced herein in detail, which is not repeated herein.
[0061] In step 104, predicted energy consumption of the first object is calculated based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model.
[0062] In some embodiments, after acquiring the simulated energy consumption sample, the computer device calculates predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling the energy consumption prediction model. The energy consumption prediction model is acquired by training based on history energy consumption data of other objects.
[0063] A method for training the energy consumption prediction model is introduced below. Optionally, the following several steps are further included before step 104.
[0064] 1. Second history data is acquired.
[0065] The second history data includes history energy consumption data of other objects associated with the first object. The other objects include objects having the same feature as the first object. The feature may be an actual function, a people flow density, a geographical location or the like, which is not limited in the embodiment of the present disclosure. For example, if the first object is an enterprise A, the other objects may be other enterprises having the same actual function as the enterprise A or other enterprises having similar staff to the enterprise A. In some embodiments, the computer device acquires the second history data by the load meter.
[0066] 2. An initial energy consumption prediction model is acquired by training based on the second history data.
[0067] In the embodiments of the present disclosure, the computer device acquires the initial energy consumption prediction model by training the second history data. In some embodiments, the computer device constructs a recurrent neural network and acquires the initial energy consumption prediction model by training the second history data.
[0068] 3. A bottom-layer parameter of the initial energy consumption prediction model is fixed, and the energy consumption prediction model is acquired by fine-adjusting a high-layer parameter of the initial energy consumption prediction model based on the simulated energy consumption sample.
[0069] In the embodiments of the present disclosure, after acquiring the initial energy consumption prediction model, the computer device fixes the bottom-layer parameter of the initial energy consumption prediction model. Further, the computer device acquires the energy consumption prediction model of the first object by fine-adjusting the high-layer parameter of the initial energy consumption prediction model based on the simulated energy consumption sample. It should be noted that for different first objects, eneigy consumption prediction models acquired finally have different parameters.
[0070] In some embodiments, referring to FIG. 3, the energy consumption prediction model includes an encoding network 31 and a decoding network 32. The computer device acquires a parameter of the encoding network 31 based on the second history data of other objects, further fix the parameter of the encoding network 31, and acquires the energy consumption prediction model of the first object by fine-adjusting a parameter of the decoding network 32 by the simulated energy consumption sample of the first object. In some embodiments, the parameter of the encoding network 31 is a weight of the encoding network. The encoding network 31 and the decoding network 32 are connected by an encoder state diagram 33. In some embodiments, parameters 1, 2, and 3 are input to the encoding network 31 and parameters 1, 2 and 3 are output from the decoding network 32.
[0071] It should be noted that the training of the neural network is as follows: first a network weight of each layer is calculated forward, and then reverse calculation is performed. In some embodiments, referring to FIG. 4, zl and z3 of the first-layer encoding network are fixed; and during reverse calculation, weights of zl and z3 are not updated again, and all that is needed is to calculate from z5 to z3.
[0072] In summary, in the technical solutions according to the embodiments of the present disclosure, the simulated energy consumption sample is acquired by expansion of the real energy consumption sample, such that the technical problem that energy consumption prediction results of newly-built enterprises are inaccurate due to less history energy consumption data of these enterprises in related art is solved and the number of samples that can be used by the energy consumption prediction model is increased. Therefore, the accuracy of the energy consumption prediction result is improved. As the predicted energy consumption is acquired by the energy consumption predication model, high convenience and efficiency are achieved and the processing overhead for energy consumption prediction by the computer device is reduced.
[0073] In addition, the real energy consumption sample is acquired by standardizing and segmenting the first history data. Thus, it is ensured that the energy consumption amount of the real energy consumption sample is in the same standard, which is convenient for the generative adversarial network to use the first history data for training. Therefore, the reliability of the generative adversarial network is improved.
[0074] In addition, after the candidate simulated energy consumption sample is generated by the generator network of the generative adversarial network, the similarity between the simulated energy consumption sample and the real energy consumption sample is improved, thereby ensuring the reliability of the simulated energy consumption sample.
[0075] In some embodiments, as shown in FIG. 5, the method according to the embodiment of the present disclosure may further include the following steps.
[0076] In step 501, a second randomly-generated noise vector is acquired.
[0077] The second noise vector refers to a random vector for training the generative adversarial network. In some embodiments, the second noise vector may be a 16 -dimensional vector or a 32-dimensional vector generated by a distribution function. The second noise vector may also be a 16-dimensional vector or a 32-dimensional vector generated randomly.
[0078] In step 502, the generative adversarial network is constructed.
[0079] In the embodiments of the present disclosure, the computer device constructs the generative adversarial network. The generative adversarial network includes the generator network and the discriminator network.
[0080] In step 503, a trained generative adversarial network is acquired by training the generative adversarial network based on the second noise vector and the real energy consumption sample.
[0081] In some embodiments, the computer device trains the generative adversarial network based on the second noise vector and the real energy consumption sample.
[0082] In some embodiments, the training of the generative adversarial network includes the following several steps.
[0083] 1. A first loss function and a second loss function are acquired.
[0084] The first loss function is a loss function of the discriminator network and the second loss function is a loss function of the generator network. Exemplarily, assuming that the discriminator network is D(x), the first loss function Li is: [log(l - D(x))]
[0085]
Figure imgf000011_0001
, and
[0086] the second loss function L2 is:
L2 — min ma xEx Pr [logD (x) ] + Ex-Pg [log (1 — D(x))]
[0087] g D [0088] wherein pr represents the real energy consumption sample and pg represents the simulated energy consumption sample. It should be noted that the higher the distinction ability of the discriminator network is, the greater the values of the left and right sides of the first loss function are.
[0089] 2. A parameter of the generator network is fixed and the discriminator network is trained. [0090] In some embodiments, after constructing the generative adversarial network, the computer device fixes the parameter of the generator network and trains the discriminator network.
[0091] 3. A parameter of the discriminator network is fixed and the generator network is trained if the first loss function satisfies a first preset condition.
[0092] In some embodiments, when the first loss function corresponding to the discriminator network satisfies the first preset condition, the training of the discriminator network is completed and the computer device fixes the parameter of the trained discriminator network and trains the generator network. The first preset condition means that a loss value of the discriminator network acquired based on the first loss function is maximized.
[0093] 4. It is determined that the training of the generator network is completed if the second loss function satisfies a second preset condition.
[0094] In some embodiments, after acquiring the trained discriminator network, the computer device fixes the parameter of the discriminator network and trains the generator network. If the second loss function corresponding to the generator network satisfies the second preset condition, it is determined that the training of the generator network is completed. In some embodiments, the second preset condition means that a loss value of the second loss function is minimized. At this time, the second loss function is a convergent function.
[0095] If the second loss function does not satisfy the second preset condition, the computer device adjusts the parameter of the generator network, and again performs the procedure from the step that the discriminator network is trained.
[0096] In some embodiments, referring to FIG. 6, a method for training the generative adversarial network is introduced hereinafter. The generator network 61 generates the simulated energy consumption sample 63 based on the second noise vector 62. Further, the discriminator network 64 determines the authenticity of the simulated eneigy consumption sample 63 based on the simulated energy consumption sample 63 and the real energy consumption sample 65. For example, assuming that the simulated energy consumption sample 63 is marked as "false" and the real energy consumption sample 65 is marked as "true," if the discriminator network 64 determines that the simulated energy consumption sample 63 is "false," it is determined that the training of the discriminator network 64 is completed; and if the discriminator network 64 determines that the simulated energy consumption sample 63 is "true," the parameter of the discriminator network 64 is modified to continuously determine the authenticity of the simulated energy consumption sample 63 until the training of the discriminator network 64 is completed. After the training of the discriminator network 64 is completed, the generator network 61 is continued to be trained, the simulated energy consumption sample 63 is continuously generated based on the generator network 61, and the trained discriminator network 64 discriminates the simulated energy consumption sample. For example, assuming that the simulated energy consumption sample 63 is marked as "false" and the real energy consumption sample 65 is marked as "true," if the discriminator network 64 determines that the simulated energy consumption sample 63 is "true," it is determined that the training of the generator network 61 is completed; and if the discriminator network 64 determines that the simulated energy consumption sample 63 is "false," the parameter of the generator network 61 is modified to continuously generate the simulated energy consumption sample 63, and the discriminator network 64 discriminates the authenticity of the simulated eneigy consumption sample 63 until the training of the generator network 61 is completed. In addition, in the process of training the generator network and the discriminator network, the steps may be repeated until the loss value of the first loss function corresponding to the discriminator network 64 is maximized, and the second loss function corresponding to the generator network 61 is the convergent function.
[0097] In summary, in the technical solution according to the embodiment of the present disclosure, the generative adversarial network is trained based on the second noise vector and the real energy consumption sample, such that the reliability of the generator network used when the simulated energy consumption sample is generated is ensured, and the accuracy of the simulated energy consumption sample is improved.
[0098] In addition, the generative adversarial network is trained alternately by the first loss function and the second loss function. Thus, the accuracy of the simulated energy consumption sample acquired by the trained generative adversarial network is further ensured.
[0099] Hereinafter, apparatus and device embodiments of the present disclosure are described. The apparatus and device may be configured to perform the methods according to the method embodiments of the present disclosure. For the details not disclosed in the apparatus and device embodiments of the present disclosure, reference may be made to the method embodiments of the present disclosure.
[00100] Referring to FIG. 7, which shows a block diagram of an apparatus for predicting energy consumption according to an embodiment of the present disclosure, the apparatus has the function of practicing the above method examples. The function may be implemented by hardware or implemented by executing corresponding software by hardware. The apparatus may be a computer device or may be disposed in a computer device. The apparatus 700 may include a first data acquiring module 701, a real sample acquiring module 702, a simulated sample acquiring module 703 and a first energy consumption predicting module 704.
[00101] The first data acquisition module 701 is configured to acquire first history data, wherein the first history data refers to history energy consumption data of a first object.
[00102] The real sample acquiring module 702 is configured to acquire a real eneigy consumption sample of the first object based on the first history data.
[00103] The simulated sample acquiring module 703 is configured to acquire a simulated energy consumption sample of the first object based on the real energy consumption sample, wherein the simulated energy consumption sample includes virtual energy consumption data acquired by expansion based on the real energy consumption sample.
[00104] The first energy consumption predicting module 704 is configured to calculate predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model. [00105] In some embodiments, the real sample acquiring module 702 is configured to standardize the first history data, wherein the standardizing is to map the history energy consumption of the first object into a preset value range; acquire a prediction time period corresponding to the predicted eneigy consumption; and acquire the real energy consumption sample by segmenting the standardized first history data based on the prediction time period. [00106] In some embodiments, the simulated sample acquiring module 703 is configured to acquire a first randomly-generated noise vector; and generate the simulated energy consumption sample based on the first noise vector by a generator network in a generative adversarial network. [00107] In some embodiments, the generative adversarial network includes a generator network and a discriminator network. As shown in FIG. 8, the apparatus 700 further includes a candidate sample generating module 705, an authenticity calculating module 706, and a simulated sample determining module 707.
[00108] The candidate sample generating module 705 is configured to generate a candidate simulated energy consumption sample based on the first noise vector by the generator network. [00109] The authenticity calculating module 706 is configured to calculate authenticity of the candidate simulated energy consumption sample by the discriminator network based on the real energy consumption sample and the candidate simulated energy consumption sample.
[00110] The simulated sample determining module 707 is configured to determine the candidate simulated energy consumption sample as the simulated energy consumption sample if the authenticity is greater than a preset threshold.
[00111] In some embodiments, as shown in FIG. 8, the apparatus 700 further includes a second vector acquiring module 708, a neural network constructing module 709, and a neural network training module 710.
[00112] The second vector acquiring module 708 is configured to acquire a second randomly-generated noise vector.
[00113] The neural network constructing module 709 is configured to construct the generative adversarial network. [00114] The neural network training module 710 is configured to acquire a trained generative adversarial network by training the generative adversarial network based on the second noise vector and the real energy consumption sample.
[00115] In some embodiments, the neural network training module 710 is configured to: acquire a first loss function and a second loss function, wherein the first loss function is a loss function of the discriminator network and the second loss function is a loss function of the generator network; fix a parameter of the generator network and train the discriminator network; fix a parameter of the discriminator network and train the generator network if the first loss function satisfies a first preset condition; determine that the training of the generator network is completed if the second loss function satisfies a second preset condition; and adjust the parameter of the generator network if the second loss function does not satisfy the second preset condition, and perform the procedure again from the step of training the discriminator network.
[00116] In some embodiments, as shown in FIG. 8, the apparatus 700 further includes a second data acquiring module 711, a prediction model training module 712 and a model parameter fine-adjusting module 713. [00117] The second data acquiring module 711 is configured to acquire second history data, wherein the second history data refers to history energy consumption data of other objects associated with the first object.
[00118] The prediction model training module 712 is configured to acquire an initial energy consumption prediction model by training based on the second history data. [00119] The model parameter fine-adjusting module 713 is configured to fix a bottom-layer parameter of the initial energy consumption prediction model, and acquire the energy consumption prediction model by fine-adjusting a high-layer parameter of the initial energy consumption prediction model based on the first history data.
[00120] In summary, in the technical solutions according to the embodiments of the present disclosure, the simulated energy consumption sample is acquired by expansion of the real energy consumption sample, such that the technical problem that energy consumption prediction results of newly-built enterprises are inaccurate due to less history energy consumption data of these enterprises in related art is solved and the number of samples that can be used by the energy consumption prediction model is increased. Therefore, the accuracy of the energy consumption prediction result is improved.
[00121] It should be noted that the apparatus according to the above embodiment is only illustrated by the division of various functional modules when implementing its function. In practical application, the functions may be assigned to be completed by different functional modules according to actual needs. That is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and the method according to the above embodiments are based on the same inventive concept, and details of the apparatus are detailed in the method embodiment, which are not repeated herein.
[00122] Referring to FIG. 9, which shows a structural block diagram of a computer device 900 according to an embodiment of the present disclosure, the computer device is configured to perform the method for predicting energy consumption according to the above embodiments. The computer device may be any electronic device having data processing and storage capabilities. [00123] Specifically, the computer device 900 includes a processing unit, such as a central processing unit (CPU), a graphics processing unit (GPU), and a field-programmable gate array (FPGA) 901, a system memory 904 including a random -access memory (RAM) 902 and a read-only memory (ROM) 903, and a system bus 905 connecting the system memory 904 and the central processing unit 901. The computer device 900 further includes a basic input/output (I/O) system 906 configured to facilitate information transmission between various components within the computer device, and a high-capacity storage device 907 configured to store an operating system 913, an application 914 and otherprogram modules 915.
[00124] The basic input/output system 906 includes a display 908 configured to display information and an input device 909, such as a mouse and a keyboard, configured to input information of a user. Both the display 908 and the input device 909 are connected to the central processing unit 901 by an input/output controller 910 connected to the system bus 905. The basic input/output system 906 may also include the input/output controller 910 configured to receive and process inputs from a plurality of other devices, such as the keyboard, the mouse, or an electronic stylus. Similarly, the input/output controller 910 further provides output to the display, a printer or other types of output devices. [00125] The high -capacity storage device 907 is connected to the central processing unit 901 by a high -capacity storage controller (not shown) connected to the system bus 905. The high -capacity storage device 907 and a computer-readable medium associated therewith provide non-volatile storage for the computer device 900. That is, the high-capacity storage device 907 may include the computer-readable medium (not shown), such as a hard disk or a compact disc read-only memory (CD-ROM) driver.
[00126] Without loss of generality, the computer-readable medium may include a computer storage medium and a communication medium. The computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as a computer-readable instruction, a data structure, a program module or other data. The computer storage medium includes a RAM, a ROM, an erasable programmable read only memory (EPROM), an electrically erasable programmable read only memory (EEPROM), a flash memory or other solid-state storage technologies, a CD-ROM, a digital video disc (DVD) or other optical storage, a tape cartridge, a magnetic tape, a disk storage or other magnetic storage devices. It will be known by a person skilled in the art that the computer storage medium is not limited hereto. The system memory 904 and the high-capacity storage device 907 may be collectively referred to as the memory.
[00127] According to the embodiments of the present disclosure, the computer device 900 may also be connected to a remote computer on a network by the network, such as the Internet, for operation. That is, the computer device 900 may be connected to the network 912 by a network interface unit 911 connected to the system bus 905, or may be connected to other types of networks or remote computer systems (not shown) with the network interface unit 911.
[00128] The memory further includes at least one instruction, at least one program, a code set, or an instruction set. The at least one instruction, the at least one program, the code set, or the instruction set, when loaded and executed by one or more processors, causes the one or more processors to perform the above method.
[00129] An embodiment of the present disclosure provides a non-transitory computer-readable storage medium. The computer-readable storage medium stores at least one instruction, at least one program, a code set, or an instruction set therein. The at least one instruction, the at least one program, the code set, or the instruction set, when loaded and executed by a processor, causes the processor to perform the above method.
[00130] In some embodiments, the non-transitory computer-readable storage medium may include a ROM, a RAM, a solid state drive (SSD), an optical disc or the like. The random-access memory may include a resistance random-access memory (ReRAM) and a dynamic random-access memory (DRAM).
[00131] In some embodiments, a computer program product is further provided. When executed by a processor, the computer program product is configured to implement the above method. [00132] It should be understood that the term "a plurality of' herein refers to two or more. The term "and/or" describes an association relationship between associated objects and indicates that there may be three relationships. For example, A and/or B may indicate that there are three cases where A exists separately, A and B exist at the same time, and B exists separately. The character "/" generally indicates that an "or" relationship exists between contextual objects. In addition, step numbers described herein merely show exemplarily a possible execution sequence of steps. In some other embodiments, the steps may also be performed out of the number sequence. For example, two steps with different numbers may be performed at the same time or two steps with different numbers may be performed based on a sequence adverse to that shown in the figure, which is not limited in the embodiments of the present disclosure.
[00133] Described above are only exemplary embodiments of the present disclosure, and are not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements and the like made within the spirit and principles of the present disclosure should be included within the scope of protection of the present disclosure.

Claims

CLAIMS What is claimed is:
1. A method for predicting energy consumption, the method being applicable to a computer device, and comprising: acquiring first history data, wherein the first history data comprises history energy consumption data of a first object; acquiring a real energy consumption sample of the first object based on the first history data; acquiring a simulated energy consumption sample of the first object based on the real energy consumption sample, wherein the simulated energy consumption sample comprises an energy consumption sample acquired by expansion based on the real energy consumption sample; and calculating predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model.
2. The method according to claim 1, wherein acquiring the real energy consumption sample of the first object based on the first history data comprises: standardizing the first history data, wherein the standardizing is to map history energy consumption of the first object into a preset value range; acquiring a prediction time period corresponding to the predicted energy consumption; and acquiring the real energy consumption sample by segmenting the standardized first history data based on the prediction time period.
3. The method according to claim 1, wherein acquiring the simulated energy consumption sample of the first object based on the real eneigy consumption sample comprises: acquiring a first randomly-generated noise vector; and generating the simulated energy consumption sample based on the first noise vector by a generator network in a generative adversarial network.
4. The method according to claim 3, wherein the generative adversarial network comprises the generator network and a discriminator network; and before generating the simulated energy consumption sample based on the first noise vector by the generator network in the generative adversarial network, the method further comprises: generating a candidate simulated energy consumption sample based on the first noise vector by the generator network; calculating authenticity of the candidate simulated energy consumption sample by the discriminator network based on the real energy consumption sample and the candidate simulated energy consumption sample; and determining the candidate simulated energy consumption sample as the simulated energy consumption sample if the authenticity is greater than a preset threshold.
5. The method according to claim 3, wherein before generating the simulated energy consumption sample based on the first noise vector by the generator network in the generative adversarial network, the method further comprises: acquiring a second randomly -generated noise vector; constructing the generative adversarial network; and acquiring a trained generative adversarial network by training the generative adversarial network based on the second noise vector and the real energy consumption sample.
6. The method according to claim 5, wherein acquiring the trained generative adversarial network by training the generative adversarial network based on the second noise vector and the real energy consumption sample comprises: acquiring a first loss function and a second loss function, wherein the first loss function comprises a loss function of the discriminator network and the second loss function comprises a loss function of the generator network; fixing a parameter of the generator network and training the discriminator network; fixing a parameter of the discriminator and training the generator network if the first loss function satisfies a first preset condition; determining that the training of the generator network is completed if the second loss function satisfies a second preset condition; and adjusting the parameter of the generator network if the second loss function does not satisfy the second preset condition, and performing the procedure again from the step of training the discriminator network.
7. The method according to claim 1, wherein before calculating the predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling the energy consumption prediction model, the method further comprises: acquiring second history data, wherein the second history data refers to history energy consumption data of other objects associated with the first object; acquiring an initial energy consumption prediction model by training based on the second history data; and fixing a bottom-layer parameter of the initial energy consumption prediction model, and acquiring the energy consumption prediction model by fine-adjusting a high-layer parameter of the initial energy consumption prediction model based on the simulated energy consumption sample.
8. An apparatus for predicting energy consumption, the apparatus being applicable to a computer device, and comprising: a first data acquiring module, configured to acquire first history data, wherein the first history data comprises history energy consumption data of a first object; a real sample acquiring module, configured to acquire a real energy consumption sample of the first object based on the first history data; a simulated sample acquiring module, configured to acquire a simulated energy consumption sample of the first object based on the real energy consumption sample, wherein the simulated energy consumption sample comprises virtual energy consumption data acquired by expansion by a generative adversarial network based on the real energy consumption sample; and a first energy consumption predicting module, configured to calculate predicted energy consumption of the first object based on the real energy consumption sample and the simulated energy consumption sample by calling an energy consumption prediction model.
9. A computer device, comprising a processor and a memory storing a computer program, wherein the computer program, when loaded and run by the processor, causes the processor to perform the method as defined in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when loaded and run by a processor, causes the processor to perform the method as defined in any one of claims 1 to 7.
PCT/SG2020/050775 2019-12-25 2020-12-23 Method and apparatus for predicting energy consumption, and device and storage medium thereof Ceased WO2021133253A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911358181.4A CN111126700B (en) 2019-12-25 2019-12-25 Energy prediction methods, devices, equipment and storage media
CN201911358181.4 2019-12-25

Publications (1)

Publication Number Publication Date
WO2021133253A1 true WO2021133253A1 (en) 2021-07-01

Family

ID=70502229

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/SG2020/050775 Ceased WO2021133253A1 (en) 2019-12-25 2020-12-23 Method and apparatus for predicting energy consumption, and device and storage medium thereof

Country Status (2)

Country Link
CN (1) CN111126700B (en)
WO (1) WO2021133253A1 (en)

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113420510A (en) * 2021-07-07 2021-09-21 广东电网有限责任公司 Energy consumption distinguishing method based on front-end sensing and learning
CN113705908A (en) * 2021-08-31 2021-11-26 江南大学 Auxiliary prediction method and system for energy consumption in organic silicon monomer fractionation process
CN113837488A (en) * 2021-10-14 2021-12-24 南京国铁电气有限责任公司 Energy consumption data prediction method, system and equipment
CN113850412A (en) * 2021-08-18 2021-12-28 华建数创(上海)科技有限公司 Prediction method of regional energy consumption in buildings based on bim model and graph convolutional neural network
CN114462719A (en) * 2022-02-25 2022-05-10 北京科技大学 Fuzzy confrontation generation network wind power prediction method and device based on evolutionary framework
CN114781716A (en) * 2022-04-18 2022-07-22 上海市嘉定自来水有限公司 Training method of water model, secondary water supply control method, system and equipment
CN115604131A (en) * 2022-12-15 2023-01-13 广州丰石科技有限公司（Cn） Link flow prediction method, system, electronic device and medium
CN115658885A (en) * 2022-08-25 2023-01-31 粤港澳大湾区数字经济研究院(福田) Text intelligent labeling method, system, intelligent terminal and storage medium
CN115730502A (en) * 2021-08-26 2023-03-03 中国移动通信集团重庆有限公司 Training method, device and equipment for battery discharge duration determination model
CN115933380A (en) * 2022-11-16 2023-04-07 中国科学院自动化研究所 Training method and related equipment for strategy prediction model of power generation process
CN116754828A (en) * 2023-08-21 2023-09-15 济南瑞源智能城市开发有限公司 Intelligent tunnel energy consumption monitoring method, device and medium
CN116780536A (en) * 2023-08-23 2023-09-19 山东未来互联科技有限公司 Energy consumption prediction method and system based on machine learning
CN117077537A (en) * 2023-08-31 2023-11-17 江苏徐工工程机械研究院有限公司 Energy consumption determination, model training methods and devices for construction machinery and equipment
CN117436848A (en) * 2023-12-21 2024-01-23 山东康吉诺技术有限公司 Motor energy consumption monitoring method based on artificial intelligence
CN118660307A (en) * 2024-07-04 2024-09-17 国网山西省电力公司经济技术研究院 5G base station power consumption prediction method and system based on conditional generative adversarial network
EP4718310A1 (en) 2024-09-30 2026-04-01 MEF MyEnergy.Farm GmbH Method for determining building-specific energy concepts for heated buildings

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112801123A (en) * 2020-09-09 2021-05-14 华北电力大学 Small sample user electricity consumption data expansion method with frequency domain distribution consistency
CN112364999B (en) * 2020-10-19 2021-11-19 深圳市超算科技开发有限公司 Training method and device for water chiller adjustment model and electronic equipment
CN112420142A (en) * 2020-11-20 2021-02-26 上海梅斯医药科技有限公司 Data acquisition method and device, computer equipment and storage medium
CN113962425A (en) * 2021-08-02 2022-01-21 北京百度网讯科技有限公司 Heating data generation method, device, equipment and computer storage medium
CN116028708A (en) * 2022-10-12 2023-04-28 支付宝(杭州)信息技术有限公司 Recommendation model training method and device
CN116245542B (en) * 2023-03-13 2026-04-14 中国建设银行股份有限公司 Method, apparatus, device and computer readable medium for processing data

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10296988B2 (en) * 2013-08-19 2019-05-21 Board Of Trustees Of Michigan State University Linear optimal power flow system and method
KR101761686B1 (en) * 2017-03-31 2017-07-31 (주)하모니앤유나이티드 Real time predicting system for energy management system using machine learning
CN110378358A (en) * 2018-04-14 2019-10-25 中国电力科学研究院有限公司 A kind of power distribution network isomeric data integration method and system
CN109117863B (en) * 2018-07-04 2021-08-13 全球能源互联网研究院有限公司 Insulator sample expansion method and device based on deep convolutional generative adversarial network
CN109116100B (en) * 2018-07-09 2019-09-27 清华大学 It is a kind of based on coding-decoding structure electric load electricity consumption decomposition method
CN109446735B (en) * 2018-12-18 2020-12-04 中国石油大学(北京) A method, device and system for generating simulated logging data
CN109993359B (en) * 2019-03-26 2022-11-18 华南理工大学 Situation prediction method based on advanced machine learning
CN110266002A (en) * 2019-06-20 2019-09-20 北京百度网讯科技有限公司 Method and apparatus for forecasting electrical load
CN110390441A (en) * 2019-07-30 2019-10-29 北京百度网讯科技有限公司 With energy prediction technique and device
CN110490659B (en) * 2019-08-21 2022-01-11 广东电网有限责任公司 GAN-based user load curve generation method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
GU YUXUAN; CHEN QIXIN; LIU KAI; XIE LE; KANG CHONGQING: "GAN-based Model for Residential Load Generation Considering Typical Consumption Patterns", 2019 IEEE POWER & ENERGY SOCIETY INNOVATIVE SMART GRID TECHNOLOGIES CONFERENCE (ISGT), 18 February 2019 (2019-02-18), pages 1 - 5, XP033593248, DOI: 10.1109/ISGT.2019.8791575 *
WANG FEI, ZHANG ZHANYAO, LIU CHUN, YU YILI, PANG SONGLING, DUIĆ NEVEN, SHAFIE-KHAH MIADREZA, CATALÃO JOÃO P.S.: "Generative adversarial networks and convolutional neural networks based weather classification model for day ahead short-term photovoltaic power forecasting", ENERGY CONVERSION AND MANAGEMENT, vol. 181, 1 February 2019 (2019-02-01), pages 443 - 462, XP055834937, ISSN: 0196-8904, DOI: 10.1016/j.enconman.2018.11.074 *
ZHANG CHI; KUPPANNAGARI SANMUKH R.; KANNAN RAJGOPAL; PRASANNA VIKTOR K.: "Generative Adversarial Network for Synthetic Time Series Data Generation in Smart Grids", 2018 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS, CONTROL, AND COMPUTING TECHNOLOGIES FOR SMART GRIDS (SMARTGRIDCOMM), 29 October 2018 (2018-10-29), pages 1 - 6, XP033478150, DOI: 10.1109/SMARTGRIDCOMM.2018.8587464 *

Cited By (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113420510B (en) * 2021-07-07 2022-06-17 广东电网有限责任公司 Energy consumption distinguishing method based on front-end sensing and learning
CN113420510A (en) * 2021-07-07 2021-09-21 广东电网有限责任公司 Energy consumption distinguishing method based on front-end sensing and learning
CN113850412A (en) * 2021-08-18 2021-12-28 华建数创(上海)科技有限公司 Prediction method of regional energy consumption in buildings based on bim model and graph convolutional neural network
CN115730502A (en) * 2021-08-26 2023-03-03 中国移动通信集团重庆有限公司 Training method, device and equipment for battery discharge duration determination model
CN113705908A (en) * 2021-08-31 2021-11-26 江南大学 Auxiliary prediction method and system for energy consumption in organic silicon monomer fractionation process
CN113837488B (en) * 2021-10-14 2023-12-26 南京国铁电气有限责任公司 Method, system and equipment for predicting energy consumption data
CN113837488A (en) * 2021-10-14 2021-12-24 南京国铁电气有限责任公司 Energy consumption data prediction method, system and equipment
CN114462719A (en) * 2022-02-25 2022-05-10 北京科技大学 Fuzzy confrontation generation network wind power prediction method and device based on evolutionary framework
CN114462719B (en) * 2022-02-25 2026-02-24 北京科技大学 Fuzzy countermeasure generation network wind power prediction method and device based on evolution framework
CN114781716A (en) * 2022-04-18 2022-07-22 上海市嘉定自来水有限公司 Training method of water model, secondary water supply control method, system and equipment
CN115658885A (en) * 2022-08-25 2023-01-31 粤港澳大湾区数字经济研究院(福田) Text intelligent labeling method, system, intelligent terminal and storage medium
CN115933380A (en) * 2022-11-16 2023-04-07 中国科学院自动化研究所 Training method and related equipment for strategy prediction model of power generation process
CN115604131A (en) * 2022-12-15 2023-01-13 广州丰石科技有限公司（Cn） Link flow prediction method, system, electronic device and medium
CN116754828B (en) * 2023-08-21 2023-12-01 济南瑞源智能城市开发有限公司 Intelligent tunnel energy consumption monitoring method, device and medium
CN116754828A (en) * 2023-08-21 2023-09-15 济南瑞源智能城市开发有限公司 Intelligent tunnel energy consumption monitoring method, device and medium
CN116780536A (en) * 2023-08-23 2023-09-19 山东未来互联科技有限公司 Energy consumption prediction method and system based on machine learning
CN116780536B (en) * 2023-08-23 2024-02-02 山东未来互联科技有限公司 Energy consumption prediction method and system based on machine learning
CN117077537A (en) * 2023-08-31 2023-11-17 江苏徐工工程机械研究院有限公司 Energy consumption determination, model training methods and devices for construction machinery and equipment
CN117436848A (en) * 2023-12-21 2024-01-23 山东康吉诺技术有限公司 Motor energy consumption monitoring method based on artificial intelligence
CN117436848B (en) * 2023-12-21 2024-03-08 山东康吉诺技术有限公司 Motor energy consumption monitoring method based on artificial intelligence
CN118660307A (en) * 2024-07-04 2024-09-17 国网山西省电力公司经济技术研究院 5G base station power consumption prediction method and system based on conditional generative adversarial network
EP4718310A1 (en) 2024-09-30 2026-04-01 MEF MyEnergy.Farm GmbH Method for determining building-specific energy concepts for heated buildings
DE102024128221A1 (en) 2024-09-30 2026-04-02 MEF MyEnergy.Farm GmbH Methods for AI-supported determination of site-specific, holistic energy concepts for heated buildings

Also Published As

Publication number Publication date
CN111126700B (en) 2023-09-15
CN111126700A (en) 2020-05-08

Similar Documents

Publication Publication Date Title
CN111126700B (en) Energy prediction methods, devices, equipment and storage media
CN111709533B (en) Distributed training method, device and computer equipment for machine learning model
KR102170105B1 (en) Method and apparatus for generating neural network structure, electronic device, storage medium
KR102831538B1 (en) Complex model scaling for neural networks
US11429434B2 (en) Elastic execution of machine learning workloads using application based profiling
US20230049747A1 (en) Training machine learning models using teacher annealing
US20220172038A1 (en) Automated deep learning architecture selection for time series prediction with user interaction
US12147886B2 (en) Predictive microservices activation using machine learning
US20210097407A1 (en) Predicting operational status of system
CN117519996B (en) Data processing method, device, equipment and storage medium
US12367350B2 (en) Random action replay for reinforcement learning
CN119249224A (en) Power grid protocol identification method, device, computer equipment, readable storage medium and program product based on knowledge graph
CN117318052A (en) Reactive power prediction method, device and computer equipment for generator set phase advance test
US20210357794A1 (en) Determining the best data imputation algorithms
CN112905987B (en) Account identification method, device, server and storage medium
US11475297B2 (en) Cross-domain homophily quantification for transfer learning
CN119518958A (en) A method and system for generating new energy power generation scenarios considering uncertainty
CN119179002A (en) Model training method, device, equipment and medium for battery health state estimation
CN117786211A (en) Travel scheme recommendation method and device
CN116701587A (en) Question answering method and device based on machine learning
CN117009028A (en) Scheduling method, device, electronic equipment and storage medium for container group
CN109285559B (en) Role transition point detection method and device, storage medium and electronic equipment
CN120909532B (en) Storage system optimization method and device, electronic equipment, medium and product
CN120564761B (en) Transformer fault voiceprint detection method based on long-term and short-term memory network optimization
CN118827413B (en) Traffic prediction methods and devices, electronic devices and storage media for IP bearer networks

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: 20908368

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: 20908368

Country of ref document: EP

Kind code of ref document: A1