CN111091217A - Building short-term load prediction method and system - Google Patents

Building short-term load prediction method and system Download PDF

Info

Publication number
CN111091217A
CN111091217A CN201811233985.7A CN201811233985A CN111091217A CN 111091217 A CN111091217 A CN 111091217A CN 201811233985 A CN201811233985 A CN 201811233985A CN 111091217 A CN111091217 A CN 111091217A
Authority
CN
China
Prior art keywords
data
building
quasi
load
correlation
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.)
Granted
Application number
CN201811233985.7A
Other languages
Chinese (zh)
Other versions
CN111091217B (en
Inventor
田世明
田英杰
卜凡鹏
苏运
李德智
龚桃荣
宫飞翔
韩凝辉
石坤
董明宇
潘明明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
State Grid Shanghai Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
State Grid Shanghai Electric Power Co 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 State Grid Corp of China SGCC, China Electric Power Research Institute Co Ltd CEPRI, State Grid Shanghai Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN201811233985.7A priority Critical patent/CN111091217B/en
Publication of CN111091217A publication Critical patent/CN111091217A/en
Application granted granted Critical
Publication of CN111091217B publication Critical patent/CN111091217B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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
    • 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
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

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

Abstract

The invention provides a building short-term load prediction method and a system, comprising the following steps: acquiring mobile base station data, meteorological data and building load data with the same time, and extracting quasi-characteristic variables of the mobile base station data; obtaining an input variable of load prediction through an entropy weight method based on the quasi-characteristic variable, the meteorological data and the building load data; and predicting the building load through a BP neural network according to the input variable. The building short-term load prediction method provided by the invention improves the prediction precision of building short-term load prediction. The method is beneficial to further understanding of the power utilization behaviors of users and exploring the development rule of the power system, and has important guiding significance for power load prediction, distribution network load early warning and safe and economic operation of the smart power grid.

Description

Building short-term load prediction method and system
Technical Field
The invention relates to the technical field of power engineering, in particular to a building short-term load prediction method and system.
Background
Along with the rapid development of social economy and the rapid progress of internet information technology, the management of electricity demand side faces new opportunities and challenges: various international conferences, large-scale events and exhibitions are called in China, such as APEC, Olympic games, Ideal races, Expo and Disney, and the large-scale venues are quickly built and operated in a short time, meanwhile, various large-scale commercial buildings and public activity places are built in succession and traffic hubs are modified to form comprehensive commercial places, the electricity consumption is huge and rapidly increased, and great impact is brought to the safe, stable and economic operation of local power grids.
Meanwhile, the prior art preliminarily realizes the integration, sharing and utilization of enterprise-level data resources. However, the technical capability of effective processing and analysis utilization of power data under the condition of big data is still not mature enough at present, and especially in the face of ubiquitous and diverse big data resources, how to effectively evaluate the quality of data power big data and external economic and social big data, how to efficiently utilize and discover various power utilization data and external economic and social big data resources to serve power distribution and utilization, and how to improve the operation quality and the benefit of a power grid company are also urgently needed to be researched and explored.
Disclosure of Invention
The invention provides a building short-term load prediction method, aiming at solving the problem that quantitative indexes affecting human activities are absent in the existing load prediction model in the prior art.
The technical scheme provided by the invention is as follows: a building short-term load prediction method comprises
Performing relevancy analysis on the acquired mobile base station data, building load data and preset quasi-characteristic variables at the same time to obtain quasi-characteristic input variables;
performing correlation analysis on the acquired meteorological data and building load data at the same time to obtain meteorological input variables;
and substituting the quasi-characteristic input variable and the meteorological input variable into a pre-trained BP neural network to predict the building load.
Preferably, the obtaining of the quasi-feature input variable by analyzing the association degree based on the obtained mobile base station data, the building load data and the preset quasi-feature variable at the same time includes:
obtaining a longitudinal correlation degree between the quasi-characteristic variable and the building load data and a transverse correlation degree between the quasi-characteristic variables through an entropy weight method;
and carrying out dimensionality reduction screening on the transverse correlation degree and the longitudinal correlation degree to obtain a quasi-characteristic variable input variable.
Preferably, the obtaining of the longitudinal correlation degree between the quasi-characteristic variable and the building load data and the transverse correlation degree between the quasi-characteristic variables by the entropy weight method includes:
obtaining an initial matrix through a grey correlation degree calculation formula based on quasi-characteristic variables and the building data load;
calculating the longitudinal correlation degree of the quasi-characteristic variable to the dependent variable based on the initial matrix to obtain a longitudinal correlation degree matrix, so as to obtain the longitudinal correlation degree of the quasi-characteristic variable to the building data load;
and calculating the transverse correlation degree between the quasi-characteristic variables by utilizing a gray correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix.
Preferably, the initial matrix is represented by the following formula:
Figure BDA0001837786380000021
in the formula, XiIs the average data per day of the ith sequence of pseudocharacteristic variables, i ═ 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
Preferably, the longitudinal correlation matrix is represented by the following formula:
Figure BDA0001837786380000022
wherein, X0y1Representing the vertical relevance of the ith pseudo-characteristic variable to j building data loads, wherein i is 0 and 1 … m; j is 1,2 … n.
Preferably, the transverse correlation matrix is represented by the following formula:
Figure BDA0001837786380000023
wherein, XixjRepresenting the transverse relevance of the ith quasi-characteristic variable to the jth quasi-characteristic variable, wherein i is 0 and 1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
Preferably, the obtaining of the longitudinal correlation degree between the quasi-characteristic variable and the building load data and the transverse correlation degree between the quasi-characteristic variables by the entropy weight method includes:
dividing quasi-feature variables corresponding to the transverse association degrees which are greater than a preset threshold into a first variable group, comparing longitudinal association degrees corresponding to the quasi-feature variables in the first variable group, and selecting a first quasi-feature variable with the maximum longitudinal association degree;
dividing quasi-characteristic variables corresponding to the transverse association degrees smaller than a preset threshold into second variable groups, comparing longitudinal association degrees corresponding to the quasi-characteristic variables in the second variable groups, and selecting the second quasi-characteristic variable with the maximum longitudinal association degree;
the first quasi characteristic variable and the second quasi characteristic variable are prediction input variables of mobile base station data.
Preferably, the correlation analysis based on the acquired meteorological data and building load data at the same time is used for obtaining meteorological input variables, including
Obtaining the longitudinal correlation degree between the meteorological data and the building load data and the transverse correlation degree between the meteorological data through an entropy weight method;
and obtaining meteorological data input variables by carrying out dimensionality reduction screening on the transverse relevance degree and the longitudinal relevance degree.
Preferably, the obtaining of the longitudinal correlation degree between the meteorological data and the building load data and the lateral correlation degree between the meteorological data by the entropy weight method includes:
obtaining a meteorological data initial matrix through a grey correlation degree calculation formula based on meteorological data and the building data load;
calculating the longitudinal correlation degree of the meteorological data to the building data load based on the meteorological data initial matrix to obtain a longitudinal correlation degree matrix;
and calculating the transverse correlation degree between the meteorological data by using a grey correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix.
Preferably, the meteorological data initial matrix is as follows:
Figure BDA0001837786380000031
in the formula, XiThe average data per day of the ith meteorological data sequence, i is 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
Preferably, the longitudinal correlation matrix is represented by the following formula:
Figure BDA0001837786380000032
wherein, X0y1Representing the vertical relevance of the ith meteorological data to the j types of building data loads, wherein i is 0 and 1 … m; j is 1,2 … n.
Preferably, the transverse correlation matrix is represented by the following formula:
Figure BDA0001837786380000041
wherein, XixjIndicating the horizontal relevance of the ith meteorological data to the jth meteorological data, wherein i is 0 and 1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
Preferably, the obtaining of the meteorological data input variables by subjecting the horizontal correlation degree and the vertical correlation degree to dimension reduction screening includes:
dividing quasi-characteristic variables corresponding to the transverse correlation degrees which are greater than a preset threshold value into a first data group, comparing longitudinal correlation degrees corresponding to meteorological data in the first data group, and selecting first meteorological data with the maximum longitudinal correlation degree;
dividing meteorological data corresponding to the transverse correlation degree smaller than a preset threshold into a second data group, comparing longitudinal correlation degrees corresponding to the meteorological data in the second data group, and selecting second meteorological data with the maximum longitudinal correlation degree;
the first meteorological data and the second meteorological data are predicted input variables of meteorological data.
Preferably, the bringing the quasi-feature input variables and the meteorological input variables into a pre-trained BP neural network to predict the building load includes:
selecting mobile base station data, meteorological data and total load data which are subjected to data preprocessing and matched with the data as an initial data set;
and dividing the initial data set into a training set and a testing set according to the ratio of 9:1, and respectively training the neural network and testing the prediction effect to obtain the trained BP neural network.
Preferably, the mobile base station data includes: time labels, number of floors reached, number of floors changed, building name and number of people changed;
the meteorological data, comprising: temperature, dew point temperature, relative humidity, wind speed, and barometric pressure;
the building load data comprises: total load, energy for lighting and sockets, energy for air conditioning, energy for power, and other energy uses.
The quasi-feature variable comprises: the number of people going upstairs, downstairs, immobilizers, variators, upstairs and immobilizers, downstairs and immobilizers, and the total number of people.
A building short term load prediction system comprising:
a data acquisition module: acquiring mobile base station data, meteorological data and building load data with the same time, and extracting quasi-characteristic variables of the mobile base station data;
an input variable acquisition module: obtaining an input variable of load prediction through an entropy weight method based on the quasi-characteristic variable, the meteorological data and the building load data;
a prediction module: and predicting the building load through a BP neural network according to the input variable.
Preferably, the input variable acquiring module includes:
a quasi-characteristic variable input quantity acquisition submodule: based on the quasi-characteristic variable, performing longitudinal correlation degree calculation and transverse correlation degree calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a mobile base station data input variable;
a meteorological data input variable acquisition submodule: and based on the meteorological data, performing longitudinal correlation calculation and transverse correlation calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a meteorological data input variable.
Preferably, the quasi-feature variable input quantity obtaining sub-module includes:
a vertical relevance calculating unit: calculating the longitudinal correlation degree of the quasi-characteristic variable to the building data load through the grey correlation analysis method;
a lateral correlation degree calculation unit: calculating the transverse correlation degree among the quasi-characteristic variables by the gray correlation analysis method;
a mobile base station data input amount acquisition unit: and determining a prediction input variable of the mobile base station data based on the transverse correlation degree, a preset threshold value and the longitudinal correlation degree.
Compared with the prior art, the invention has the beneficial effects that:
acquiring mobile base station data, meteorological data and building load data with the same time, and extracting quasi-characteristic variables of the mobile base station data; obtaining an input variable of load prediction through an entropy weight method based on the quasi-characteristic variable, the meteorological data and the building load data; and predicting the building load through a BP neural network according to the input variable. The data of the mobile base station is used as a quantitative index of the influence of human activities on the building load, so that the understanding of the influence mechanism of the human activities on the load is deepened. The mobile base station data characteristic variables are used as input variables and added into building short-term load prediction, the prediction precision of the building short-term load prediction is improved, the power utilization behaviors of users are further understood, the development rule of a power system is explored, and the method has important guiding significance for power load prediction, distribution network load early warning and safe and economic operation of a smart power grid.
Drawings
FIG. 1 is a flow chart of a building short term load prediction method of the present invention;
FIG. 2 is a flow chart of a gray correlation analysis with entropy weight;
FIG. 3 is a flowchart of a predictor variable selection method based on cross-vertical gray correlation calculation according to the present invention;
FIG. 4 is a graph of a prediction without consideration of a mobile station according to an embodiment of the present invention;
fig. 5 is a graph of building short term load prediction in accordance with an embodiment of the present invention.
Detailed Description
For a better understanding of the present invention, reference is made to the following description taken in conjunction with the accompanying drawings and examples.
Example 1:
fig. 1 is a flowchart of a building short-term load prediction method according to the present invention, and as shown in the drawing, the building short-term load prediction method according to the present invention includes:
s1: performing relevance analysis based on the obtained mobile base station data, building load data and preset quasi-characteristic variable at the same time to obtain quasi-characteristic input variable
The method comprises the steps of obtaining mobile base station data and building load data, preprocessing the data, matching the data, and obtaining the mobile base station data and the building load data with the same time tag;
the changing floor number in the mobile base station data records the floor change condition of the mobile crowd in the current time period, the changing floor number is positive when the mobile crowd goes upstairs, the specific numerical value represents the changing floor number, and the quasi-characteristic variable of the mobile base station data is extracted according to the positive and negative of the changing floor number as a judgment basis;
the mobile base station data includes: time labels, number of floors reached, number of floors changed, building name, number of people changed; the meteorological data includes: temperature, dew point temperature, relative humidity, wind speed, barometric pressure; the building load data includes: total load, energy for lighting and sockets, energy for air conditioning, energy for power, and energy for other uses.
And extracting the quasi-characteristic variable of the mobile base station data by taking the positive and negative of the number of the varying layers in the mobile base station data as a judgment basis: when the number of the changed floors is more than 0, the corresponding number of people is the number of people going upstairs, when the number of the changed floors is equal to 0, the number of people is regarded as the number of people not moving, when the number of the changed floors is less than 0, the number of people going downstairs is regarded as the number of people going downstairs, the number of people going upstairs and the number of people going downstairs are summed to obtain the number of people going upstairs and the number of people not moving upstairs, the number of people going downstairs, the number of people not moving are summed to obtain the total number, and then 7 pseudo-characteristic variables of the moving base station data are obtained.
S2: performing correlation analysis on the acquired meteorological data and building load data at the same time to obtain meteorological input variables;
acquiring meteorological data and building load data, preprocessing the data, matching the data, and acquiring the meteorological data and the building load data with the same time tag;
s3: and substituting the quasi-characteristic input variable and the meteorological input variable into a pre-trained BP neural network to predict the building load.
Taking the degree of association of the independent variables to the dependent variables as longitudinal degree of association, considering the degree of influence of the independent variables to the dependent variables, taking the degree of association of one of the independent variables to other independent variables as transverse degree of association, considering the similarity between the respective variables, and analyzing the data hierarchy of the mobile base station and the meteorological data hierarchy and selecting the predictive variables by using a predictive variable selection method calculated based on the transverse and longitudinal grey degree of association;
and (4) predicting the total load of the building by using a BP neural network in combination with the input variables in the steps.
The predictive variable selection method based on horizontal and vertical grey correlation calculation analyzes and selects predictive variables of a mobile base station data hierarchy and a meteorological data hierarchy, and comprises the following specific steps:
(1) respectively calculating the longitudinal relevance of each quasi-characteristic variable of the mobile base station data and the meteorological data to each subentry load, wherein the subentry load with the highest longitudinal relevance for the same mobile base station data is a main way for influencing the total load by the mobile base station data, and the meteorological data are similar;
(2) and respectively calculating the transverse association degree among the quasi-characteristic variables of the mobile base station data, wherein the transverse association degree larger than 0.9 can be regarded as a similar factor, comparing the longitudinal association degree of the similar factor to the total load, keeping the highest longitudinal association degree as an input variable of load prediction to achieve the purpose of reducing the dimension of the input variable, then calculating the transverse association degree among the meteorological data variables, and the screening principle is the same as that of the mobile base station data.
The method for predicting the total load of the building by using the BP neural network comprises the following specific steps:
(1) reading historical load data and input variable data, and performing data normalization processing;
(2) training the BP neural network by using the data after the normalization processing;
(3) and inputting the input variable of the day to be predicted into the trained BP neural network for prediction.
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 2, the process diagram of gray correlation analysis based on entropy weight method specifically includes the following steps:
(1) inputting a numerical data set and generating an initial matrix.
Figure BDA0001837786380000081
Wherein, XiColumn is the average data per day of the ith sequence of influencing factors, Y0Columns are the daily average data for the sequence of target factors.
(2) Normalization processing, namely performing normalization processing on data by adopting a linear normalization function, wherein the linear normalization function is as follows:
Figure BDA0001837786380000082
(3) generating a difference matrix, calculating the difference to obtain the difference matrix, and extracting the maximum difference M and the minimum difference M, wherein the correlation formula is as follows:
Figure BDA0001837786380000083
M＝max{Δij}
m＝min{Δij}
wherein, Deltaij＝|xi(j)-y0(j)|。
(4) Generating a correlation coefficient matrix, wherein the correlation coefficient matrix is as follows:
Figure BDA0001837786380000084
wherein the content of the first and second substances,
Figure BDA0001837786380000085
usually, ρ is 0.5.
(5) Calculating entropy and weight coefficient of each index, and specifically comprises the following steps:
(a) calculating the characteristic proportion of the jth influence factor under the ith transaction, wherein the characteristic proportion calculation formula is as follows:
Figure BDA0001837786380000086
(b) and calculating the entropy value of the j index, wherein the entropy value calculation formula is as follows:
Figure BDA0001837786380000087
in the formula, m is the number of transactions. When P is presentijWhen equal to 0, take PijlnPij＝0。
(c) Calculating a weight coefficient of the j index, wherein the weight coefficient calculation formula is as follows:
Figure BDA0001837786380000088
(6) and calculating the weighted gray correlation degree, and summing the elements of each column of the correlation coefficient matrix according to the weighted weight coefficient to obtain the weighted gray correlation degree.
As shown in fig. 3, a flowchart of a predictor variable selection method based on horizontal and vertical gray correlation calculation specifically includes the following steps:
(1) and inputting the quasi-feature variable and the dependent variable to form an initial matrix.
Figure BDA0001837786380000091
(2) Calculating quasi-characteristic variable X by utilizing grey correlation analysis algorithm based on entropy weight methodiObtaining a longitudinal correlation matrix for the longitudinal correlation of the dependent variable Y, wherein XiyjAnd the correlation degree of the ith independent variable to the jth dependent variable is shown.
Figure BDA0001837786380000092
(3) Calculating quasi-characteristic variable X by utilizing grey correlation analysis algorithm based on entropy weight methodiThe transverse correlation degree between the two groups of the data blocks to obtain a transverse correlation degree matrix, wherein XixjIndicates the degree of association of the ith argument with the jth argument, and if i equals j, X isixj＝0。
Figure BDA0001837786380000093
(4) And dividing the quasi-characteristic variables with the transverse association degree larger than 0.9 into a group, comparing the longitudinal association degrees of all the variables in the group to the dependent variable, and selecting the quasi-characteristic variable with the maximum longitudinal association degree.
And (4) outputting the quasi-feature variable with the transverse correlation degree less than or equal to 0.9, taking the quasi-feature variable with the maximum longitudinal correlation degree in the step (4) as the feature variable of the independent variable data set, and outputting the corresponding longitudinal correlation degree.
Example 2:
based on the same inventive concept, the invention also provides a building short-term load prediction system, which comprises:
a data acquisition module: acquiring mobile base station data, meteorological data and building load data with the same time tag, and extracting quasi-characteristic variables of the mobile base station data;
an input variable acquisition module: obtaining an input variable of load prediction through an entropy weight method based on the quasi-characteristic variable, the meteorological data and the building load data;
a prediction module: and predicting the building load through a BP neural network according to the input variable.
The input variable acquisition module comprises:
a quasi-characteristic variable input quantity acquisition submodule: based on the quasi-characteristic variable, performing longitudinal correlation degree calculation and transverse correlation degree calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a mobile base station data input variable;
a meteorological data input variable acquisition submodule: and based on the meteorological data, performing longitudinal correlation calculation and transverse correlation calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a meteorological data input variable.
The quasi-characteristic variable input quantity acquisition submodule comprises:
a vertical relevance calculating unit: calculating the longitudinal correlation degree of the quasi-characteristic variable to the building data load through the grey correlation analysis method;
a lateral correlation degree calculation unit: calculating the transverse correlation degree among the quasi-characteristic variables by the gray correlation analysis method;
a mobile base station data input amount acquisition unit: and determining a prediction input variable of the mobile base station data based on the transverse correlation degree, a preset threshold value and the longitudinal correlation degree.
The longitudinal correlation degree calculation unit obtains the longitudinal correlation degree of the quasi-characteristic variable to the building data load through the following formula:
forming an initial matrix through a grey correlation degree calculation formula based on quasi-characteristic variables and the building data load;
the initial matrix is shown as follows:
Figure BDA0001837786380000101
in the formula, XnN is 1,2 … i; y is0Is the building data load;
calculating the longitudinal association degree of the quasi-characteristic variable to the dependent variable by utilizing a grey association analysis algorithm based on an entropy weight method to obtain a longitudinal association degree matrix, so as to obtain the longitudinal association degree of the quasi-characteristic variable to the building data load;
the longitudinal correlation matrix is shown as follows:
Figure BDA0001837786380000102
in the formula, XiIs the average data per day of the ith sequence of pseudocharacteristic variables, i ═ 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
The transverse relevance calculating unit calculates the transverse relevance among the quasi-feature variables according to the following formula:
calculating the transverse correlation degree between quasi-characteristic variables by utilizing a gray correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix;
the transverse correlation matrix is shown as follows:
Figure BDA0001837786380000111
wherein, XixjRepresenting the transverse relevance of the ith quasi-characteristic variable to the jth quasi-characteristic variable, wherein i is 0 and 1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
The mobile base station data input amount acquisition unit comprises:
a first quasi-characteristic variable acquisition subunit: dividing all the quasi-characteristic variables with the transverse association degrees larger than a preset threshold into a first variable group, comparing the longitudinal association degrees of the quasi-characteristic variables in the first variable group to the building data load, and selecting a first quasi-characteristic variable with the maximum longitudinal association degree;
a second quasi-characteristic variable obtaining subunit: dividing all the quasi-characteristic variables with the transverse association degrees smaller than a preset threshold into a second variable group, comparing the longitudinal association degrees of the building data loads of the quasi-characteristic variables in the second variable group, and selecting the second quasi-characteristic variable with the maximum longitudinal association degree;
the mobile base station input variable acquisition subunit: the first quasi characteristic variable and the second quasi characteristic variable are prediction input variables of mobile base station data.
The meteorological data input variable acquisition submodule based on the meteorological data comprises:
a vertical relevance calculating unit: the grey correlation analysis method is used for calculating the longitudinal correlation degree of the meteorological data to the building data load;
a lateral correlation degree calculation unit: calculating the transverse correlation degree between the meteorological data according to the grey correlation analysis method;
meteorological data input quantity acquisition unit: and determining a prediction input variable of the meteorological data based on the transverse correlation degree, a preset threshold value and the longitudinal correlation degree.
The longitudinal relevance calculating unit obtains the longitudinal relevance of the meteorological data to the building data load according to the following formula:
forming an initial matrix through a grey correlation degree calculation formula based on meteorological data and the building data load;
the initial matrix is shown as follows:
Figure BDA0001837786380000121
in the formula, XnN is 1,2 … i; y is0Is the building data load;
calculating the longitudinal correlation degree of the meteorological data to the building data load by utilizing a grey correlation analysis algorithm based on an entropy weight method to obtain a longitudinal correlation degree matrix;
the longitudinal correlation matrix is shown as follows:
Figure BDA0001837786380000122
in the formula, XiThe average data per day of the ith meteorological data sequence, i is 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
The lateral correlation degree calculation unit calculates the lateral correlation degree of the meteorological data by the following formula:
calculating the transverse correlation degree between meteorological data by using a grey correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix;
the transverse correlation matrix is shown as follows:
Figure BDA0001837786380000123
wherein, XixjIndicating the horizontal relevance of the ith meteorological data to the jth meteorological data, wherein i is 0 and 1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
The meteorological data input quantity acquisition unit comprises:
a first weather data acquisition subunit: dividing all the meteorological data with the transverse association degree larger than a preset threshold into a first data group, comparing the longitudinal association degree of each meteorological data in the first data group to the building data load, and selecting first meteorological data with the maximum longitudinal association degree;
the second meteorological data acquisition subunit: dividing all the meteorological data with the cross line correlation degree smaller than a preset threshold into a second data group, comparing the longitudinal correlation degree of each meteorological data in the second data group to the building data load, and selecting the second meteorological data with the maximum longitudinal correlation degree;
a predicted input variable acquisition subunit: the first meteorological data and the second meteorological data are predicted input variables of meteorological data.
A prediction module comprising:
reading historical load data and input variable data, and performing normalization processing;
and training the BP neural network by using the normalized data to obtain the trained BP neural network.
And substituting the prediction input variable of the mobile base station data and the prediction input variable of the meteorological data into the trained prediction neural network for prediction to obtain the prediction result of the short-term load of the building.
The data acquisition module acquires the following variables:
the mobile base station data comprises: time labels, number of floors reached, number of floors changed, building name and number of people changed;
the meteorological data, comprising: temperature, dew point temperature, relative humidity, wind speed, and barometric pressure;
the building load data comprises: total load, energy for lighting and sockets, energy for air conditioning, energy for power and other energy;
the quasi-feature variable comprises: the number of people going upstairs, downstairs, immobilizers, variators, upstairs and immobilizers, downstairs and immobilizers, and the total number of people.
Example 3:
and acquiring the measured mobile base station data, the meteorological data and the load data of a certain building from 0 point 15 in 12/1/31/23 point 45 in 2017. In the above measured data, the mobile base station data and the load data are acquired once for 15 minutes, the total number of measured data is 96 points all day, and the meteorological data is acquired once for 30 minutes. Wherein the mobile base station data comprises: time labels, number of floors reached, number of floors changed, building name, number of people changed; the meteorological data includes: temperature, dew point temperature, relative humidity, wind speed, barometric pressure; the building load data includes: total load, energy for lighting and sockets, energy for air conditioning, energy for power, and energy for other uses. Examples of the data are shown in tables 1 to 3.
TABLE 1 example of data for a building moving base station
Figure BDA0001837786380000131
TABLE 2 weather data example of a building
Figure BDA0001837786380000141
TABLE 3 example of load data of a building
Figure BDA0001837786380000142
Step (2) is carried out: and carrying out data preprocessing on the data of the mobile base station to remove bad data. Then, the number of people going upstairs, the number of people going downstairs and the number of people not moving at each time point are extracted according to the positive and negative of the number of the changed floors marked at the same time, and the 3 groups of data are combined in a cross mode to obtain 7 quasi-feature variables of the number of the changed people, the number of the people going upstairs and the number of the people not moving, the number of the people going downstairs and the number of the people not moving.
Carrying out step (3): calculating the longitudinal relevance of the quasi-characteristic variables and the meteorological data to the load data; and calculating the transverse correlation degree between the mobile base station and the meteorological data. The correlation results are shown in tables 4 to 6.
TABLE 4 results of vertical grey correlation for a building
Figure BDA0001837786380000143
Figure BDA0001837786380000151
TABLE 5 horizontal correlation results of data of a mobile base station of a building
Figure BDA0001837786380000152
TABLE 6 lateral correlation results of meteorological data for a building
Figure BDA0001837786380000153
And vertically comparing, wherein the correlation degree of the air conditioning energy is the largest in the correlation degrees of the mobile base station data and various loads. Therefore, the mobile base station data mainly influences the total energy consumption through air conditioning energy, and then the mobile base station data comprises power energy, illumination and socket energy and other energy.
In the horizontal comparison, in the data hierarchy of the moving base station, the number of the fluctuating persons is closely related to 5 variables except the number of the persons going downstairs, so that the number of the fluctuating persons with higher degree of relation with the total load and the number of the persons going downstairs with less close relation with the number of the fluctuating persons can be selected as the feature data of the moving base station data. In the meteorological data hierarchy, the relation between each meteorological variable is not close, so that the temperature, the dew point temperature, the relative humidity, the air pressure and the wind speed are all used as characteristic variables of the meteorological data.
Carrying out the step (4): selecting the number of the variable people and the number of the people going downstairs as characteristic variables of the mobile base station, selecting a model only considering meteorological factors and a model considering the characteristic data and the meteorological data of the mobile base station, and performing short-term load prediction test on a certain building by using a BP neural network. And selecting the mobile base station data, the meteorological data and the total load data which are subjected to data preprocessing and matched with the data as an initial data set. The initial data set is divided into a training set and a testing set according to the ratio of 9:1, and the training set and the testing set are respectively used for training a neural network and testing a prediction effect. The obtained prediction results are shown in fig. 4 and 5.
As can be seen from the comparison results, the prediction results are extremely poor when the mobile base station is not considered, and do not conform to the curve trend of the true values; when the mobile base station is considered, the predicted result has better consistency with the trend of the true value curve. Therefore, the accuracy of the building electrical load curve prediction can be improved by considering the data of the mobile base station.
It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The present invention is not limited to the above embodiments, and any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are included in the scope of the claims of the present invention which are filed as the application.

Claims (18)

1. A building short-term load prediction method is characterized by comprising
Performing relevancy analysis on the acquired mobile base station data, building load data and preset quasi-characteristic variables at the same time to obtain quasi-characteristic input variables;
performing correlation analysis on the acquired meteorological data and building load data at the same time to obtain meteorological input variables;
and substituting the quasi-characteristic input variable and the meteorological input variable into a pre-trained BP neural network to predict the building load.
2. The building short-term load prediction method as claimed in claim 1, wherein the obtaining of the pseudo-characteristic input variable based on the correlation degree analysis of the acquired mobile base station data, building load data and the preset pseudo-characteristic variable at the same time comprises:
obtaining a longitudinal correlation degree between the quasi-characteristic variable and the building load data and a transverse correlation degree between the quasi-characteristic variables through an entropy weight method;
and carrying out dimensionality reduction screening on the transverse correlation degree and the longitudinal correlation degree to obtain a quasi-characteristic variable input variable.
3. The method as claimed in claim 2, wherein said finding the vertical correlation between said pseudo-characteristic variables and said building load data and the horizontal correlation between said pseudo-characteristic variables by entropy weight method comprises:
obtaining an initial matrix through a grey correlation degree calculation formula based on quasi-characteristic variables and the building data load;
calculating the longitudinal correlation degree of the quasi-characteristic variable to the dependent variable based on the initial matrix to obtain a longitudinal correlation degree matrix, so as to obtain the longitudinal correlation degree of the quasi-characteristic variable to the building data load;
and calculating the transverse correlation degree between the quasi-characteristic variables by utilizing a gray correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix.
4. The method as claimed in claim 3, wherein the initial matrix is expressed as follows:
Figure FDA0001837786370000011
in the formula, XiIs the average data per day of the ith sequence of pseudocharacteristic variables, i ═ 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
5. The method as claimed in claim 3, wherein the vertical correlation matrix is expressed as follows:
Figure FDA0001837786370000021
wherein, X0y1Representing the vertical relevance of the ith pseudo-characteristic variable to j building data loads, wherein i is 0 and 1 … m; j is 1,2 … n.
6. The method as claimed in claim 3, wherein the transverse correlation matrix is represented by the following equation:
Figure FDA0001837786370000022
wherein, XixjRepresenting the ith quasi-characteristic variable pair of the jth stationThe cross correlation degree of the pseudo-characteristic variable, i is 0,1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
7. The method as claimed in claim 2, wherein said finding the vertical correlation between said pseudo-characteristic variables and said building load data and the horizontal correlation between said pseudo-characteristic variables by entropy weight method comprises:
dividing quasi-feature variables corresponding to the transverse association degrees which are greater than a preset threshold into a first variable group, comparing longitudinal association degrees corresponding to the quasi-feature variables in the first variable group, and selecting a first quasi-feature variable with the maximum longitudinal association degree;
dividing quasi-characteristic variables corresponding to the transverse association degrees smaller than a preset threshold into second variable groups, comparing longitudinal association degrees corresponding to the quasi-characteristic variables in the second variable groups, and selecting the second quasi-characteristic variable with the maximum longitudinal association degree;
the first quasi characteristic variable and the second quasi characteristic variable are prediction input variables of mobile base station data.
8. The method as claimed in claim 1, wherein said correlation analysis based on weather data and building load data obtained at the same time to obtain weather input variables comprises
Obtaining the longitudinal correlation degree between the meteorological data and the building load data and the transverse correlation degree between the meteorological data through an entropy weight method;
and obtaining meteorological data input variables by carrying out dimensionality reduction screening on the transverse relevance degree and the longitudinal relevance degree.
9. The method as claimed in claim 8, wherein said determining the vertical correlation between said weather data and said building load data and the horizontal correlation between said weather data by entropy weight method comprises:
obtaining a meteorological data initial matrix through a grey correlation degree calculation formula based on meteorological data and the building data load;
calculating the longitudinal correlation degree of the meteorological data to the building data load based on the meteorological data initial matrix to obtain a longitudinal correlation degree matrix;
and calculating the transverse correlation degree between the meteorological data by using a grey correlation analysis algorithm based on an entropy weight method to obtain a transverse correlation degree matrix.
10. The method of claim 9, wherein the initial matrix of weather data is as follows:
Figure FDA0001837786370000031
in the formula, XiThe average data per day of the ith meteorological data sequence, i is 1,2 … i; y is0Is the daily average data of the building data load sequence, and j is the jth building data load.
11. The method as claimed in claim 9, wherein the vertical correlation matrix is expressed by the following formula:
Figure FDA0001837786370000032
wherein, X0y1Representing the vertical relevance of the ith meteorological data to the j types of building data loads, wherein i is 0 and 1 … m; j is 1,2 … n.
12. The method as claimed in claim 9, wherein said lateral correlation matrix is represented by the following equation:
Figure FDA0001837786370000033
wherein, XixjIndicating the horizontal relevance of the ith meteorological data to the jth meteorological data, wherein i is 0 and 1 … m; j is 0,1 … m; if i is j, then Xixj＝0。
13. The building short-term load forecasting method as claimed in claim 8, wherein the step of subjecting the horizontal correlation degree and the vertical correlation degree to dimension reduction screening to obtain meteorological data input variables comprises the steps of:
dividing quasi-characteristic variables corresponding to the transverse correlation degrees which are greater than a preset threshold value into a first data group, comparing longitudinal correlation degrees corresponding to meteorological data in the first data group, and selecting first meteorological data with the maximum longitudinal correlation degree;
dividing meteorological data corresponding to the transverse correlation degree smaller than a preset threshold into a second data group, comparing longitudinal correlation degrees corresponding to the meteorological data in the second data group, and selecting second meteorological data with the maximum longitudinal correlation degree;
the first meteorological data and the second meteorological data are predicted input variables of meteorological data.
14. The method as claimed in claim 1, wherein said introducing said pseudo-characteristic input variables and said meteorological input variables into a pre-trained BP neural network to predict said building load, previously comprises:
selecting mobile base station data, meteorological data and total load data which are subjected to data preprocessing and matched with the data as an initial data set;
and dividing the initial data set into a training set and a testing set according to the ratio of 9:1, and respectively training the neural network and testing the prediction effect to obtain the trained BP neural network.
15. The method as claimed in claim 1, wherein the short-term building load prediction method,
the mobile base station data comprises: time labels, number of floors reached, number of floors changed, building name and number of people changed;
the meteorological data, comprising: temperature, dew point temperature, relative humidity, wind speed, and barometric pressure;
the building load data comprises: total load, energy for lighting and sockets, energy for air conditioning, energy for power, and other energy uses.
The quasi-feature variable comprises: the number of people going upstairs, downstairs, immobilizers, variators, upstairs and immobilizers, downstairs and immobilizers, and the total number of people.
16. A building short term load prediction system, comprising:
a data acquisition module: acquiring mobile base station data, meteorological data and building load data with the same time, and extracting quasi-characteristic variables of the mobile base station data;
an input variable acquisition module: obtaining an input variable of load prediction through an entropy weight method based on the quasi-characteristic variable, the meteorological data and the building load data;
a prediction module: and predicting the building load through a BP neural network according to the input variable.
17. The building short term load prediction system as claimed in claim 16, wherein said input variable acquisition module comprises:
a quasi-characteristic variable input quantity acquisition submodule: based on the quasi-characteristic variable, performing longitudinal correlation degree calculation and transverse correlation degree calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a mobile base station data input variable;
a meteorological data input variable acquisition submodule: and based on the meteorological data, performing longitudinal correlation calculation and transverse correlation calculation through a gray correlation analysis method of the entropy weight method to obtain a transverse analysis result as a meteorological data input variable.
18. The building short term load forecasting system as claimed in claim 17, wherein said pseudo-characteristic variable input quantity obtaining sub-module includes:
a vertical relevance calculating unit: calculating the longitudinal correlation degree of the quasi-characteristic variable to the building data load through the grey correlation analysis method;
a lateral correlation degree calculation unit: calculating the transverse correlation degree among the quasi-characteristic variables by the gray correlation analysis method;
a mobile base station data input amount acquisition unit: and determining a prediction input variable of the mobile base station data based on the transverse correlation degree, a preset threshold value and the longitudinal correlation degree.
CN201811233985.7A 2018-10-23 2018-10-23 Building short-term load prediction method and system Active CN111091217B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811233985.7A CN111091217B (en) 2018-10-23 2018-10-23 Building short-term load prediction method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811233985.7A CN111091217B (en) 2018-10-23 2018-10-23 Building short-term load prediction method and system

Publications (2)

Publication Number Publication Date
CN111091217A true CN111091217A (en) 2020-05-01
CN111091217B CN111091217B (en) 2022-07-08

Family

ID=70392125

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811233985.7A Active CN111091217B (en) 2018-10-23 2018-10-23 Building short-term load prediction method and system

Country Status (1)

Country Link
CN (1) CN111091217B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014104829A1 (en) * 2012-12-28 2014-07-03 주식회사 효성 Method for predicting power load and apparatus using the method
CN104778503A (en) * 2015-03-16 2015-07-15 国家电网公司 Method for forecasting public building air conditioner short-time base wire load with consideration of real-time weather factors
CN106022954A (en) * 2016-05-16 2016-10-12 四川大学 Multiple BP neural network load prediction method based on grey correlation degree
KR20180053452A (en) * 2016-11-11 2018-05-23 한국에너지기술연구원 Method for predicting energy demand, and distributed energy management method and system
CN108416366A (en) * 2018-02-06 2018-08-17 武汉大学 A kind of power-system short-term load forecasting method of the weighting LS-SVM based on Meteorological Index

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014104829A1 (en) * 2012-12-28 2014-07-03 주식회사 효성 Method for predicting power load and apparatus using the method
CN104778503A (en) * 2015-03-16 2015-07-15 国家电网公司 Method for forecasting public building air conditioner short-time base wire load with consideration of real-time weather factors
CN106022954A (en) * 2016-05-16 2016-10-12 四川大学 Multiple BP neural network load prediction method based on grey correlation degree
KR20180053452A (en) * 2016-11-11 2018-05-23 한국에너지기술연구원 Method for predicting energy demand, and distributed energy management method and system
CN108416366A (en) * 2018-02-06 2018-08-17 武汉大学 A kind of power-system short-term load forecasting method of the weighting LS-SVM based on Meteorological Index

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
NASPI,F;ARNESANO,M;ZAMPETTI,L;ET AL: "Experimental study on occupants" interaction with windows and lights in Mediterranean offices during the non-heating season", 《BUILDING AND ENVIRONMENT》 *
王雁凌；吴梦凯；周子青；马洪宇: "基于改进灰色关联度的电力负荷影响因素量化分析模型", 《电网技术》 *

Also Published As

Publication number Publication date
CN111091217B (en) 2022-07-08

Similar Documents

Publication Publication Date Title
Yu et al. Low-carbon city pilot and carbon emission efficiency: Quasi-experimental evidence from China
CN107730054B (en) Gas load combined prediction method based on support vector regression
CN112508118B (en) Target object behavior prediction method aiming at data offset and related equipment thereof
CN103093285A (en) Short-term load forecast method based on artificial neural network
CN109657962B (en) Method and system for evaluating sound quantity assets of brands
CN110689162B (en) Bus load prediction method, device and system based on user side classification
Bassi et al. Building energy consumption forecasting: A comparison of gradient boosting models
CN103700030A (en) Grey rough set-based power grid construction project post-evaluation index weight assignment method
Zekić-Sušac et al. Cluster analysis and artificial neural networks in predicting energy efficiency of public buildings as a cost-saving approach
CN115330130A (en) Agricultural park comprehensive energy system evaluation method based on improved AHP-CRITIC
Dönmez et al. Developing statistical optimization models for Urban Competitiveness Index: Under the boundaries of econophysics approach
CN113887809A (en) Power distribution network supply and demand balance method, system, medium and computing equipment under double-carbon target
CN111091217B (en) Building short-term load prediction method and system
CN111080037A (en) Short-term power load prediction method and device based on deep neural network
Guo et al. Mobile user credit prediction based on lightgbm
CN114240102A (en) Line loss abnormal data identification method and device, electronic equipment and storage medium
CN109753990B (en) User electric energy substitution potential prediction method, system and storage medium
CN113688120A (en) Quality detection method and device for data warehouse and electronic equipment
Liu Effective combination and application path of computer big data analysis and cloud computing network technology
Deng et al. A survey of stock forecasting model based on artificial intelligence algorithm
CN116702059B (en) Intelligent production workshop management system based on Internet of things
CN112365280B (en) Electric power demand prediction method and device
CN117311244B (en) Energy-saving regulation and control method and system based on equipment working condition prediction
Yang et al. Research on the development level evaluation of smart city in China
Zhou et al. Combined grey forecast model with particle swarm optimizer

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant