CN105631539A - Intelligent building energy consumption prediction method based on support vector machine - Google Patents
Intelligent building energy consumption prediction method based on support vector machine Download PDFInfo
- Publication number
- CN105631539A CN105631539A CN201510987926.9A CN201510987926A CN105631539A CN 105631539 A CN105631539 A CN 105631539A CN 201510987926 A CN201510987926 A CN 201510987926A CN 105631539 A CN105631539 A CN 105631539A
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- building
- data
- regression model
- day
- 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.)
- Pending
Links
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"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an intelligent building energy consumption prediction method based on a support vector machine, relating to the technical field of energy management. The technical problems of improving prediction precision and shortening a training time are solved. According to the method, the itemized collection of the building energy consumption data of a field and indoor and outdoor temperature and humidity, and missed building energy consumption data is filled; then a normalization method is used to convert the building energy consumption data into a dimensionless index evaluation value, and then a building energy consumption prediction sample is constructed according to the normalized data, the weather data of a prediction date and the date attribute of the prediction data; then according to the building energy consumption prediction sample, a svmtrain function is employed to create and train a SVR regression model; and then the SVR regression model is used to predict building energy consumption. The method provided by the invention is suitable for the energy consumption prediction of an intelligent building.
Description
Technical field
The present invention relates to energy conservation technology, particularly relate to the technology of a kind of intelligent building energy consumption Forecasting Methodology based on support vector machine.
Background technology
Intelligent building Building Energy Analysis is one of the study hotspot in building energy saving field, is the basis realizing reasonable energy-saving scheme, is realize the precondition that energy resource system optimization runs. The important process analyzing energy consumption of building is the own characteristic according to building construction thing, sets up the forecast model of reflection difference subitem energy consumption change. Energy consumption forecast model accurately, can predict and grasp the situation of its energy consumption in advance, thus to building with making planning in advance so that national energy policy is better implemented.
The prediction of current energy consumption can be largely classified into two big classes, and the traditional method that a class is representative with time series is another kind of, be representative with neutral net novel artificial intelligent method. If adding the factors of influence such as major holiday, accident, climate change, the precision of tradition energy consumption Forecasting Methodology has bigger deviation, and neural net prediction method can not thoroughly solve a difficult problem for network structure design, and the training time is longer.
Summary of the invention
For the defect existed in above-mentioned prior art, it is high that the technical problem to be solved is to provide a kind of precision of prediction, and the intelligent building energy consumption Forecasting Methodology based on support vector machine that the training time is short.
In order to solve above-mentioned technical problem, a kind of intelligent building energy consumption Forecasting Methodology based on support vector machine provided by the present invention, it is characterised in that specifically comprise the following steps that
1) the energy consumption of building data of subitem collection site, and indoor and outdoor temperature, humidity, and the energy consumption of building data of disappearance are filled up, specifically filling up formula is:
Wherein,For yDayThe energy consumption of building data that moment lacks,For y d-1 dayThe energy consumption of building data in moment,For y d-2 dayThe energy consumption of building data in moment,��For numerical value weight set in advance;
2) adopt method for normalizing that energy consumption of building data are converted to index without dimension test and appraisal value so that each desired value in energy consumption of building data is all in same number of levels;
3) according to the data after normalization, the meteorological data of prediction day and the day attribute of prediction day, the energy consumption of building forecast sample of timing structure is built;
4) according to energy consumption of building forecast sample, svmtrain function creation SVR regression model is adopted;
5) SVR regression model is trained;
6) the return value of SVR regression model and energy consumption of building actual value are contrasted, if difference between the two is more than range of error set in advance, then returns step 5, otherwise then go to step 7;
7) SVR forecast of regression model energy consumption of building is utilized, and using the return value of SVR regression model as energy consumption of building predictive value.
Intelligent building energy consumption Forecasting Methodology based on support vector machine provided by the invention, energy consumption data is obtained by building metering separate, by the reparation to historical data abnormity point, normalized, in conjunction with climate characteristic, indoor and outdoor temperature, the factors such as festivals or holidays are analyzed, characteristic according to energy consumption data, set up the forecast sample of timing structure, realize establishment and the training of Support Vector Machines for Regression model, the model finally utilizing establishment carries out energy consumption prediction, depending on precision of prediction, supporting vector machine model is modified, there is precision of prediction height, the feature that training time is short, better can provide foundation for the scheduling strategy of building powering device, realize building construction farthest energy-conservation.
Accompanying drawing explanation
Fig. 1 is the flow chart of the intelligent building energy consumption Forecasting Methodology based on support vector machine of the embodiment of the present invention.
Detailed description of the invention
Illustrate embodiments of the invention are described in further detail below in conjunction with accompanying drawing; but the present embodiment is not limited to the present invention; the analog structure of every employing present invention and similar change thereof; protection scope of the present invention all should be listed in; pause mark in the present invention all represents the relation of sum, and the English alphabet in the present invention is case sensitive.
As it is shown in figure 1, a kind of intelligent building energy consumption Forecasting Methodology based on support vector machine that the embodiment of the present invention provides, it is characterised in that specifically comprise the following steps that
1) the energy consumption of building data of subitem collection site, and indoor and outdoor temperature, humidity, and the energy consumption of building data of disappearance are filled up, specifically filling up formula is:
Wherein,For yDayThe energy consumption of building data that moment lacks,For y d-1 dayThe energy consumption of building data in moment,For y d-2 dayThe energy consumption of building data in moment,��For numerical value weight set in advance, the representative value of a1 is the representative value of 0.6, a2 is 0.4, a1, the value of a2 can be adjusted according to actual items;
The gatherer process of data there will be the situation of shortage of data or data exception, the existence of these abnormal datas causes the loss of substantial amounts of useful information, the definitiveness composition contained in system is made to be more difficult to hold, even can make generation entanglement in prediction process, need the data gathered are carried out pretreatment, repair the disappearance in history energy consumption sequence and abnormal data;
According to the energy consumption of building data gathered, meteorological data and the impact on energy consumption of building characteristic variations trend of the day attribute (Monday to Sunday) of prediction day, draw the relation curve of energy consumption of building and environmental factors, and according to relation curve, distributing the weight of these influence factors, meteorological data includes max. daily temperature, Daily minimum temperature, mean daily temperature, weather conditions (fine, cloudy, cloudy, rain);
2) adopt method for normalizing that energy consumption of building data are converted to index without dimension test and appraisal value, make each desired value in energy consumption of building data all in same number of levels, adopting the method that energy consumption of building data are converted to index without dimension test and appraisal value by method for normalizing is prior art, and concrete conversion formula is:
WhereinFor the energy consumption of building test and appraisal value after normalization,For energy consumption of building measured value,For input vector dimension (namely affecting the quantity of the factor of energy consumption of building),For the maximum of sample data,Minima for sample data;
3) according to the data after normalization, the meteorological data of prediction day and the day attribute of prediction day, building the energy consumption of building forecast sample of timing structure, the data structure of energy consumption of building forecast sample is:
,,;
Wherein,For the set of the energy consumption of building data strong with prediction time relatedness set in advance,For predicting the set of the meteorological data of day,For predict day day attribute set;
Owing to energy consumption data has timing, therefore forecast model is the structure of timing, sample input includes three parts: the load data stronger with prediction time relatedness, the meteorological data (max. daily temperature, Daily minimum temperature, mean daily temperature and weather conditions) of prediction day, it was predicted that the day attribute of day;
4) according to energy consumption of building forecast sample, svmtrain(training modeling is adopted) function creation SVR regression model, the creation method of SVR regression model is prior art;
5) SVR regression model being trained, the training method of SVR regression model is prior art;
6) the return value of SVR regression model and energy consumption of building actual value are contrasted, if difference between the two is more than range of error set in advance, then returns step 5, otherwise then go to step 7;
7) SVR forecast of regression model energy consumption of building is utilized, and using the return value of SVR regression model as energy consumption of building predictive value.
In the embodiment of the present invention, adopt svmtrain function creation and train comprising the following steps that of SVR regression model;
4.1) set containingThe training set sample of individual training sample to for,It isThe input column vector of individual training,,,ForTraining output valve, by these data projections to high latitude feature space, the linear regression function of foundation is:
Wherein,For linear regression function,For nonlinear mapping function,For weight vectors,For position side-play amount;
4.2) defining linear insensitive loss function is:
Wherein,ForCorresponding actual value,For error amount set in advance,Representative value be 0.0162;
4.3) slack variable is introduced, and will find��Problem formula be described as:
Wherein,For penalty factor,More big expression to training error more thanSample punishment more big,Define the error requirements of regression function,The error of more little expression regression function is more little;
4.4) Largrange(Lagrange is adopted) function, the formula of step 4.3 is converted to:
Wherein,For kernel function;
4.5) formula of solution procedure 4.4, obtaining optimal solution is, then have:
Wherein,For supporting vector number;
4.6) obtaining SVR regression model is:
Wherein, only argument sectionIt is not zero, corresponding sampleIt is support vector machine;
Consider the type of kernel function, the feature that the value of parameter is bigger to the performance impact of regression model, the method adopting cross validation finds best penalty factor c and parameter g (variance in kernel function), and give tacit consent to employing RBF kernel function, then the parameter training SVR regression model of the best is utilized, after SVR regression model is set up, by svmpredict function, regression model is carried out emulation testing, SVR regression model can be evaluated by error amount and the coefficient of determination that this function returns, if being not reaching to requirement, model parameter can be modified, the methods such as kernel function re-establish regression model, until meeting requirement.
Claims (1)
1. the intelligent building energy consumption Forecasting Methodology based on support vector machine, it is characterised in that specifically comprise the following steps that
1) the energy consumption of building data of subitem collection site, and indoor and outdoor temperature, humidity, and the energy consumption of building data of disappearance are filled up, specifically filling up formula is:
Wherein,For yDayThe energy consumption of building data that moment lacks,For y d-1 dayThe energy consumption of building data in moment,For y d-2 dayThe energy consumption of building data in moment,��For numerical value weight set in advance;
2) adopt method for normalizing that energy consumption of building data are converted to index without dimension test and appraisal value so that each desired value in energy consumption of building data is all in same number of levels;
3) according to the data after normalization, the meteorological data of prediction day and the day attribute of prediction day, the energy consumption of building forecast sample of timing structure is built;
4) according to energy consumption of building forecast sample, svmtrain function creation SVR regression model is adopted;
5) SVR regression model is trained;
6) the return value of SVR regression model and energy consumption of building actual value are contrasted, if difference between the two is more than range of error set in advance, then returns step 5, otherwise then go to step 7;
7) SVR forecast of regression model energy consumption of building is utilized, and using the return value of SVR regression model as energy consumption of building predictive value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510987926.9A CN105631539A (en) | 2015-12-25 | 2015-12-25 | Intelligent building energy consumption prediction method based on support vector machine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510987926.9A CN105631539A (en) | 2015-12-25 | 2015-12-25 | Intelligent building energy consumption prediction method based on support vector machine |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105631539A true CN105631539A (en) | 2016-06-01 |
Family
ID=56046446
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510987926.9A Pending CN105631539A (en) | 2015-12-25 | 2015-12-25 | Intelligent building energy consumption prediction method based on support vector machine |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105631539A (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106250905A (en) * | 2016-07-08 | 2016-12-21 | 复旦大学 | A kind of real time energy consumption method for detecting abnormality of combination colleges and universities building structure feature |
CN106910144A (en) * | 2017-01-11 | 2017-06-30 | 华南理工大学 | Based on timesharing carve it is actual with can coefficient heavy construction by when energy consumption on-line prediction method |
CN106991504A (en) * | 2017-05-09 | 2017-07-28 | 南京工业大学 | Building energy consumption Forecasting Methodology, system and building based on metering separate time series |
CN107315884A (en) * | 2017-07-04 | 2017-11-03 | 北京首钢自动化信息技术有限公司 | A kind of building energy consumption modeling method based on linear regression |
CN107590565A (en) * | 2017-09-08 | 2018-01-16 | 北京首钢自动化信息技术有限公司 | A kind of method and device for building building energy consumption forecast model |
CN107679660A (en) * | 2017-09-30 | 2018-02-09 | 山东建筑大学 | Based on SVMs by when building energy consumption Forecasting Methodology |
CN108317670A (en) * | 2018-03-02 | 2018-07-24 | 深圳市华之任有限公司 | A kind of refrigeration system energy-saving control method and system based on machine learning |
CN109035067A (en) * | 2018-07-19 | 2018-12-18 | 北京泰豪智能工程有限公司 | Building energy consumption processing method and processing device based on RF and ARMA algorithm |
CN110046751A (en) * | 2019-03-27 | 2019-07-23 | 上海建坤信息技术有限责任公司 | Multi-energy system dispatching method based on the prediction of radial base energy consumption and real-time energy efficiency |
CN110276480A (en) * | 2019-05-31 | 2019-09-24 | 武汉烽火富华电气有限责任公司 | A kind of analyzing and predicting method and system for building energy consumption management |
CN110826817A (en) * | 2019-11-18 | 2020-02-21 | 北京燃气能源发展有限公司 | Building distributed energy multi-energy collaborative operation strategy generation method and system |
CN112365029A (en) * | 2019-09-03 | 2021-02-12 | 深圳市得益节能科技股份有限公司 | Missing value processing method for air conditioner load prediction and air conditioner load prediction system |
CN112434787A (en) * | 2020-10-28 | 2021-03-02 | 西安交通大学 | Terminal space energy consumption prediction method based on building total energy consumption, medium and equipment |
CN113505537A (en) * | 2021-07-27 | 2021-10-15 | 上海上实龙创智能科技股份有限公司 | Building energy consumption detection method and device, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101494567A (en) * | 2008-08-29 | 2009-07-29 | 北京理工大学 | Detection method for distributed abnegation service aggression based on load prediction |
US20100211224A1 (en) * | 2008-12-19 | 2010-08-19 | EnaGea LLC | Heating and cooling control methods and systems |
CN102705957A (en) * | 2012-06-07 | 2012-10-03 | 华南理工大学 | Method and system for predicting hourly cooling load of central air-conditioner in office building on line |
CN102779228A (en) * | 2012-06-07 | 2012-11-14 | 华南理工大学 | Method and system for online prediction on cooling load of central air conditioner in marketplace buildings |
-
2015
- 2015-12-25 CN CN201510987926.9A patent/CN105631539A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101494567A (en) * | 2008-08-29 | 2009-07-29 | 北京理工大学 | Detection method for distributed abnegation service aggression based on load prediction |
US20100211224A1 (en) * | 2008-12-19 | 2010-08-19 | EnaGea LLC | Heating and cooling control methods and systems |
CN102705957A (en) * | 2012-06-07 | 2012-10-03 | 华南理工大学 | Method and system for predicting hourly cooling load of central air-conditioner in office building on line |
CN102779228A (en) * | 2012-06-07 | 2012-11-14 | 华南理工大学 | Method and system for online prediction on cooling load of central air conditioner in marketplace buildings |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106250905A (en) * | 2016-07-08 | 2016-12-21 | 复旦大学 | A kind of real time energy consumption method for detecting abnormality of combination colleges and universities building structure feature |
CN106250905B (en) * | 2016-07-08 | 2020-05-12 | 复旦大学 | Real-time energy consumption abnormity detection method combined with building structure characteristics of colleges and universities |
CN106910144A (en) * | 2017-01-11 | 2017-06-30 | 华南理工大学 | Based on timesharing carve it is actual with can coefficient heavy construction by when energy consumption on-line prediction method |
CN106910144B (en) * | 2017-01-11 | 2020-04-28 | 华南理工大学 | Large building time-by-time energy consumption online prediction method based on time-division actual energy consumption coefficient |
CN106991504A (en) * | 2017-05-09 | 2017-07-28 | 南京工业大学 | Building energy consumption Forecasting Methodology, system and building based on metering separate time series |
CN107315884A (en) * | 2017-07-04 | 2017-11-03 | 北京首钢自动化信息技术有限公司 | A kind of building energy consumption modeling method based on linear regression |
CN107315884B (en) * | 2017-07-04 | 2020-12-18 | 北京首钢自动化信息技术有限公司 | Building energy consumption modeling method based on linear regression |
CN107590565A (en) * | 2017-09-08 | 2018-01-16 | 北京首钢自动化信息技术有限公司 | A kind of method and device for building building energy consumption forecast model |
CN107679660A (en) * | 2017-09-30 | 2018-02-09 | 山东建筑大学 | Based on SVMs by when building energy consumption Forecasting Methodology |
CN108317670B (en) * | 2018-03-02 | 2020-01-07 | 深圳市华之任有限公司 | Refrigeration system energy-saving control method and system based on machine learning |
CN108317670A (en) * | 2018-03-02 | 2018-07-24 | 深圳市华之任有限公司 | A kind of refrigeration system energy-saving control method and system based on machine learning |
CN109035067A (en) * | 2018-07-19 | 2018-12-18 | 北京泰豪智能工程有限公司 | Building energy consumption processing method and processing device based on RF and ARMA algorithm |
CN110046751A (en) * | 2019-03-27 | 2019-07-23 | 上海建坤信息技术有限责任公司 | Multi-energy system dispatching method based on the prediction of radial base energy consumption and real-time energy efficiency |
CN110276480A (en) * | 2019-05-31 | 2019-09-24 | 武汉烽火富华电气有限责任公司 | A kind of analyzing and predicting method and system for building energy consumption management |
CN112365029A (en) * | 2019-09-03 | 2021-02-12 | 深圳市得益节能科技股份有限公司 | Missing value processing method for air conditioner load prediction and air conditioner load prediction system |
CN112365029B (en) * | 2019-09-03 | 2021-08-17 | 深圳市得益节能科技股份有限公司 | Missing value processing method for air conditioner load prediction and air conditioner load prediction system |
CN110826817A (en) * | 2019-11-18 | 2020-02-21 | 北京燃气能源发展有限公司 | Building distributed energy multi-energy collaborative operation strategy generation method and system |
CN110826817B (en) * | 2019-11-18 | 2024-02-20 | 北京燃气能源发展有限公司 | Building distributed energy multi-energy collaborative operation strategy generation method and system |
CN112434787A (en) * | 2020-10-28 | 2021-03-02 | 西安交通大学 | Terminal space energy consumption prediction method based on building total energy consumption, medium and equipment |
CN112434787B (en) * | 2020-10-28 | 2024-04-05 | 西安交通大学 | Terminal space energy consumption prediction method, medium and equipment based on total energy consumption of building |
CN113505537A (en) * | 2021-07-27 | 2021-10-15 | 上海上实龙创智能科技股份有限公司 | Building energy consumption detection method and device, computer equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105631539A (en) | Intelligent building energy consumption prediction method based on support vector machine | |
Ahmad et al. | A review on applications of ANN and SVM for building electrical energy consumption forecasting | |
Aman et al. | Improving energy use forecast for campus micro-grids using indirect indicators | |
Jamil et al. | Income and price elasticities of electricity demand: Aggregate and sector-wise analyses | |
Gaudard et al. | Climate change impacts on hydropower management | |
CN103218675B (en) | A kind of based on the short-term load forecasting method of cluster and moving window | |
CN104484723A (en) | Power transformer economic life prediction method based on life data | |
CN110110912B (en) | Photovoltaic power multi-model interval prediction method | |
CN109858728B (en) | Load prediction method based on industry-divided electricity utilization characteristic analysis | |
CN102945508B (en) | Model correction based wind power forecasting method | |
CN109659933A (en) | A kind of prediction technique of power quality containing distributed power distribution network based on deep learning model | |
CN105760970A (en) | Method for predicting AQI | |
CN102183621A (en) | Aquaculture dissolved oxygen concentration online forecasting method and system | |
CN102930352B (en) | Power grid basic construction project cost prediction method based on multi-core support vector regression | |
CN103606022A (en) | Short-term load prediction method | |
CN107426759A (en) | The Forecasting Methodology and system of newly-increased base station data portfolio | |
CN104504619B (en) | Two kinds consider that the monthly system of temperature and economic growth factor calls power predicating method | |
CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
CN103226736A (en) | Method for predicting medium and long term power load based on cluster analysis and target theory | |
CN105447770A (en) | Assessment method for applying power grid monitoring data to refined weather forecast | |
CN102509173A (en) | Markov chain based method for accurately forecasting power system loads | |
CN114118571A (en) | Heating heat load prediction method based on LSTM deep learning | |
CN103679289B (en) | Methods of electric load forecasting based on multiple regression extrapolation | |
CN103293424A (en) | Data center power usage efficiency (PUE) determination method | |
CN104063810A (en) | Electricity utilization factor influence model based on big data |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160601 |