CN105205558B - A kind of BP neural network model configuration method towards building energy consumption prediction - Google Patents
A kind of BP neural network model configuration method towards building energy consumption prediction Download PDFInfo
- Publication number
- CN105205558B CN105205558B CN201510613729.0A CN201510613729A CN105205558B CN 105205558 B CN105205558 B CN 105205558B CN 201510613729 A CN201510613729 A CN 201510613729A CN 105205558 B CN105205558 B CN 105205558B
- Authority
- CN
- China
- Prior art keywords
- prediction
- network model
- training
- energy consumption
- neural network
- 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.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of BP neural network model configuration methods towards building energy consumption prediction, including（A）, create the BP neural network model of configurableization；（B）, the BP neural network model of the embedded power consumption mode of training, formation prediction training pattern；（C）, carry out building energy consumption data prediction.The BP neural network model configuration method towards building energy consumption prediction of the present invention, complicated building energy consumption prediction Correlative Influence Factors, parameter, algorithm are carried out decoupling with building energy consumption actual prediction function and detached, realize that configurableization of arbitrary prediction model creates, reduce the development and maintenance cost of building energy consumption forecasting system, so that the Energy Consumption Prediction System built on a large scale is easily achieved, have a good application prospect.
Description
Technical field
The present invention relates to a kind of BP neural network model configuration methods towards building energy consumption prediction, and it is pre- to belong to building energy consumption
Survey technology field.
Background technology
Sharp increase with Chinese population and people's living standards continue to improve, China's construction area are multiplied,
Building energy consumption expense is also continuously increased.According to statistics, China's building energy consumption accounts for about the 1/3 of whole society's total energy consumption.Currently, building energy consumption
Prediction mainly according to history consumption information and relevant context information, is predicted for the possibility of future architecture with energy situation.It builds
Administrative staff's reasonable arrangement system operation mode can not only be helped by building energy consumption prediction, but also can utilize actual value and prediction
The comparison of value carries out building energy consumption abnormality detection, to ensure that the safety with energy carries out.
In recent years, the method that being much mainly concerned with has occurred in the research in relation to building energy consumption prediction technique has BP refreshing
Through network algorithm, support vector machines, Time series analysis method etc..But existing method is mostly with single building, single energy consumption
Implement progress for object, result of study all finally provides a preferable experimental model as optimum prediction model.Though in addition,
The theoretical method research of right building energy consumption prediction is more, however, blank but occurs in actual building energy consumption forecasting system, mainly
When being predicted due to practical building energy consumption, for architectural object scale it is more, energy consumption wide variety, for different buildings, no
Same classification subitem energy consumption prediction model is then different, and the single prediction model given by theoretical research is difficult to meet actual complex
Therefore how changeable forecasting system requirement establishes a kind of effective, flexible, configurable prediction model configuration design side
Case meets the needs of actual prediction system complex is changeable, is current urgent problem.
Invention content
It is difficult to meet the changeable prediction system of actual complex the purpose of the invention is to overcome existing Individual forecast model
The problem of system requires.The BP neural network model configuration method towards building energy consumption prediction of the present invention, by complicated building energy
Consumption prediction Correlative Influence Factors, parameter, algorithm carry out decoupling with building energy consumption actual prediction function and detach, and realize arbitrary prediction mould
Configurableization of type creates, and reduces the development and maintenance cost of building energy consumption forecasting system so that the energy consumption built on a large scale
Forecasting system is easily achieved, and is had a good application prospect.
In order to achieve the above object, the technical solution adopted in the present invention is：
A kind of BP neural network model configuration method towards building energy consumption prediction, it is characterised in that：Include the following steps,
Step (A) creates the BP neural network model of configurableization, including establishes input configuration, output configuration, network knot
Structure configures, and carries out model checking, the preservation of basic network model；
Step (B), the BP neural network model of the embedded power consumption mode of training form prediction training pattern, including prediction instruction
Practice model selection, the prediction input data time is arranged, building energy consumption is predicted, prediction result preserves；
Step (C) carries out building energy consumption data prediction, including the selection of prediction training pattern, prediction input data time are set
It sets, building energy consumption is predicted, prediction result preserves.
BP neural network model configuration method above-mentioned towards building energy consumption prediction, it is characterised in that：Step (A), wound
The BP neural network model of configurableization, including following procedure are built,
(A1) influence factor of energy consumption prediction is abstracted, defines input storage mode, the solution of BP neural network model
Analysis mode and data load mode；
(A2) storage mode, analysis mode and the data load mode of the output of BP neural network model are defined；
(A3) the range screening of BP neural network model is carried out；
(A4) input configuration is established, the input factor realization that gamut is screened according to the range of BP neural network model is arbitrary
The selection of input factor, arbitrary input number selection, meanwhile, the related data extraction involved by each input factor is related to
Parameter be configured；
(A5) output configuration is established, the output factor realization that gamut is screened according to the range of BP neural network model is arbitrary
The selection of output factor, arbitrary output number selection, meanwhile, the related data extraction involved by each output factor is related to
Parameter be configured；
(A6) network structure configuration is established, includes the neuron number to input layer, hidden layer neuron number, output layer
Neuron number, the configuration of cycle of training, predetermined period；
(A7) model checking is carried out, the BP neural network model being configured is verified, including input and output are matched
Set the matching verification with network structure, the configuration verification of inputoutput data extracting parameter；
(A8) basic network model preservation is carried out, including preserves network structure configuration information, input configuration information, input and match
Confidence ceases.
BP neural network model configuration method above-mentioned towards building energy consumption prediction, it is characterised in that：Step (B), instruction
The BP neural network model for practicing embedded power consumption mode forms prediction training pattern, including following procedure,
(B1) power consumption mode configuration is filtered to training data according to the energy consumption model of festivals or holidays, situation of whether going to work
Extraction training, ensures the accuracy of prediction BP neural network model；Training parameter is arranged, including weight correction factor, study speed
The setting of rate, training algebraically；Training data extraction time is arranged, and is carried out with the end time at the beginning of being extracted to training data
Setting, when carrying out hands-on, then the time range has been configured according to the BP neural network model of selection input, export into
The corresponding data extraction of row；
(B2) basic network model selects, and selection is carried out to the BP neural network model kept in step (A)
Selection；
(B3) to the basic network model training of selection, prediction training pattern is formed, according to the energy consumption mould being configured
Formula, training parameter, training data extraction time range select the basic network model of selection using BP neural network algorithm
Take training；
(B4) preservation for predicting training pattern, the related data before training, after training is preserved, related data includes
Basic network model ID, training time, training data extraction time started, training data extraction end time, weight amendment system
The weight information of number, learning rate, normalization maximum value, normalization minimum value, training algebraically, input layer to hidden layer, implies
Layer arrives the weight information of output layer.
BP neural network model configuration method above-mentioned towards building energy consumption prediction, it is characterised in that：Step (C), into
Row building energy consumption data prediction, including following procedure,
(C1) prediction training pattern selection, to step (B) trained prediction training pattern selection, and to need into
The period of row prediction is configured；
(C2) building energy consumption is predicted, according to the prediction training pattern of selection, it is pre- to carry out energy consumption using BP neural network algorithm
It surveys；
(C3) prediction result preserves, and is preserved to the data that building energy consumption predicts, and shows and tracks convenient for data, packet
Include the training pattern ID of selection, predicted time, prediction of energy consumption value.
The beneficial effects of the invention are as follows：The BP neural network model configuration method towards building energy consumption prediction of the present invention,
Complicated building energy consumption prediction Correlative Influence Factors, parameter, algorithm are carried out decoupling with building energy consumption actual prediction function and detached,
It realizes that configurableization of arbitrary prediction model creates, reduces the development and maintenance cost of building energy consumption forecasting system so that big
The Energy Consumption Prediction System of scale building is easily achieved, and is had a good application prospect.
Description of the drawings
Fig. 1 is the flow chart for the BP neural network model configuration method of the present invention predicted towards building energy consumption.
Fig. 2 is the flow chart of the formation prediction training pattern of the present invention.
Fig. 3 is the flow chart of the progress building energy consumption data prediction of the present invention.
Specific implementation mode
Below in conjunction with Figure of description, the invention will be further described.Following embodiment is only used for clearly
Illustrate technical scheme of the present invention, and not intended to limit the protection scope of the present invention.
The BP neural network model configuration method towards building energy consumption prediction of the present invention, complicated building energy consumption is predicted
Correlative Influence Factors, parameter, algorithm and building energy consumption actual prediction function carry out decoupling and detach, and that realizes arbitrary prediction model can
Configurationization creates, and reduces the development and maintenance cost of building energy consumption forecasting system so that the energy consumption prediction system built on a large scale
System is easily achieved, as shown in Figure 1, include the following steps,
Step (A) creates the BP neural network model of configurableization, including establishes input configuration, output configuration, network knot
Structure configures, and carries out model checking, the preservation of basic network model, and establishment process is as follows,
(A1) influence factor of energy consumption prediction is abstracted, defines input storage mode, the solution of BP neural network model
Analysis mode and data load mode；
(A2) storage mode, analysis mode and the data load mode of the output of BP neural network model are defined；
(A3) the range screening of BP neural network model is carried out；
(A4) input configuration is established, the input factor realization that gamut is screened according to the range of BP neural network model is arbitrary
The selection of input factor, arbitrary input number selection, meanwhile, the related data extraction involved by each input factor is related to
Parameter be configured；
(A5) output configuration is established, the output factor realization that gamut is screened according to the range of BP neural network model is arbitrary
The selection of output factor, arbitrary output number selection, meanwhile, the related data extraction involved by each output factor is related to
Parameter be configured；
(A6) network structure configuration is established, includes the neuron number to input layer, hidden layer neuron number, output layer
Neuron number, the configuration of cycle of training, predetermined period；
(A7) model checking is carried out, the BP neural network model being configured is verified, including input and output are matched
Set the matching verification with network structure, the configuration verification of inputoutput data extracting parameter；
(A8) basic network model preservation is carried out, including preserves network structure configuration information, input configuration information, input and match
Confidence ceases；
Step (B), the BP neural network model of the embedded power consumption mode of training form prediction training pattern, including prediction instruction
To practice model selection, predicts that the setting of input data time, building energy consumption prediction, prediction result preserve, forming process is as follows,
(B1) power consumption mode configuration is filtered to training data according to the energy consumption model of festivals or holidays, situation of whether going to work
Extraction training, ensures the accuracy of prediction BP neural network model；Training parameter is arranged, including weight correction factor, study speed
The setting of rate, training algebraically；Training data extraction time is arranged, and is carried out with the end time at the beginning of being extracted to training data
Setting, when carrying out hands-on, then the time range has been configured according to the BP neural network model of selection input, export into
The corresponding data extraction of row；
(B2) basic network model selects, and selection is carried out to the BP neural network model kept in step (A)
Selection；
(B3) to the basic network model training of selection, prediction training pattern is formed, according to the energy consumption mould being configured
Formula, training parameter, training data extraction time range select the basic network model of selection using BP neural network algorithm
Take training；
Step (C) carries out building energy consumption data prediction, including the selection of prediction training pattern, prediction input data time are set
Set, building energy consumption prediction, prediction result preserve, prediction process it is as follows,
(C1) prediction training pattern selection, to step (B) trained prediction training pattern selection, and to need into
The period of row prediction is configured；
(C2) building energy consumption is predicted, according to the prediction training pattern of selection, it is pre- to carry out energy consumption using BP neural network algorithm
It surveys；
(C3) prediction result preserves, and is preserved to the data that building energy consumption predicts, and is convenient for data tracing, including selection
Training pattern ID, predicted time, prediction of energy consumption value.
Above-mentioned step (A)-step (C), the BP neural network model of foundation form prediction training pattern, selection prediction
Training pattern, has sequencing relationship when execution, and the call relation between model is one-to-many relationship, i.e. a BP nerve
Network model can correspond to multiple prediction training patterns, and a prediction training pattern can carry out the prediction of multiple data.
As shown in Fig. 2, the BP neural network model of the embedded power consumption mode of step (B) training, forms prediction training pattern
Flow chart parses basic network model, according to the input and output factor combined training data extraction time after parsing into
Row data are extracted, then progress data prediction, output data normalize successively, power consumption mode filters, BP neural network is trained,
Training pattern preservation and etc..
As shown in figure 3, step (C) carries out the flow chart of building energy consumption data prediction, instructed according to trained prediction
Practice model, is predicted using BP neural network algorithm.
BP neural network model configuration method according to the present invention towards building energy consumption prediction, introduces an embodiment,
1) the input factor of BP neural network model is described, the classification for defining BP neural network mode input has
Two layers, first layer is classified according to factor classification, such as temperature data, energy consumption data, and the second layer Main Basiss time carries out
Subitem, input factor description are as shown in table 1：
Table 1
2) output of BP neural network model is described, defines the classification of prediction model output with two layers, first
Layer is classified according to attribute classification, such as single building energy consumption data, garden energy consumption data, the second layer Main Basiss time into
Row subitem, output factor description are as shown in table 2：
Table 2
3) using unified format definition, storage and parsing, input building mode is, different for the input of network structure configuration
It is separated with ' & ' between input, inputs in inside and separated using ' | ' between relevant parameter, first item indicates in parameter
Classification, Section 2 indicate subitem, and first two constitute the data item for needing to extract altogether, and third to Section 6 indicates extraction data
The required parameter of item, the number for inputting neuron is the sum of ' & '.
Such as, Wei &1 is inputted | 1 | 110101 | | | &2 | 1 | 320100A000 | 01000 | |, indicate that input parameter shares two
?；First input is the temperature data of current hour, and the temperature data needed belongs to the temperature in 110101 cities；Second defeated
Enter for the energy consumption data of previous hour, specific data come from 01000 electricity consumption that building number is 320100A000；
The analytic method of input is corresponding with construction method, according to relevant character ' & ' ginseng related to the separation of ' | ' extraction
It is several；
The load mode of input data is：The number for needing to extract is determined according to the classification of the input after parsing and subitem information
According to table Table_A and extraction data row Column_A, it is related to parameter row Column_B and Column_C, then according to the 3rd to the 4th
Given parameter Param1, Param2 of item extracts data；
Extracting format is：Select Column_A from Table_A Where Column_B like"Param1"
and Column_C like"Param2"；
4) output of network structure configuration is defined using unified format, storage and parsing, output building mode are：Different
It is separated with ' & ' between input, exports in inside and separated using ' | ' between relevant parameter, first item indicates in parameter
Classification, Section 2 indicate subitem, and first two constitute the data item for needing to extract altogether, and third to Section 6 indicates extraction data
The required parameter of item；
Such as, Wei &1 is exported | 1 | 320100A000 | 01000 | |, indicate that the data building number that needs are predicted is
320100A000, energy consumption are classified as 01000 power load；
The analytic method of output is corresponding with construction method, according to relevant character ' & ' ginseng related to the separation of ' | ' extraction
It is several；
The load mode of output data is：The number for needing to extract is determined according to the classification of the output after parsing and subitem information
According to table Table_A and extraction data row Column_A, it is related to parameter row Column_B and Column_C, then according to the 3rd to the 4th
Given parameter Param1, Param2 of item extracts data；
Extracting format is：Select Column_A from Table_A Where Column_B like"Param1"
and Column_C like"Param2"；
5) weight, the weight of hidden layer to output layer of input layer to hidden layer are defined using unified format, store reconciliation
Analysis is separated between the row and row of weight matrix using character ' & ', being separated between the column and the column using character ' | ' in row, ' & ' word
The number of symbol is total line number of weight matrix, and the number of ' | ' character between front and back continuous two characters ' ' is weight matrix
Total columns,
Such as, the weight matrix of input layer to hidden layer is w：
After being built, weight storage form is：
&0.948|0.802|-2.052|2.662|&0.199|0.989|6.614|0.628|&0.576|0.755|0.712
|0.650|
The analytic method of weight is corresponding with construction method, and dependency number is extracted in the separation according to relevant character ' & ' and ' | '
According to item；
6) it defines power consumption mode to be indicated with 5 character bits, front two indicates time started hour, the 3rd, 4 expression hour
End time, the 5th is indicate whether festivals or holidays, as shown in table 3：
Table 3
The prediction model with festivals or holidays or the power consumption modes such as next of going to work is considered, in order to ensure the compatibility of prediction model
Property and scalability, after data prediction, according to energy consumption model carry out data filtering.As power consumption mode is
20081, then the total evidence of extracted setting period is filtered, the period will be more than or equal at 20 points, and be less than 08
The data of the festivals or holidays of point retain, other data removals.
7) it when manual model training, from top to bottom can successively be held according to Fig. 2 building energy consumption prediction model training managing flows
Row；Automodel is trained, model training uses period treatment mechanism, the adhoc basis defined in basic model saving interface
Model i items cycle of trainingWith the next training timeUsing the testing mechanism of the short time less than cycle of training,
Constantly training is then trained when reaching the next training time, is changed the next training time after training and isContinue to train detection；
8) when the test of manual energy consumption data prediction, can according to Fig. 3 building energy consumptions prediction model data prediction process flow by
It is executed successively under above；For automatic building energy consumption data prediction, using period treatment mechanism, in basic model saving interface
Define predetermined period item of adhoc basis model iWith next predicted time itemUsing short less than predetermined period
The testing mechanism of time is constantly predicted, when reaching next predicted time, is predicted, prediction next time is changed after prediction
Time isContinue to train detection.
The basic principles and main features and advantage of the present invention have been shown and described above.The technical staff of the industry should
Understand, the present invention is not limited to the above embodiments, and the above embodiments and description only describe the originals of the present invention
Reason, without departing from the spirit and scope of the present invention, various changes and improvements may be made to the invention, these changes and improvements
It all fall within the protetion scope of the claimed invention.The claimed scope of the invention is by appended claims and its equivalent circle
It is fixed.
Claims (1)
1. a kind of BP neural network model configuration method towards building energy consumption prediction, it is characterised in that：Include the following steps,
Step（A）, the BP neural network model of configurableization is created, including foundation input configures, exports configuration, network structure is matched
It sets, and carries out model checking, the preservation of basic network model；
Step（B）, the BP neural network model of the embedded power consumption mode of training forms and predicts training pattern, including prediction training mould
Type selection, the setting of prediction input data time, building energy consumption prediction, prediction result preserve；
Step（C）, building energy consumption data prediction is carried out, including the selection of prediction training pattern, prediction input data time are arranged, build
Build energy consumption prediction, prediction result preserves；
Wherein, step（A）, the BP neural network model of configurableization, including following procedure are created,
（A1）The influence factor of energy consumption prediction is abstracted, input storage mode, the parsing side of BP neural network model are defined
Formula and data load mode；
（A2）Define storage mode, analysis mode and the data load mode of the output of BP neural network model；
（A3）Carry out the range screening of BP neural network model；
（A4）Input configuration is established, the input factor that gamut is screened according to the range of BP neural network model realizes arbitrary input
Factor selection, arbitrary input number selection, meanwhile, the ginseng being related to is extracted to the related data involved by each input factor
Number is configured；
（A5）Output configuration is established, the output factor that gamut is screened according to the range of BP neural network model realizes arbitrary output
Factor selection, arbitrary output number selection, meanwhile, the ginseng being related to is extracted to the related data involved by each output factor
Number is configured；
（A6）Network structure configuration is established, includes the neuron number to input layer, hidden layer neuron number, output layer nerve
First number, the configuration of cycle of training, predetermined period；
（A7）Carry out model checking, the BP neural network model being configured is verified, including input and output configuration and
The matching of network structure verifies, the configuration verification of inputoutput data extracting parameter；
（A8）Basic network model preservation is carried out, including preserves network structure configuration information, input configuration information, output with confidence
Breath；
Step（B）, the BP neural network model of the embedded power consumption mode of training, formation prediction training pattern, including following procedure,
（B1）Power consumption mode configuration is to be filtered extraction to training data according to the energy consumption model of festivals or holidays, situation of whether going to work
Training ensures the accuracy of prediction BP neural network model；Training parameter is arranged, including weight correction factor, learning rate, instruction
Practice the setting of algebraically；Training data extraction time is arranged, and is configured with the end time at the beginning of being extracted to training data,
When carrying out hands-on, then the time range has been configured according to the BP neural network model of selection input, output progress phase
The data extraction answered；
（B2）Basic network model selects, and selection is to step（A）In the BP neural network model kept selected；
（B3）To the basic network model training of selection, prediction training pattern is formed, according to the power consumption mode being configured, is instructed
Practice parameter, training data extraction time range, selection instruction is carried out to the basic network model of selection using BP neural network algorithm
Practice；
（B4）The preservation for predicting training pattern preserves the related data before training, after training, and related data includes basis
Network model ID, it the training time, the training data extraction time started, the training data extraction end time, weight correction factor, learns
The weight information of rate, normalization maximum value, normalization minimum value, training algebraically, input layer to hidden layer is practised, hidden layer is to defeated
Go out the weight information of layer；
Step（C）, building energy consumption data prediction, including following procedure are carried out,
（C1）Training pattern selection is predicted, to step（B）Trained prediction training pattern selection, and it is pre- to needing to carry out
The period of survey is configured；
(C2) building energy consumption is predicted, according to the prediction training pattern of selection, energy consumption prediction is carried out using BP neural network algorithm；
(C3) prediction result preserves, and is preserved to the data that building energy consumption predicts, and shows and tracks convenient for data, including choosing
The training pattern ID selected, predicted time, prediction of energy consumption value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510613729.0A CN105205558B (en) | 2015-09-23 | 2015-09-23 | A kind of BP neural network model configuration method towards building energy consumption prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510613729.0A CN105205558B (en) | 2015-09-23 | 2015-09-23 | A kind of BP neural network model configuration method towards building energy consumption prediction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105205558A CN105205558A (en) | 2015-12-30 |
| CN105205558B true CN105205558B (en) | 2018-11-06 |
Family
ID=54953227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510613729.0A Active CN105205558B (en) | 2015-09-23 | 2015-09-23 | A kind of BP neural network model configuration method towards building energy consumption prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105205558B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106199174A (en) * | 2016-07-01 | 2016-12-07 | 广东技术师范学院 | Extruder energy consumption predicting abnormality method based on transfer learning |
| CN106874581B (en) * | 2016-12-30 | 2021-03-30 | 浙江大学 | Building air conditioner energy consumption prediction method based on BP neural network model |
| 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 |
| CN106933173A (en) * | 2017-03-06 | 2017-07-07 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | A kind of equipment for dyeing and finishing monitoring system based on CORBA |
| CN107844859B (en) * | 2017-10-31 | 2020-08-07 | 深圳达实智能股份有限公司 | Large medical equipment energy consumption prediction method based on artificial intelligence and terminal equipment |
| CN109919308B (en) | 2017-12-13 | 2022-11-11 | 腾讯科技（深圳）有限公司 | Neural network model deployment method, prediction method and related equipment |
| CN109102103A (en) * | 2018-06-26 | 2018-12-28 | 上海鲁班软件股份有限公司 | A kind of multi-class energy consumption prediction technique based on Recognition with Recurrent Neural Network |
| CN109033595B (en) * | 2018-07-16 | 2024-03-29 | 清华大学 | New method and device for building green performance simulation analysis |
| 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 |
| CN110569981A (en) * | 2019-07-22 | 2019-12-13 | 国网冀北电力有限公司电力科学研究院 | Method and device for determining fuel consumption of scattered coal |
| CN111752870B (en) * | 2020-05-28 | 2022-02-22 | 苏州浪潮智能科技有限公司 | Compatible method and system of server interface based on neural network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102938092A (en) * | 2012-10-08 | 2013-02-20 | 珠海派诺科技股份有限公司 | Prediction method of building energy consumption in festivals and holidays based on neural network |
| US20130325377A1 (en) * | 2012-05-31 | 2013-12-05 | Johnson Controls Technology Company | Systems and methods for measuring and verifying energy usage in a building |
| CN104239982A (en) * | 2014-10-12 | 2014-12-24 | 刘岩 | Method for predicting energy consumption of buildings during holidays and festivals on basis of time series and neural networks |
-
2015
- 2015-09-23 CN CN201510613729.0A patent/CN105205558B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130325377A1 (en) * | 2012-05-31 | 2013-12-05 | Johnson Controls Technology Company | Systems and methods for measuring and verifying energy usage in a building |
| CN102938092A (en) * | 2012-10-08 | 2013-02-20 | 珠海派诺科技股份有限公司 | Prediction method of building energy consumption in festivals and holidays based on neural network |
| CN104239982A (en) * | 2014-10-12 | 2014-12-24 | 刘岩 | Method for predicting energy consumption of buildings during holidays and festivals on basis of time series and neural networks |
Non-Patent Citations (3)
| Title |
|---|
| Rajesh Kumar 等.Energy analysis of a building using artificial neural network: A review.《Energy and Buildings》.2013,第65卷第352-358页. * |
| 刘长俊.基于遗传BP神经网络技术的大型公建能耗分析模型的研究与应用.《中国优秀硕士学位论文全文数据库工程科技II辑》.2013,(第02期),第C038-73页. * |
| 李然然 等.基于BP神经网络的建筑物用电能耗预测.《山东建筑大学学报》.2011,第26卷(第2期),第162-165页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105205558A (en) | 2015-12-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105205558B (en) | A kind of BP neural network model configuration method towards building energy consumption prediction | |
| CN107682216B (en) | A kind of network traffics protocol recognition method based on deep learning | |
| CN107066445B (en) | The deep learning method of one attribute emotion word vector | |
| CN107861942B (en) | Suspected power complaint work order identification method based on deep learning | |
| CN106650725A (en) | Full convolutional neural network-based candidate text box generation and text detection method | |
| CN110929918B (en) | 10kV feeder fault prediction method based on CNN and LightGBM | |
| CN108448610A (en) | A kind of short-term wind power prediction method based on deep learning | |
| CN107169575A (en) | A kind of modeling and method for visualizing machine learning training pattern | |
| CN107133960A (en) | Image crack dividing method based on depth convolutional neural networks | |
| CN107145559A (en) | Intelligent classroom Knowledge Management Platform and method based on semantic technology and gameization | |
| CN109272123A (en) | It is a kind of based on convolution-Recognition with Recurrent Neural Network sucker rod pump operating condition method for early warning | |
| CN107563122A (en) | The method of crime prediction of Recognition with Recurrent Neural Network is locally connected based on interleaving time sequence | |
| CN106295799A (en) | A kind of implementation method of degree of depth study multilayer neural network | |
| CN109101948A (en) | A kind of more attention mechanism video presentation methods based on space-time and channel | |
| CN107909206A (en) | A kind of PM2.5 Forecasting Methodologies based on deep structure Recognition with Recurrent Neural Network | |
| CN107784397A (en) | A kind of power network material requirements forecasting system and its Forecasting Methodology | |
| CN107882541B (en) | Water drive flow field generation method and device | |
| CN109783629A (en) | A kind of micro-blog event rumour detection method of amalgamation of global event relation information | |
| CN109977747A (en) | A kind of coal rock detection system of comprehensive space-time characteristic | |
| CN104182803B (en) | Wind-powered electricity generation data preprocessing method and wind power forecasting method and system | |
| CN107465664A (en) | Intrusion detection method based on parallel more artificial bee colony algorithms and SVMs | |
| CN109558484A (en) | Electric power customer service work order emotion quantitative analysis method based on similarity word order matrix | |
| CN106919164B (en) | Water conservancy unit failure analysis methods based on storehouse autocoder | |
| CN110364186A (en) | A kind of emotion identification method across language voice end to end based on confrontation study | |
| CN110516742A (en) | A kind of distribution terminal fault distinguishing method and system based on Combination neural network model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |