CN102073785A - Daily gas load combination prediction method based on generalized dynamic fuzzy neural network - Google Patents
Daily gas load combination prediction method based on generalized dynamic fuzzy neural network Download PDFInfo
- Publication number
- CN102073785A CN102073785A CN 201010561217 CN201010561217A CN102073785A CN 102073785 A CN102073785 A CN 102073785A CN 201010561217 CN201010561217 CN 201010561217 CN 201010561217 A CN201010561217 A CN 201010561217A CN 102073785 A CN102073785 A CN 102073785A
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- historical
- daily load
- fuzzy neural
- 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
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Feedback Control In General (AREA)
Abstract
The invention provides a daily gas load combination prediction method based on a generalized dynamic fuzzy neural network, comprising the following steps: (1) acquiring historical urban gas record data which are used as historical time series data; (2) judging and processing abnormal data in the historical time series data; (3) carrying out differential processing on historical load time series, i.e. sample data, and predicting by adopting a generalized regressive neural network; (4) taking one-time accumulation generation data input into the historical load time series as the input of the network by using a gray neural network, outputting the one-time accumulation generation data corresponding to the predicted daily gas load, training the network, and finally carrying out one-time degressive inverse generation processing on the output value; and (5) taking predicted values obtained in the steps (3) and (4) as the input of the generalized dynamic fuzzy neural network, and grouping the data. The daily gas load combination prediction method is adopted according to the characteristics of randomness, instability, periodicity and the like of the daily gas load, thereby having higher prediction precision.
Description
Technical field
What the present invention relates to is a kind of combustion gas daily load combining prediction method, a kind of specifically new method of using multiple neural network model to carry out the forecast of combustion gas daily load and finally by the broad sense dynamic fuzzy neural network combustion gas being carried out combining prediction.
Background technology
The combustion gas Study on Forecast depends on characteristics such as periodicity that rock gas has, tendency, randomness.The factor that influences the combustion gas load is also a lot, for example, temperature, weather conditions, festivals or holidays, resident living level, habits and customs, energy policy and socio-economic development level etc., these influences make how accurately prediction is more challenging.
The combustion gas prediction has had certain development in China, but because China's natural gas is also not long service time as a kind of new forms of energy, can be also very not enough for data and the experience that prediction is used.
Traditional Forecasting Methodology has regression analysis, time series analysis method, neural network method, gray theory method, fuzzy logic method, methods such as combined prediction at present.The whole bag of tricks respectively has the characteristics of oneself.Have the principle maturation, use advantages such as easy, required historical data is few, workload is few such as time series method; Gray prediction has that the requirement load data is few, and modeling is fairly simple, and is very effective when data deficiency.
But these methods also have certain disadvantages and limitation.Such as time series method having relatively high expectations of pair data arranged, can not consider easily that weather condition etc. has the correlative factor of material impact to load, only is devoted to the match of data, to the shortcomings such as undertreatment of regularity.Gray prediction relatively is fit to have the prediction of the load of exponential increase rule, when load increasing speed slower, i.e. Dui Ying exponential function x=be
-atIn | a| hour, precision of prediction is higher, | when a| is big, the precision of prediction variation.
The combustion gas load forecasting method is a lot, but the applicable elements of each method is different with characteristics, and the combustion gas load had both had a kind of characteristics of method sometimes in each schedule periods, has the characteristics of another kind of method again, so just several method can be optimized combination, its prediction effect will be greatly enhanced.It can avoid single Forecasting Methodology to lose the defective of useful information, reduces randomness.Improve precision of prediction.
Summary of the invention
The object of the present invention is to provide a kind of forecast accuracy and precision height as a result, the combustion gas daily load combining prediction method that relative error is little based on the broad sense dynamic fuzzy neural network.
The object of the present invention is achieved like this:
(1) gathers the gas historical record data as the historical time sequence data by the city gas dispatch center;
(2) the historical time sequence data is carried out that abnormal data is judged and handle, abnormal data is judged and is comprised with processing: with reference to similar day data, carry out level and smooth, correction, normalization, composition historical load time series; Described similar day promptly be from its nearest have close weather, temperature, what day;
(3) utilizing generalized regression nerve networks, is that sample data is carried out difference processing and predicted by network to the historical load time series, approaches implicit prediction mapping relations according to sample data, obtains predicting the daily load value;
(4) utilize grey neural network, input historical load seasonal effect in time series one-accumulate is generated the input of data as network, and the one-accumulate of the corresponding prediction daily load of output generates data, training network, at last output valve is carried out once the tired contrary processing that generates that subtracts, obtain predicting the daily load value;
(5) predicted value that (3) (4) are obtained is as the input of broad sense dynamic fuzzy neural network, and data are divided into groups, and finally obtains predicting the daily load value.
The present invention can also comprise:
1, the gas historical record data of described collection is meant the urban gas daily load gas consumption;
Do not set fuzzy rule when 2, the broad sense dynamic fuzzy neural network that is used for combined prediction begins, system is by the on-line automatic generation of study and prune fuzzy rule, and parameter adaptive calculates, and and Structure Identification carry out simultaneously.
3, described generalized regression nerve networks, the various single Forecasting Methodologies of grey neural network based on difference processing need not to weather and week type carry out special consideration.
4, described packet form is meant: the combustion gas daily load data that the broad sense dynamic fuzzy neural network is adopted will be according to the situation grouping of festivals or holidays.
Of the present inventionly carry out combined prediction based on the broad sense dynamic fuzzy neural network, less demanding to data, need not to consider factors such as weather condition, can approach (combustion gas daily load) seasonal effect in time series Changing Pattern.
The back partition combustion gas dispatching system just that the present invention is directed to software platform in the gas ductwork dispatching system is carried out the part of prediction, propose a kind of new combustion gas daily load prediction method, utilized the intelligent adaptive combination forecasting method to obtain relation between combustion gas daily load historical data and the following daily load amount to be predicted.
At first adopted the single Forecasting Methodology of generalized regression nerve networks, grey neural network to predict respectively in the invention based on difference processing, then it is predicted the outcome and carry out combined prediction, obtain final intelligent adaptive combination forecasting method as the input of broad sense dynamic fuzzy neural network.
Principal feature of the present invention comprises:
1, adopt two kinds of single Forecasting Methodologies to predict respectively, and to festivals or holidays situation do not do special consideration;
2, at the higher combination forecasting method of the design accuracy that predicts the outcome of each single Forecasting Methodology;
3, use the broad sense dynamic fuzzy neural network and carry out combined prediction;
4, different packet forms is adopted in combined prediction result's check;
5, with the output of broad sense dynamic fuzzy neural network output, and utilize historical data to test, obtain relative error, and the relative error result is analyzed as whole prognoses system.
Characteristics of the present invention can also comprise:
1, the single Forecasting Methodology of described kind is: generalized regression nerve networks Forecasting Methodology, grey neural network Forecasting Methodology, and when utilizing generalized regression nerve networks to predict, (combustion gas daily load) time series is carried out difference processing;
2, the combination forecasting method that described accuracy is higher is meant: with the output of generalized regression nerve networks, each single Forecasting Methodology of the grey neural network input as combined prediction, and realize combined prediction by the broad sense dynamic fuzzy neural network;
3, described packet form is meant: the check data that the broad sense dynamic fuzzy neural network is adopted will be according to the situation grouping of festivals or holidays.And each is tested to single Forecasting Methodology and need not packet;
4, described error analysis is meant: adopt the broad sense dynamic fuzzy neural network to carry out relative error and various single Forecasting Methodology relative error that combined prediction produced and compare, and then estimate the quality of each Forecasting Methodology.
Principle of work of the present invention is: core of the present invention is design, realization and the check of combustion gas daily load combination forecasting method.This prognoses system is based on the generalized regression nerve networks of difference processing, the combustion gas daily load prediction system that each single Forecasting Methodology of grey neural network makes up by the broad sense dynamic fuzzy neural network.
Described combustion gas load forecasting method specifically may further comprise the steps:
1. consider that a kind of Forecasting Methodology can not reflect all data messages fully, determined self-adapting intelligent combination forecasting method based on the broad sense dynamic fuzzy neural network, see Fig. 1, and export as its input with the single Forecasting Methodology of generalized regression nerve networks, grey neural network based on difference processing.The data pretreatment process is seen Fig. 2.
2. set up generalized regression nerve networks output input relation based on difference processing.
(Generalized Regres sion Neural Network, GRNN) structure is seen Fig. 3 to generalized regression nerve networks.It is output as:
And
Wherein ω 2
i, ω 1
JiBe network weight coefficient, b1
iBe network threshold, they utilize gradient method by historical data, by changing step-length, increase the momentum term self-adaptation and determine y
qBe network output,
Be the network input value, i=1,2,, n, j=1,2,, m
Urban gas daily load has very strong randomness, and data are not steady comparatively speaking, and also can have certain error bring difficulty for prediction when concrete statistics.The present invention is by removing non-stationary to the seasonal effect in time series difference.
Through after the difference processing, can eliminate the section data error, improve the precision of data, help reducing the relative error of prediction, improve accuracy of predicting.
Concrete steps are:
(1) pretreated time series is carried out the new signal that difference obtains.χ
(1)(k)=χ
(0)(k)-χ
(0)(k-1) as the net training time sequence;
(2) set up and train GRNN;
When (3) predicting, the network output valve is the first order difference of prediction daily load amount, promptly obtains the daily load amount of predicting through reduction.
3. utilize grey neural network to predict
Gray system theory is combined with generalized regression nerve networks, the one-accumulate of raw data in the gray theory is generated data as generalized regression network training sample data, because cumulative data has dull increase trend, makes the non-linear excitation function in the neural network be easy to approach.
Concrete steps are:
(1) with pretreated time series { χ
(0)(1), χ
(0)(2), χ
(0)(n) } carry out one-accumulate and generate processing, obtain χ
(1)={ χ
(1)(1), χ
(1)(2), χ
(1)(n) }, as the net training time sequence,
Wherein,
K=1,2,, n
(2) set up and train GRNN;
When (3) predicting, with its output valve
Do the tired reduction that subtracts:
Obtain predicting the daily load value.
4. utilize the broad sense dynamic fuzzy neural network to carry out combined prediction
Broad sense dynamic fuzzy neural network (Generalized dynamic fuzzy neural network, GD-FNN) be to utilize a kind of learning algorithm based on expansion radial base neural net, online self-organization, its function is equivalent to the Takagi-Sufeno-Kang fuzzy system, this learning algorithm is when beginning, system does not have fuzzy rule, by learning online generation and pruning fuzzy rule, the adjustment of its parameter and Structure Identification are carried out simultaneously.
The learning algorithm flow process of GD-FNN is seen Fig. 4.K wherein
dAnd k
eBe pre-set threshold, η
lBe the importance of l bar rule, k
ErrIt is the predefine parameter of initialization system.
Description of drawings
Fig. 1 is the combination forecasting method block diagram;
Fig. 2 is data pretreatment process figure;
Fig. 3 is generalized regression nerve networks (GR-FNN) structural drawing;
Fig. 4 is the learning algorithm process flow diagram of broad sense dynamic fuzzy neural network (GD-FNN);
Fig. 5 is 1～7 day approximating curve of combined prediction;
Fig. 6 is 8～32 days approximating curves of combined prediction;
Fig. 7 is 33～46 days approximating curves of combined prediction;
Fig. 8 is 47～60 days approximating curves of combined prediction;
Fig. 9 is the relative error value based on the generalized regression nerve networks prediction of difference processing;
Figure 10 is the relative error value of grey neural network prediction;
Figure 11 is 1～7 day relative error value (regular number is 4) of combined prediction;
Figure 12 is 16 consequent parameter values of 1～7 day of combined prediction (regular number is 4);
Figure 13 is 8～32 days relative error values (regular number is 8) of combined prediction;
Figure 14 is 32 consequent parameter values of 8～32 days of combined prediction (regular number is 8);
Figure 15 is 33～46 days relative error values (regular number is 7) of combined prediction;
Figure 16 is 28 consequent parameter values of 33～46 days of combined prediction (regular number is 7);
Figure 17 is 47～60 days relative error values (regular number is 5) of combined prediction;
Figure 18 is 20 consequent parameter values of 47～60 days of combined prediction (regular number is 5).
Embodiment
For example the present invention is done description in more detail below in conjunction with accompanying drawing:
1. utilize generalized regression nerve networks, grey neural network that the combustion gas daily load is predicted respectively based on difference processing.
General step is:
(1) data pre-service.Promptly at first before carrying out neural metwork training, carry out the abnormal data judgement and handle (similar day data of reference, similar day promptly be from its nearest have close weather, temperature, what day etc.), then data are carried out normalized, the present invention arrives (0,1) interval with data normalization;
(2) carry out the selection of network input and output node.Consider combustion gas day the gas consumption periodicity and the tendency that have, select day combustion gas load value before the previous day and seven days of prediction day as the input node here respectively, imports nodes for totally two.Select prediction day combustion gas load value as output node;
(3) selection of training sample, observation sample.Experimental data is divided into training sample and test sample book.The present invention utilizes the historical record data of urban gas daily load to predict as experimental data, 74 data instances of combustion gas daily load with the 2008.4.17-2008.6.29 of Harbin City, here select preceding 14 groups of data as training sample, and the 60 groups of data in back are as test sample book;
(4) set up neural network structure, initial parameter is set, calculate relative error and each parameter is made amendment;
(5) network performance that utilizes evaluating data group (test sample book) inspection to train.Through check, the network performance of being set up is good.
Generalized regression nerve networks combustion gas daily load prediction relative error based on difference processing sees Table 1, and grey neural network prediction relative error sees Table 2.
2. utilize the broad sense dynamic fuzzy neural network that the combustion gas daily load is carried out combined prediction.
Based on the predicted value of the generalized regression nerve networks of difference processing, the grey neural network input quantity as the broad sense dynamic fuzzy neural network, its output quantity is the combined prediction value, and GD-FNN is trained and verification, carries out result's analysis.Process is as follows:
At distributing festivals or holidays to the influence of gas consumption, 2008.4.17-2008.6.29 middle 5.1 long holidays of existence and 6.1 Children's Day, but 6.1 Children's Day are little to the influence of combustion gas daily load prediction, so 60 groups of test sample book data are divided into four groups: 1～7,8～32,33～46,47～60.Set up four independently GD-FNN respectively, draw the GD-FNN of four periods.
Programme with the M language in MATLAB, operation result is as follows:
To 1～7 data (being 2008.5.1-2008.5.7), emulation experiment parameter preset: ε
Min=0.5, ε
Max=0.8, e
Max=0.5, e
Min=0.08, k
Mf=0.1, k
s=0.95, k
Err=0.003
It is 4 that training finishes back generation rule number, simulation time 0.1082s.The combined prediction approximating curve is seen Fig. 4, and the relative error value sees Table 3, and the application of rules degree is: 0.0329 .0002 .0067 .3683,16 of consequent parameters see Table 4.
In like manner, to 8～32 data (being 2008.5.8-2008.6.1), can draw the predicted value based on GD-FNN, its parameter preset is provided with as follows: ε
Min=0.5, ε
Max=0.8, e
Max=02, e
Min=0.07, k
Mf=0.11, k
s=0.9, k
Err=0.002
After training finished, it was 8 that GD-FNN generates number of fuzzy rules, and program runtime is 0.1107s.The application of rules degree is: 0.1962,0.0064,0.0042,0.0003,0.0091,0.0803,0.2773,0.4263.Relative error sees Table 5, and 32 of consequent parameters see Table 6.Fig. 5 is its approximating curve.
To 33～46 data (being 2008.6.2-2008.6.15), parameter preset is provided with as follows: ε
Min=0.5, ε
Max=0.8, e
Max=0.02, e
Min=0.009, k
Mf=0.015, k
s=0.9, k
Eer=0.001.
After training finished, generating number of fuzzy rules was 7, and program runtime is 0.1107s.The application of rules degree is: 0.0013,0.0147,0.0159,0.0252,0.8916,0.0050,0.0462.The combined prediction relative error sees Table 7, and 28 of consequent parameters see Table 8.Fig. 6 is its approximating curve.
To 47～60 data (being 2008.6.16-2008.6.29), parameter preset is provided with as follows: ε
Min=0.5, ε
Max=0.8, e
Max=02, e
Min=0.007, k
Mf=0.06, k
s=0.9, k
Eer=0.001.After training finished, generating number of fuzzy rules was 5, and program runtime is 0.0952s.The application of rules degree is: 0.0091,0.0005,0.5039,0.2392,0.2473.Relative error sees Table 9, and 20 of consequent parameters see Table 10.Fig. 7 is its approximating curve.
Be the Forecasting Methodology that patent of the present invention is set up by the constructed combination forecasting method of above step, by test (check) data and final approximating curve as can be seen this method have well and predict the outcome.
The basic thought of combustion gas daily load prediction method is to take all factors into consideration to use the gas rule, simultaneously according to gas consumption reflected in historical day meteorology, festivals or holidays, week etc. information and and following day gas consumption between the characteristics that concern, thereby select suitable combination forecasting method that a following day gas consumption is carried out scientific forecasting.This method has been given full play to generalized regression nerve networks fitting precision height, need not set up the advantage of function model, and data are through difference processing, and it is non-stationary to overcome data, advantage of high precision; Grey neural network can also can provide the advantage of higher forecast precision when data deficiency; The online generation of broad sense dynamic fuzzy neural network and pruning fuzzy rule and the adjustment of its parameter and the plurality of advantages that Structure Identification is carried out simultaneously.
Combustion gas daily load prediction method provided by the invention can quick and precisely reflect the variation of combustion gas daily load and need not the influence in extra consideration meteorology, week etc., can online self-organized learning, fast, accurately following combustion gas load is predicted.Because therefore the vital role of combustion gas daily load prediction aspect combustion gas scheduling, gas distributing system construction carry out effective the combination with the present invention with gas distribution system, gas network, form combustion gas load prediction software.To increase precision of prediction greatly, improve the efficiency of management.Thereby bring great economic benefit, reduce the wasting of resources.
The present invention has provided the combustion gas daily load prediction data that are fit to based on combustion gas day gas consumption intelligent Combination Forecasting, and compared with prior art, the present invention has following technique effect:
(1) adopt combination forecasting method, precision of prediction is higher;
(2) adopt the broad sense dynamic fuzzy neural network to carry out combined prediction, it can online self-organized learning, has improved operation efficiency, precision greatly, dimension disaster can not occur again;
(3) adopt and to predict respectively and it is predicted the outcome based on generalized regression nerve networks, the grey neural network of difference processing and import as combined prediction, can give full play to the advantage of each single Forecasting Methodology, overcome the shortcoming and defect of single Forecasting Methodology, guarantee not drop-out again to the full extent;
(4) only need provide combustion gas daily load historical data, need not to consider factors such as meteorology, week, the information that not only can more effective comprehensive embodiment historical data be provided has also reduced undulatory property, the randomness of data, improved the validity of prediction, used also more extensive.
Claims (4)
1. combustion gas daily load combining prediction method based on the broad sense dynamic fuzzy neural network is characterized in that specifically may further comprise the steps:
(1) gathers the gas historical record data as the historical time sequence data by the city gas dispatch center;
(2) the historical time sequence data is carried out that abnormal data is judged and handle, abnormal data is judged and is comprised with processing: with reference to similar day data, carry out level and smooth, correction, normalization, composition historical load time series; What day be meant from its nearest having close weather, temperature, date in described similar day;
(3) utilizing generalized regression nerve networks, is that sample data is carried out difference processing and predicted by network to the historical load time series, approaches implicit prediction mapping relations according to sample data, obtains predicting the daily load value;
(4) utilize grey neural network, input historical load seasonal effect in time series one-accumulate is generated the input of data as network, and the one-accumulate of the corresponding prediction daily load of output generates data, training network, at last output valve is carried out once the tired contrary processing that generates that subtracts, obtain predicting the daily load value;
(5) predicted value that step (3), (4) are obtained is as the input of broad sense dynamic fuzzy neural network, and data are divided into groups, and finally obtains predicting the daily load value.
2. the combustion gas daily load combining prediction method based on the broad sense dynamic fuzzy neural network according to claim 1, it is characterized in that: the gas historical record data of described collection is meant the urban gas daily load gas consumption.
3. the combustion gas daily load combining prediction method based on the broad sense dynamic fuzzy neural network according to claim 2, when beginning, the broad sense dynamic fuzzy neural network that it is characterized in that being used for combined prediction do not set fuzzy rule, system is by the on-line automatic generation of study and prune fuzzy rule, parameter adaptive calculates, and carries out simultaneously with Structure Identification.
4. the combustion gas daily load combining prediction method based on the broad sense dynamic fuzzy neural network according to claim 3 is characterized in that described packet form is meant: the situation grouping of the broad sense dynamic fuzzy neural network is adopted combustion gas daily load data based festivals or holidays.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010561217 CN102073785B (en) | 2010-11-26 | 2010-11-26 | Daily gas load combination prediction method based on generalized dynamic fuzzy neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010561217 CN102073785B (en) | 2010-11-26 | 2010-11-26 | Daily gas load combination prediction method based on generalized dynamic fuzzy neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102073785A true CN102073785A (en) | 2011-05-25 |
| CN102073785B CN102073785B (en) | 2013-02-27 |
Family
ID=44032323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201010561217 Expired - Fee Related CN102073785B (en) | 2010-11-26 | 2010-11-26 | Daily gas load combination prediction method based on generalized dynamic fuzzy neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102073785B (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102354376A (en) * | 2011-06-30 | 2012-02-15 | 内蒙古电力勘测设计院 | Method for supplementing and correcting wind measurement data |
| CN102968669A (en) * | 2011-08-31 | 2013-03-13 | 富士通株式会社 | Method and device for predicating load |
| CN103020725A (en) * | 2011-09-20 | 2013-04-03 | 株式会社日立制作所 | Electric power demand forecast system and method |
| CN103258243A (en) * | 2013-04-27 | 2013-08-21 | 杭州电子科技大学 | Tube explosion predicting method based on grey neural network |
| CN103853894A (en) * | 2014-03-26 | 2014-06-11 | 上海航天电子通讯设备研究所 | RBF (radial basis function) algorithm-based bait bullet delay casting time precision calculating method |
| CN104008164A (en) * | 2014-05-29 | 2014-08-27 | 华东师范大学 | Generalized regression neural network based short-term diarrhea multi-step prediction method |
| CN104036362A (en) * | 2014-06-23 | 2014-09-10 | 国家电网公司 | Rapid detection method of transformer power load abnormal data |
| CN104123591A (en) * | 2014-07-14 | 2014-10-29 | 上海师范大学 | Smart grid classification and fuzzy neural network based natural gas load prediction method |
| CN104182854A (en) * | 2014-09-18 | 2014-12-03 | 北京清晖翔科技有限公司 | Mixed energy consumption measuring method for enterprise energy management system |
| CN105426956A (en) * | 2015-11-06 | 2016-03-23 | 国家电网公司 | Ultra-short-period photovoltaic prediction method |
| CN105894111A (en) * | 2016-03-30 | 2016-08-24 | 天鸿泰（北京）科技有限公司 | Energy consumption prediction method and device based on complementary fuzzy neural network |
| CN105894113A (en) * | 2016-03-31 | 2016-08-24 | 中国石油天然气股份有限公司规划总院 | Natural gas short-period demand prediction method |
| CN105976057A (en) * | 2016-05-04 | 2016-09-28 | 三峡大学 | Medium and long-term load forecasting method based on improved grey theory |
| CN103258243B (en) * | 2013-04-27 | 2016-11-30 | 杭州电子科技大学 | Tube explosion prediction method based on grey neural network |
| CN103729528B (en) * | 2012-10-15 | 2017-06-16 | 富士通株式会社 | The apparatus and method processed sequence |
| CN108364060A (en) * | 2018-01-31 | 2018-08-03 | 天津市市政工程设计研究院 | A kind of classification design method of highway |
| CN109997164A (en) * | 2016-10-21 | 2019-07-09 | 瑞士再保险有限公司 | Measurement based on interarrival time and the measuring system and method for the imminent natural disaster sexual behavior part of prediction are provided |
| CN110633846A (en) * | 2019-09-02 | 2019-12-31 | 北京市燃气集团有限责任公司 | Gas load prediction method and device |
| CN110852496A (en) * | 2019-10-29 | 2020-02-28 | 同济大学 | Natural gas load prediction method based on LSTM recurrent neural network |
| CN110866672A (en) * | 2019-10-10 | 2020-03-06 | 重庆金融资产交易所有限责任公司 | Data processing method, device, terminal and medium |
| CN111144649A (en) * | 2019-12-26 | 2020-05-12 | 哈尔滨工程大学 | Combined prediction method based on information fusion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080208487A1 (en) * | 2007-02-23 | 2008-08-28 | General Electric Company | System and method for equipment remaining life estimation |
| CN101482941A (en) * | 2008-01-09 | 2009-07-15 | 新奥(廊坊)燃气技术研究发展有限公司 | Urban gas daily load prediction method |
| CN101788692A (en) * | 2009-11-24 | 2010-07-28 | 西北电网有限公司 | Short-time wind speed forecasting method based on neural network |
-
2010
- 2010-11-26 CN CN 201010561217 patent/CN102073785B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080208487A1 (en) * | 2007-02-23 | 2008-08-28 | General Electric Company | System and method for equipment remaining life estimation |
| CN101482941A (en) * | 2008-01-09 | 2009-07-15 | 新奥(廊坊)燃气技术研究发展有限公司 | Urban gas daily load prediction method |
| CN101788692A (en) * | 2009-11-24 | 2010-07-28 | 西北电网有限公司 | Short-time wind speed forecasting method based on neural network |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102354376B (en) * | 2011-06-30 | 2014-04-02 | 内蒙古电力勘测设计院 | Method for supplementing and correcting wind measurement data |
| CN102354376A (en) * | 2011-06-30 | 2012-02-15 | 内蒙古电力勘测设计院 | Method for supplementing and correcting wind measurement data |
| CN102968669A (en) * | 2011-08-31 | 2013-03-13 | 富士通株式会社 | Method and device for predicating load |
| CN102968669B (en) * | 2011-08-31 | 2015-11-25 | 富士通株式会社 | To the method and apparatus that load is predicted |
| CN103020725B (en) * | 2011-09-20 | 2016-10-05 | 株式会社日立制作所 | Electric power demand forecast system and method |
| CN103020725A (en) * | 2011-09-20 | 2013-04-03 | 株式会社日立制作所 | Electric power demand forecast system and method |
| CN103729528B (en) * | 2012-10-15 | 2017-06-16 | 富士通株式会社 | The apparatus and method processed sequence |
| CN103258243A (en) * | 2013-04-27 | 2013-08-21 | 杭州电子科技大学 | Tube explosion predicting method based on grey neural network |
| CN103258243B (en) * | 2013-04-27 | 2016-11-30 | 杭州电子科技大学 | Tube explosion prediction method based on grey neural network |
| CN103853894A (en) * | 2014-03-26 | 2014-06-11 | 上海航天电子通讯设备研究所 | RBF (radial basis function) algorithm-based bait bullet delay casting time precision calculating method |
| CN103853894B (en) * | 2014-03-26 | 2017-04-12 | 上海航天电子通讯设备研究所 | RBF (radial basis function) algorithm-based bait bullet delay casting time precision calculating method |
| CN104008164A (en) * | 2014-05-29 | 2014-08-27 | 华东师范大学 | Generalized regression neural network based short-term diarrhea multi-step prediction method |
| CN104036362A (en) * | 2014-06-23 | 2014-09-10 | 国家电网公司 | Rapid detection method of transformer power load abnormal data |
| CN104123591A (en) * | 2014-07-14 | 2014-10-29 | 上海师范大学 | Smart grid classification and fuzzy neural network based natural gas load prediction method |
| CN104182854A (en) * | 2014-09-18 | 2014-12-03 | 北京清晖翔科技有限公司 | Mixed energy consumption measuring method for enterprise energy management system |
| CN105426956B (en) * | 2015-11-06 | 2019-01-22 | 国家电网公司 | A kind of ultra-short term photovoltaic prediction technique |
| CN105426956A (en) * | 2015-11-06 | 2016-03-23 | 国家电网公司 | Ultra-short-period photovoltaic prediction method |
| CN105894111A (en) * | 2016-03-30 | 2016-08-24 | 天鸿泰（北京）科技有限公司 | Energy consumption prediction method and device based on complementary fuzzy neural network |
| CN105894111B (en) * | 2016-03-30 | 2020-02-04 | 天鸿泰（北京）科技有限公司 | Energy consumption prediction method and device based on complementary fuzzy neural network |
| CN105894113A (en) * | 2016-03-31 | 2016-08-24 | 中国石油天然气股份有限公司规划总院 | Natural gas short-period demand prediction method |
| CN105976057A (en) * | 2016-05-04 | 2016-09-28 | 三峡大学 | Medium and long-term load forecasting method based on improved grey theory |
| CN105976057B (en) * | 2016-05-04 | 2021-07-20 | 三峡大学 | Medium-and-long-term load prediction method based on improved grey theory |
| CN109997164A (en) * | 2016-10-21 | 2019-07-09 | 瑞士再保险有限公司 | Measurement based on interarrival time and the measuring system and method for the imminent natural disaster sexual behavior part of prediction are provided |
| CN108364060A (en) * | 2018-01-31 | 2018-08-03 | 天津市市政工程设计研究院 | A kind of classification design method of highway |
| CN110633846A (en) * | 2019-09-02 | 2019-12-31 | 北京市燃气集团有限责任公司 | Gas load prediction method and device |
| CN110866672A (en) * | 2019-10-10 | 2020-03-06 | 重庆金融资产交易所有限责任公司 | Data processing method, device, terminal and medium |
| CN110852496A (en) * | 2019-10-29 | 2020-02-28 | 同济大学 | Natural gas load prediction method based on LSTM recurrent neural network |
| CN111144649A (en) * | 2019-12-26 | 2020-05-12 | 哈尔滨工程大学 | Combined prediction method based on information fusion |
| CN111144649B (en) * | 2019-12-26 | 2023-04-07 | 哈尔滨工程大学 | Urban gas daily load combined prediction method based on information fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102073785B (en) | 2013-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102073785B (en) | Daily gas load combination prediction method based on generalized dynamic fuzzy neural network | |
| CN102567391B (en) | Method and device for building classification forecasting mixed model | |
| CN103295075B (en) | A kind of ultra-short term load forecast and method for early warning | |
| CN110222882A (en) | A kind of prediction technique and device of electric system Mid-long Term Load | |
| CN102129511B (en) | System for forecasting short-term wind speed of wind power station based on MATLAB | |
| CN103310286A (en) | Product order prediction method and device with time series characteristics | |
| CN104715292A (en) | City short-term water consumption prediction method based on least square support vector machine model | |
| Zhu et al. | Short‐Term Electricity Consumption Forecasting Based on the EMD‐Fbprophet‐LSTM Method | |
| Huang et al. | Genetic Algorithms for Optimization of Resource Allocation in Large Scale Construction Project Management. | |
| CN106096771A (en) | The parameter selection method of support vector machine and system and load forecasting method and system | |
| Jakobsen et al. | A standardized Approach to Multi-criteria Assessment of CCS Chains | |
| CN105956247A (en) | Open, fair and impartial scheduling model building method based on Gini coefficient | |
| CN107239850A (en) | A kind of long-medium term power load forecasting method based on system dynamics model | |
| CN103345663A (en) | Combinatorial optimization method of electric power system set considering creep speed constraints | |
| Shtunder et al. | Sustainable development of the economy in the conditions of the energy crisis. | |
| Faber et al. | Economy-environment interactions in the long-run: a neo-Austrian approach | |
| CN104182854A (en) | Mixed energy consumption measuring method for enterprise energy management system | |
| Arida et al. | building energy modeling using artificial Neural Networks | |
| Makri et al. | Short term net imbalance volume forecasting through machine and deep learning: a UK case study | |
| Jiao et al. | Multi-step traffic flow prediction method based on the Conv1D+ LSTM | |
| Yang et al. | Risk prediction of city distribution engineering based on BP | |
| Ru-tao et al. | A Machine learning pipeline generation approach for data analysis | |
| Yan | A building management evaluation method based on BP neural networks | |
| Song et al. | 3Es System Optimization under Uncertainty Using Hybrid Intelligent Algorithm: A Fuzzy Chance‐Constrained Programming Model | |
| Guojun et al. | Prediction of ecotourism population based on exponential smoothing and ARIMA mixed 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130227 Termination date: 20191126 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |