CN108376294A - A kind of heat load prediction method of energy supply feedback and meteorologic factor - Google Patents
A kind of heat load prediction method of energy supply feedback and meteorologic factor Download PDFInfo
- Publication number
- CN108376294A CN108376294A CN201810048118.XA CN201810048118A CN108376294A CN 108376294 A CN108376294 A CN 108376294A CN 201810048118 A CN201810048118 A CN 201810048118A CN 108376294 A CN108376294 A CN 108376294A
- Authority
- CN
- China
- Prior art keywords
- temperature
- layer
- return water
- input
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 23
- 230000001537 neural effect Effects 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 18
- 239000011159 matrix material Substances 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 3
- 238000000205 computational method Methods 0.000 claims description 2
- 238000010438 heat treatment Methods 0.000 abstract description 11
- 238000012546 transfer Methods 0.000 abstract description 3
- 238000013528 artificial neural network Methods 0.000 description 19
- 230000005540 biological transmission Effects 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 5
- 238000013507 mapping Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000004393 prognosis Methods 0.000 description 2
- 230000005619 thermoelectricity Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 239000000686 essence Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 229910052709 silver Inorganic materials 0.000 description 1
- 239000004332 silver Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention is applied to multiple-energy-source joint network heating power and supplies field, the heat load prediction method for energizing feedback and meteorologic factor for determining the demand supplied for heating power load side in heating power supply, and this method includes:Step 1, the influence by wind speed to outdoor temperature is converted into corresponding temperature variation, equivalent outdoor temperature in the case of calculating is calm;Step 2, according to historical data first-level pipeline network return water temperature, first-level pipeline network return water flow velocity, secondary pipe network return water temperature, secondary pipe network intake flow velocity, a pipe network inflow temperature and equivalent outdoor temperature as input, inflow temperature once to net is trained RBF neural as output;Step 3, the RBF neural after training predicts primary net inflow temperature.Solve the problems, such as that thermic load complex model is difficult to set up, and the lag delay for also avoiding heat transfer brings the problem of being difficult to set up the thermic load equilibrium of supply and demand.
Description
Technical field
The present invention be applied to multiple-energy-source joint network heating power supply field, for determine heating power supply in load side for
The heat load prediction method of the energy supply feedback and meteorologic factor of the demand of heating power supply.
Background technology
The problem of due to energy shortages, environmental pollution it is increasingly serious, how clean and effective utilize natural resources
Have become the Important Problems of current scientific research.Therefore, energy internet, multiple-energy-source, which are combined, has become each energy industry development
Inexorable trend, as the element task of multiple-energy-source scheduling, the accuracy of load prediction can largely influence traffic control
It can be smoothed out.
Its internal heat energy transmittance process of heating power network is extremely complex, is difficult to through specific mathematical expression in non-linear relation
Formula illustrates the relationship between each physical quantity in diabatic process, therefore most predictions is established between each parameter in conjunction with historical data
Relationship.The method of heat load prediction, which can be divided into, to be predicted by environment weather parameter prediction with by return water temperature.Pass through environment
Meteorologic parameter (such as temperature, humidity, intensity of sunshine, wind speed and direction) prediction using artificial intelligence and data mining, it is non-linear from
Regression model, adaptive Kalman filter, adaptive line time series models, and obtain preferable precision of prediction.Such side
Method can obtain preferable prediction effect, but have higher want to comprehensive and weather prognosis data the accuracy of Consideration
It asks, is difficult to provide preferable prediction result in the case where data are limited or weather prognosis precision is not high;It is pre- by return water temperature
Survey in fact, the runing adjustment strategy of central heating system operational management is achieved by a variety of factors, the different periods by
It in the difference of environmental parameter, adjusts strategy and is different, very big influence caused on thermic load, between the heating parameter of thermic load
Relationship be continually changing, therefore, the application of this method is limited.
In addition, different-energy form transmission characteristic is different, compared to power transmission, thermodynamic transport has apparent lag
Property, it is difficult to use identical two kinds of energy form United Dispatchings of time scale pair.
Invention content
Technical problem to be solved by the present invention lies in provide a kind of heat load prediction side of energy supply feedback and meteorologic factor
Method solves the problems, such as that thermic load complex model is difficult to set up, and the lag delay for also avoiding heat transfer brings and is difficult to set up
The problem of thermic load equilibrium of supply and demand.
The invention is realized in this way a kind of heat load prediction method of energy supply feedback and meteorologic factor, this method include:
Step 1, the influence by wind speed to outdoor temperature is converted into corresponding temperature variation, equivalent in the case of calculating is calm
Outdoor temperature;
Step 2, using heat supply return water temperature as energy supply feedback data, according to historical data first-level pipeline network return water temperature, one
Grade pipe network return water flow velocity, secondary pipe network return water temperature, secondary pipe network water inlet flow velocity, a pipe network inflow temperature and equivalent outdoor
As input, the inflow temperature once to net is trained RBF neural as output temperature;
Step 3, the RBF neural after training predicts primary net inflow temperature.
Further, the weights of RBF neural input layer to hidden layer are fixed as 1.
Further, the computational methods of equivalent outdoor temperature are:Work as v<When 2m/s, TF=T;
As 2m/s≤v<When 5m/s, TF=T- (v-2);
As 5m/s≤v<When 10m/s, TF=T- (v-5) × [3+0.2 × (v-5)] -3;
As v >=10m/s, TF=T-23;
Wherein, v is wind speed, TFFor equivalent outdoor temperature, T is outdoor actual temperature.
Further, in step 2 RBF neural be trained including:
The initial data of input is normalized input layer;
Hidden layer determines number, center and the width of hidden layer neuron using the method for randomly selecting network center, with
And hidden layer is to the connection weight between input layer, and export output matrix;
Input layer into linear transformation is done, exports output matrix by linear weighted function and to input variable.
Compared with prior art, the present invention advantageous effect is:The present invention is used return water temperature and environment weather parameter
Accurate real-time prediction result will be obtained to load progress by considering.It is different to solve different energy sources scheduling time scale
Problem, institute's extracting method of the present invention data acquisition with prediction be in synchronization heat demand can be predicted in real time and
Without considering the decaying and delay of thermodynamic transport, and then the scheduling of heating power network and the scheduling of electric power networks are placed in same time
Under scale, multiple-energy-source joint United Dispatching is realized.It is difficult to unifiedly calculate and individually examine to solve different meteorologic parameters
Consider and predict error caused by meteorologic factor or return water temperature, it is proposed that consider wind speed, outdoor temperature equivalent outdoor temperature
Degree, consider the return water temperature of heat supply network, equivalent environment temperature, heat supply network water inlet flow velocity to inflow temperature predicted (control) with
Achieve the purpose that heat load prediction.
Inflow temperature is determined by the temperature of return water.Solve the problems, such as that thermic load complex model is difficult to set up, and
The lag delay for also avoiding heat transfer brings the problem of being difficult to set up the thermic load equilibrium of supply and demand.
Description of the drawings
Fig. 1 is method flow diagram provided in an embodiment of the present invention;
Fig. 2 is RBF neural schematic diagram calculation provided in an embodiment of the present invention;
Fig. 3 is REF neural metwork trainings flow chart provided in an embodiment of the present invention.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
The present invention is divided into two stages, and the first stage is to calculate the equivalent outdoor temperature stage, by wind speed to outdoor temperature
Influence is converted into corresponding temperature variation, so calculate its it is calm in the case of equivalent outdoor temperature;Second stage be combine into
Water flow velocity, return water temperature and equivalent outdoor temperature predict inflow temperature, in conjunction with Various types of data historical data with to REF
Neural network is trained, and is finally predicted inflow temperature according to available data.
Consider the outdoor equivalent temperature of outdoor environment first referring to Fig. 1
Experiments have shown that wind speed, outdoor temperature are the principal element being had an impact to the demand of thermic load, in order to will be a variety of
Factor is unifiedly calculated, the present invention propose it is a kind of by the outdoor environment for actually having wind by the way that wind speed to be converted to the change of temperature
Change, and then is converted into calm equivalent outdoor temperature.
First, wind speed conversion, which is equivalent to outdoor temperature, reduces the equivalent temperature value of actual temperature.Equivalent temperature value is by building
Build the outdoor temperature variable quantity that object wind speed is quantified as the influence for building temperature in building.This is an equivalent outdoor temperature,
Rather than actual outside air temperature.The expression formula of comprehensive outdoor temperature is:
Work as v<When 2m/s, TF=T
As 2m/s≤v<When 5m/s, TF=T- (v-2)
As 5m/s≤v<When 10m/s, TF=T- (v-5) × [3+0.2 × (v-5)] -3
As v >=10m/s, TF=T-23
Influence of the wind speed for building load is considered according to the calculated comprehensive outdoor temperature of above formula.Due to wind
It is little to the influence for building load, and in the time scale of dynamic load variation, the variation of wind direction will not be especially bright
It is aobvious, it is believed that the influence of wind direction can approximation ignore.
According to the calculated outdoor combined air temperature of above-mentioned formula, influence of the wind speed to built-loading is considered.Due to wind direction pair
Heat load influences less, and changes less in thermic load dynamic change time scale, therefore the influence of wind direction can neglect
Slightly.
The neural network prediction of return water parameter and equivalent temperature
Therrmodynamic system and electrical power system transmission property difference are larger, be difficult to establishing accurate therrmodynamic system model by
The discrepant system call problem of two kinds of thermoelectricity is placed under same time scale.In addition, there is one in itself between thermoelectricity load
It is fixed that interaction relationship, the weather conditions such as temperature, humidity, wind speed are even more further to increase to build an important factor for influencing load
The difficulty of vertical accurate model.
In the constraint of " electricity determining by heat ", the determination that heat is contributed is assigned important meaning for the energy of co-generation unit
Justice, but the model of thermic load by factors influenced it is therein interaction it is sufficiently complex,
The present invention, which uses, combines REF neural networks, predicts inflow temperature.
RBF neural is divided into input layer, output layer and hidden layer, the transmission of information be from input layer to hidden layer, then
From hidden layer to output layer.Wherein, between input layer and hidden layer it is direct mapping relations, hidden layer to output layer is then linear
Weighting is summed again.The mode of intelligence transmission of this direct mapping and weighted linear sum greatly accelerates the arithmetic speed of RBF networks,
And it avoids and is absorbed in local minimum.Hidden layer unit number can automatically adjust as needed in RBF neural, solve
Hidden layer unit number is difficult to determining problem;The weights of input layer to hidden layer are fixed as 1, eliminate initial weight determination and power
It is worth the trouble adjusted.
Referring to Fig. 2 and Fig. 3, used RBF neural calculation process is:
1 determines training sample set;
2 initialization weights, learning rate and smooth silver;
3 calculate output error, adjust weights;
4 take next training sample;
5 neural network forward operations;
6 errors are superimposed;
7 feedbacks calculate;
8 judge that all sample calculating terminate, if otherwise return to step 4;If then carrying out in next step;
9 correct connection weight;
10 learning rates, smoothing factor adjustment;
11 judge whether to be more than maximum iteration, if otherwise return to step 3, if algorithm is received in stipulated number
It holds back.
RBF neural of the present invention can be divided into three parts from structure, and the input layer of data processing, forecast period are hidden
The output layer of layer, output data.
The input layer of data processing
To reduce influence of the measurement error to prediction result, this patent is first normalized data.In RBF god
Through in network, there is no the weight matrixs of input layer to hidden layer, input layer to be responsible for input data and be transferred to neural network by outside
Inside, and input data is normalized and is selected.
The initial data of input is converted into [- 1 ,+1] section first;Then it is converted into initial data in output layer.Specifically
Calculation formula is as follows:
Wherein, x is initial data, xmaxFor maximum value in initial data, xminFor initial data minimum value, y is normalization
Data afterwards.
The hidden layer of forecast period
The topological structure of RBF neural has a huge impact its network performance on very big depth, hidden layer god
The generalization ability that neural network is not only affected through first number has an effect on the complexity of neural network.Simultaneously because each implicit
The center position and width of layer neuron determine that respective part mapping is responded, and neural network can reflect sample number
According to spatially different demarcation, mapping result will be had a huge impact when selecting inappropriate center position,
That is neural metwork training result is very sensitive to the selection of center position.
The width of RBF neural hidden layer is also a key factor for influencing its predictive ability of classifying, central point
Width determine the sphere of action hidden layer that input variable is responded to the weights between input layer be realize whole network it is final
One key variables of mapping result.Therefore it is as the difficult point for establishing neural network:Determine the correlation of hidden layer neuron
Parameter (the i.e. number of hidden layer neuron, center and width and hidden layer to the connection weight between input layer.
The present invention randomly chooses or selects whole works from input sample using the method for randomly selecting network center
For the center of the basic function of hidden layer neuron, so that it is determined that the center position of neural network, after determining center position,
The position and number of central point will all immobilize.The width (variance) of neural network is calculated according to the conventional method.When
After center position and width all determine, the output matrix of hidden layer is exactly known, by solving system of linear equations
Method determine the connection weight of neural network.Here it is the general steps randomly selected.
The output layer of output data
For output layer input matrix as simple linear transformation, by linear weighted function and to input variable into defeated
Go out.RBF neural output layer selection function is set as F, i.e.,:
Wherein, c is the center of basic function, | | | | it is Euclid norm, ωiFor weight vector,
For the set of N number of radial basis function, entire formula, that is, function F radial basis functionCarry out linear combination to approach.Central diameter of the present invention
Gaussian kernel function is used to base kernel function
The result of prediction normalize, the actual value of prediction data is exported.Calculation formula is as follows:
The RBF neural built up is trained, to determine one RBF nerve net of Weight Training of hidden layer
Network actually adjusts the weight of network and biases the two parameters, and the training process of neural network is divided into two parts:
Fl transmission, successively awave transmission output valve;
Reverse feedback, reversely successively adjusts weight and biasing;
1) fl transmission
Before training network, need random initializtion weight and biasing, to each weight take one of [- 1,1] it is random
Real number, each biasing take a random real number of [0,1], then begin to carry out fl transmission.The training of neural network is
It is completed by successive ignition, iteration will use all records of training set each time, and train network to only use each time
One record, is described as follows:
The output valve of input layer is set first, sets the number of attribute as 200, then the nerve of input layer is just arranged in we
Unit number is 200, the node N of input layeriFor the attribute value x in record i-th dimensioni.To the easy to operate of input layer, later
Every layer just more complex, and the input value of other each layers in addition to input layer is that last layer input value is added by weight accumulation result value
Biasing, the output valve etc. of each node are converted with the input value of the node.
The output layer calculating process of fl transmission, formula are as follows:
Output valve will be calculated in the way of figure as above to each of hidden layer and output layer node, to biography before completing
The process broadcast.
2) reverse feedback
Reverse feedback is since last layer i.e. output layer, and the purpose that training neural network is made to classify is desirable to last
The output of layer can describe the classification of data record.In feed-forward, the weight of whole network and biasing are all that we are random
Take, so the output of network can't describe the classification of record certainly, it is therefore desirable to adjust the parameter of network, i.e., weighted value and
Bias, the foundation of adjustment be exactly the difference between the output valve of the output layer of network and classification, is reduced by adjusting parameter
This difference, here it is the optimization aims of BP neural network.For output layer:
Ej=Oj(1-Oj)(Oj-yj)
Wherein:EjIndicate the error amount of j-th of node, OjIndicate the output valve of j-th of node, yjRecord output valve.It is intermediate
Hidden layer do not come into contacts with the classification of data record directly, but it is tired by weight by next layer of all node errors
Add, calculation formula is as follows:
Ej=Oj(1-Oj)∑kEkWjk
Wherein:WjkIndicate the weighted value of the node k of the node j to next layer of current layer, EkIt is the mistake of next layer of node k
Rate.After calculating error rate, so that it may to be updated to weight and biasing with error rate, the update status of weight is looked first at, it is public
Formula is as follows
ΔWjk=η EjOi
Wjk=Wjk+ΔWjk
Wherein:η indicates that learning rate, value are 0 to 1 section, and learning rate is arranged big, and training restrains faster, still
It is easily trapped into locally optimal solution, learning rate is arranged smaller, and convergence rate is slower, but can approach global optimum step by step
Solution.After updating weight, last parameter, which needs to update, to be biased, and formula is as follows
Δθj=λ Ej
θj=θj+Δθj
So far, the training process for completing a neural network, by constantly being instructed using all data of data set
Practice, to obtain a model.
3) training end condition
Each round training all uses all records of data set, and there are two types of stop conditions:
(1) maximum iterations are set, such as using number of data sets according to deconditioning after iteration 200 times.
(2) predictablity rate of the training set in neural network is calculated, deconditioning after error is met.
Embodiment predicts neural network using 24 hours one day data of Shenyang City's heating plant, and passes through network
Simulation result is as follows to be predicted to 15 hour datas of future:
Prediction result is passed through with actual numerical value compared with, and precision degree reaches 98% or more.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement etc., should all be included in the protection scope of the present invention made by within refreshing and principle.
Claims (4)
1. a kind of energy supply feedback and the heat load prediction method of meteorologic factor, which is characterized in that this method includes:
Step 1, the influence by wind speed to outdoor temperature is converted into corresponding temperature variation, equivalent outdoor in the case of calculating is calm
Temperature;
Step 2, using heat supply return water temperature as energy supply feedback data, according to historical data first-level pipeline network return water temperature, level-one pipe
Net return water flow velocity, secondary pipe network return water temperature, secondary pipe network water inlet flow velocity, a pipe network inflow temperature and equivalent outdoor temperature
As input, the inflow temperature once to net is trained RBF neural as output;
Step 3, the RBF neural after training predicts primary net inflow temperature.
2. according to the method for claim 1, which is characterized in that the connection weight of RBF neural input layer to hidden layer
It is fixed as 1.
3. according to the method for claim 1, which is characterized in that the computational methods of equivalent outdoor temperature are:
Work as v<When 2m/s, TF=T;
As 2m/s≤v<When 5m/s, TF=T- (v-2);
As 5m/s≤v<When 10m/s, TF=T- (v-5) × [3+0.2 × (v-5)] -3;
As v >=10m/s, TF=T-23;
Wherein, v is wind speed, TFFor equivalent outdoor temperature, T is outdoor actual temperature.
4. according to the method for claim 1, which is characterized in that in step 2 RBF neural be trained including:
The initial data of input is normalized input layer;
Hidden layer determines number, center and the width of hidden layer neuron, Yi Jiyin using the method for randomly selecting network center
Containing layer to the connection weight between input layer, and export output matrix;
Input layer into linear transformation is done, exports output matrix by linear weighted function and to input variable.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810048118.XA CN108376294A (en) | 2018-01-18 | 2018-01-18 | A kind of heat load prediction method of energy supply feedback and meteorologic factor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810048118.XA CN108376294A (en) | 2018-01-18 | 2018-01-18 | A kind of heat load prediction method of energy supply feedback and meteorologic factor |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108376294A true CN108376294A (en) | 2018-08-07 |
Family
ID=63016527
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810048118.XA Pending CN108376294A (en) | 2018-01-18 | 2018-01-18 | A kind of heat load prediction method of energy supply feedback and meteorologic factor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108376294A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109489117A (en) * | 2018-11-21 | 2019-03-19 | 国网青海省电力公司 | The control method and device of accumulation of heat heating system, accumulation of heat heating system |
CN109740803A (en) * | 2018-12-24 | 2019-05-10 | 北京航天智造科技发展有限公司 | A kind of heating network operation optimization method of data-driven |
CN109948824A (en) * | 2018-11-09 | 2019-06-28 | 北京华源热力管网有限公司 | A method of thermal substation thermic load is predicted using mode identification technology |
CN109949180A (en) * | 2019-03-19 | 2019-06-28 | 山东交通学院 | A kind of the cool and thermal power load forecasting method and system of ship cooling heating and power generation system |
CN114707765A (en) * | 2022-05-10 | 2022-07-05 | 浙江大学 | Dynamic weighted aggregation-based federated learning load prediction method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102419827A (en) * | 2011-11-02 | 2012-04-18 | 昆明理工大学 | Radial basis function (RBF) neural network-based boiling heat exchanging prediction method |
CN103591637A (en) * | 2013-11-19 | 2014-02-19 | 长春工业大学 | Centralized heating secondary network operation adjustment method |
-
2018
- 2018-01-18 CN CN201810048118.XA patent/CN108376294A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102419827A (en) * | 2011-11-02 | 2012-04-18 | 昆明理工大学 | Radial basis function (RBF) neural network-based boiling heat exchanging prediction method |
CN103591637A (en) * | 2013-11-19 | 2014-02-19 | 长春工业大学 | Centralized heating secondary network operation adjustment method |
Non-Patent Citations (2)
Title |
---|
王文标: "基于气象因素的集中供热系统热负荷预测研究", 《计算机测量与控制》 * |
袁闪闪: "集中供热系统的热动态特性研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109948824A (en) * | 2018-11-09 | 2019-06-28 | 北京华源热力管网有限公司 | A method of thermal substation thermic load is predicted using mode identification technology |
CN109948824B (en) * | 2018-11-09 | 2021-09-07 | 北京华源热力管网有限公司 | Method for predicting heat load of heating power station by using pattern recognition technology |
CN109489117A (en) * | 2018-11-21 | 2019-03-19 | 国网青海省电力公司 | The control method and device of accumulation of heat heating system, accumulation of heat heating system |
CN109740803A (en) * | 2018-12-24 | 2019-05-10 | 北京航天智造科技发展有限公司 | A kind of heating network operation optimization method of data-driven |
CN109949180A (en) * | 2019-03-19 | 2019-06-28 | 山东交通学院 | A kind of the cool and thermal power load forecasting method and system of ship cooling heating and power generation system |
CN114707765A (en) * | 2022-05-10 | 2022-07-05 | 浙江大学 | Dynamic weighted aggregation-based federated learning load prediction method |
CN114707765B (en) * | 2022-05-10 | 2024-08-20 | 浙江大学 | Federal learning load prediction method based on dynamic weighted aggregation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ye et al. | Predicting electricity consumption in a building using an optimized back-propagation and Levenberg–Marquardt back-propagation neural network: Case study of a shopping mall in China | |
CN108376294A (en) | A kind of heat load prediction method of energy supply feedback and meteorologic factor | |
Kalogirou | Applications of artificial neural networks in energy systems | |
More et al. | Forecasting wind with neural networks | |
Tian et al. | Predictive model of energy consumption for office building by using improved GWO-BP | |
CN104484715A (en) | Neural network and particle swarm optimization algorithm-based building energy consumption predicting method | |
Cui | District heating load prediction algorithm based on bidirectional long short-term memory network model | |
CN109711617A (en) | A kind of medium-term and long-term Runoff Forecast method based on BLSTM deep learning | |
CN110866640A (en) | Power load prediction method based on deep neural network | |
Ruiz-Moreno et al. | Model predictive control based on deep learning for solar parabolic-trough plants | |
CN116415728B (en) | Carbon emission prediction method, system, electronic equipment and storage medium | |
CN113237135B (en) | SD-LSTMs-based apartment independent heating system heat demand prediction and energy-saving control method | |
Kofinas et al. | Daily multivariate forecasting of water demand in a touristic island with the use of artificial neural network and adaptive neuro-fuzzy inference system | |
Wang et al. | Multi-criteria comprehensive study on predictive algorithm of heating energy consumption of district heating station based on timeseries processing | |
CN113962454A (en) | LSTM energy consumption prediction method based on dual feature selection and particle swarm optimization | |
Gwaivangmin et al. | Water demand prediction using artificial neural network for supervisory control | |
Zheng et al. | Digital twin modeling for district heating network based on hydraulic resistance identification and heat load prediction | |
CN117933458A (en) | Intelligent heat supply load prediction method, intelligent heat supply load prediction system and heat supply system regulation and control method | |
Boussaid et al. | Evaluation of graph neural networks as surrogate model for district heating networks simulation | |
CN116485582A (en) | Heat supply optimization regulation and control method and device based on deep learning | |
Yang et al. | A novel exhaust gas temperature prediction method of hot blast stove | |
Ma et al. | Traffic flow forecasting research based on Bayesian normalized Elman neural network | |
Lv et al. | Research on short-term load forecasting approach for smart grid | |
Ramsami et al. | Neural network frameworks for electricity forecasting in mauritius and rodrigues Islands | |
Groumpos et al. | New advanced technology methods for energy efficiency of buildings |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180807 |
|
RJ01 | Rejection of invention patent application after publication |