CN104331737A - Office building load prediction method based on particle swarm neural network - Google Patents
Office building load prediction method based on particle swarm neural network Download PDFInfo
- Publication number
- CN104331737A CN104331737A CN201410675567.9A CN201410675567A CN104331737A CN 104331737 A CN104331737 A CN 104331737A CN 201410675567 A CN201410675567 A CN 201410675567A CN 104331737 A CN104331737 A CN 104331737A
- Authority
- CN
- China
- Prior art keywords
- neural network
- office building
- load
- particle
- value
- 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
Abstract
The invention discloses an office building load prediction method based on a particle swarm neural network. The method includes the following steps of: determining the input feature variable and the output target vector of an office building load prediction neural network; initializing a particle swarm solution set; calculating the fitness value of each particle; updating the local optimal position and the global optimal position of each particle; updating speeds and positions of particles; judging ending conditions; is the ending conditions are met, outputting the current optimal position; assigning the neural network and simulating the neural network, and predicting the load of an office building. Through the office building load prediction method based on the neutral network, all internal disturbance and external disturbance factors influencing fluctuation of the official building load are comprehensively considered. Meanwhile, aiming at the special periodic electricity consumption characteristic of the office building, the periodic load change is also considered; the high-precision load prediction of the office building is achieved by using manually simulating the neutral network; the office building load prediction method based on the particle swarm neural network has the advantages of high load prediction precision and simple and easy to implement.
Description
Technical field
The present invention relates to a kind of load forecasting method, relate to a kind of office building load forecasting method based on PSO Neural Network specifically.
Background technology
When power supply node fault and For Congestion appear in electrical network imbalance of supply and demand, some areas, the scheduling pressure increase of power scheduling department, cannot maintain the equilibrium of supply and demand of grid side.On the other hand, current intelligent grid faces serious problems in regional development, cause the contradiction between region peak load and plant factor, and this contradiction to cause with integral load level is on the low side because local localised load is developed at a too fast speed, be all that the bracing wire of sliver is rationed the power supply before in order to alleviate the problems, very large impact be created on the electricity consumption of user side.The more important thing is, current office building user power utilization efficiency and plant factor are low, electricity consumption peak-valley difference widens gradually, office building user has more and more higher need for electricity and demand for services to power grid enterprises.Under this present situation, the automatic demand response of user side becomes the inexorable trend of intelligent grid development, and also becomes the most important thing as the high precision Load Forecast Algorithm of one of its gordian technique.Due to the real-time dynamic response mechanism of demand response, cause load curtailment strategy to the influence of fluctuations of load prediction, high-precision Load Forecast Algorithm can be next step and judges that demand response control strategy provides decision-making foundation in advance.
In order to realize the high precision load prediction of office building, current numerous scholars have done deep research to its load prediction technology.Main technical schemes is, by analyzing on the basis of reference design standard for energy efficiency of buildings specification and lot of documents, sets up the Typical Office Building analytical model that meets this area energy consumption level.With this BUILDINGS MODELS for carrier, respectively general analyzes is carried out to building energy consumption influence factor according to building load, air-conditioning system and refrigeration plant three modules, then simulation test is arranged to the conspicuousness energy consumption factor, regretional analysis is carried out to test figure, thus set up the energy consumption equation of building, its load is predicted.Also have some scholars to utilize the history energy consumption of building, predicted by the load of seasonal effect in time series exponential smoothing to subsequent time, all achieve higher precision of prediction.
One of above-mentioned load forecasting method mainly utilizes simulation of energy consumption software, sets up the building analysis model and the energy consumption equation that meet energy consumption level, then predicts building load.The method relates to very complicated Building Energy Analysis, and is designed into relevant simulation of energy consumption software, only has the architect of a seldom part to grasp at present.For general technician, the method complex, is not easy to realize.And for time series forecasting, although implementation procedure is relatively simple, but owing to only considered the historical load data influence factor of building, the outdoor medial temperature of other important factor in order picture, outdoor medial humidity, personnel's variation etc. are not all considered, cause the standard that load prediction precision does not reach high precision Load Forecast Algorithm.
Summary of the invention
Technical matters to be solved by this invention is the defect overcoming prior art, and provide a kind of office building load forecasting method based on PSO Neural Network, load prediction precision is high, is simple and easy to realize.
In order to realize the high-precision forecast of office building load, the present invention proposes a kind of office building load forecasting method based on PSO Neural Network, mainly comprises following steps:
Step 1: establish the input feature vector variable of office building load prediction neural network model and export object vector.Wherein, the input feature vector variable of neural network model comprises: the load value in T-1 moment, the load value in T-2 moment, the load value in T-3 moment, the load value in T-24 moment, Zhou Bianliang, outdoor medial temperature, outdoor medial humidity, air-conditioner host 1 opening, air-conditioner host 2 opening, air-conditioner host 3 opening, chilled water temperature; Export the predicted load that object vector is the T moment.
Step 2: initialization population disaggregation x
i=(a
1i, a
2i, b
1i, b
2i, c
i), wherein a
1iand b
1irepresent the weights and threshold between neural network input layer and hidden layer respectively, a
2iand b
2irepresent the weights and threshold between neural network hidden layer and output layer respectively, c
irepresent the number of neural network hidden layer, respectively initialization is carried out to it.
Step 3: the fitness value calculating each particle, target function value f (x) namely under current particulate.
Step 4: the more local optimum position p of new particle
i(t) and global optimum position p
g(t).
Step 5: the speed v upgrading particulate
ijand position x (t+1)
ij(t+1).With reference to following formula:
v
ij(t+1)＝ω·v
ij(t)+d
1r
1j(t)(p
ij(t)-x
ij(t))+d
2r
2j(t)(p
gj(t)-x
ij(t))
x
ij(t+1)＝x
ij(t)+v
ij(t+1)
Wherein, ω is inertia weight, d
1and d
2for aceleration pulse, r
1j(t) and r
2jt () is two separate random numbers between 0 and 1, t refers to t generation, v
ij(t) and x
ijt () represents speed and the position in t generation respectively, v
ijand x (t+1)
ij(t+1) speed and the position in t+1 generation is represented respectively.
Step 6: judge termination condition.
If not yet meet termination condition, then return step 3; If meet termination condition, then export current optimal location p
g(t).
Step 7: assignment neural network also carries out neuron network simulation, prediction office building load.
Because office building has typical load period, every day and weekly same load curve are in the daytime periodic Changing Pattern, and out door climatic parameter and indoor occupant variation are the main causes causing office building load fluctuation with equipment start-stop.The present invention has considered affects the major influence factors of office building load variations and the periodicity of load, and forecast model only needs input feature vector variable can realize the high-precision forecast of building load after setting up, and method is easy, easily realizes.
The beneficial effect that the present invention reaches:
By the office building load forecasting method based on neural network, considered affect office building load fluctuation change all in disturb and disturb factor outward, be i.e. out door climatic parameter, indoor occupant variation and equipment start-stop factor.Simultaneously for the periodicity electrical characteristics that office building is special, its cyclical demand fluctuation is taken into account in the lump, utilize artificial Neural Network Simulation, realize the high precision load prediction of office building, there is the beneficial effect that load prediction precision is high, be simple and easy to realization.
Accompanying drawing explanation
Fig. 1 is the office building Load Forecast Algorithm process flow diagram based on PSO Neural Network
Fig. 2 is the optimum office building load prediction neural network model after training terminates.
Embodiment
Below in conjunction with accompanying drawing, the invention will be further described.Following examples only for technical scheme of the present invention is clearly described, and can not limit the scope of the invention with this.
As shown in Figure 1, it mainly comprises following seven steps to the solution of the present invention process flow diagram, is specifically implemented as follows:
Step 1: determine input feature vector variable and export object vector
This embodiment is a certain office building in Beijing.Generally speaking, building load is caused with interior disturbing by outer the disturbing of building, and out door climatic parameter and indoor occupant variation are the main causes causing building loading to fluctuate with equipment start-stop.For office building, personnel change comparatively rule, and building load also often presents periodic Changing Pattern, and every day and weekly same load curve in the daytime have similarity.
Simulate the load prediction of this office building in the present invention by the BP neural network of three layers, it is regarded as a nonlinearity system.First the determination of model structure is will determine input feature vector variable and export object vector.The load value in the T-1 moment in invention in input feature vector variables choice historical load, the load value in T-2 moment, the load value in T-3 moment, the load value in T-24 moment; Outdoor medial temperature in out door climatic parameter, outdoor medial humidity; Air-conditioner host 1 opening in equipment start-stop factor, air-conditioner host 2 opening, air-conditioner host 3 opening, chilled water temperature; Consider the load period of office building uniqueness, input feature vector variable also should add the parameter of reflection load period change, is referred to as Zhou Bianliang (this variable-value is the integer of 1 to 7, represents Monday to Sunday).
The object vector of load forecasting model of the present invention is the load in office building T moment, so the output object vector of model is chosen to be the predicted load in T moment.
To sum up, in the present invention, the input layer of neural network prediction model adopts 11 neural network unit, and input feature vector variable is: the load value in T-1 moment, the load value in T-2 moment, the load value in T-3 moment, the load value in T-24 moment, Zhou Bianliang, outdoor medial temperature, outdoor medial humidity, air-conditioner host 1 opening, air-conditioner host 2 opening, air-conditioner host 3 opening, chilled water temperature; Output layer adopts 1 neural network unit, exports the predicted load that object vector is the T moment.Collect the sample value of this building 200 input feature vector variablees and output object vector, as the sample data of next step neuron network simulation.
Step 2: initialization population disaggregation.
Population disaggregation x is set
i=(a
1i, a
2i, b
1i, b
2i, c
i), wherein a
1iand b
1irepresent the weights and threshold between neural network input layer and hidden layer respectively, a
2iand b
2irepresent the weights and threshold between neural network hidden layer and output layer respectively, c
irepresent the number of neural network hidden layer.Learnt by step 1, the input layer number of neural network is 11, and output layer number is 1, so, a
1for c
i× 11 rank matrixes, a
2be 1 × c
irank matrix, b
1for c
i× 1 rank matrix, b
2it is 1 × 1 rank matrix.
Setting population scale is s, and evolution greatest iteration number is t
max, to each particle x
icarry out the initialization of position and speed.To arbitrary i, j (j=1,2 ..., 4), between scope [0,1], all obey to be uniformly distributed producing x
ijand v
ij; To arbitrary i, j (j=5), namely the 5th dimension element of population disaggregation represents the number of neural network hidden layer, due to according to hidden layer determination publicity
(wherein n
1for hidden layer unit number, n is input block number, and m is output unit number, and a is the constant between [1,9]) can determine that its scope is for [2,13], therefore obedience is uniformly distributed generation x between [2,13]
ijand v
ij.For Arbitrary Particles, set the initial value p of its desired positions
i=x
i.
Step 3: calculate fitness value.
Each disaggregation x of population
i(t), t=(1,2,3 ..., t
max) all correspond to one group of weights of neural network, threshold value and hidden layer number, these values are assigned to neural network respectively and carry out network training.Fitness value f (the x of particle is calculated by the inverse of the training error calculating neural network
i(t)).Computing formula is as follows
wherein, y
jfor the target output value of neural network, y '
jfor the real output value of neural network, n is the number of train samples.
Step 4: the more local of new particle and global optimum position.
For each particle, by its adaptive value and the desired positions p lived through
it the adaptive value of () compares, the local optimum location updating of particle is with reference to following formula:
Global optimum position in colony is designated as p
g(t), p
g(t) ∈ { p
0(t), p
1(t) ..., p
s(t) }, and
f(P
g(t))＝min{f(P
0(t)),f(P
1(t)),…,f(P
s(t))}。Global optimum's location updating formula is as follows:
Step 5: the more speed of new particle and position.With reference to following formula:
v
ij(t+1)＝ω·v
ij(t)+d
1r
1j(t)(p
ij(t)-x
ij(t))+d
2r
2j(t)(p
gj(t)-x
ij(t))
x
ij(t+1)＝x
ij(t)+v
ij(t+1)
Wherein, ω is inertia weight, d
1and d
2for aceleration pulse, r
1j(t) and r
2jt () is two separate random numbers between 0 and 1, t refers to current iteration number of times, and t+1 refers to following iteration number of times.Value ω=0.729 in the present embodiment, d
1=d
2=1.49.
Step 6: judge termination condition.
If f is (p
g(t))-f (p
g(t+1))≤ε, or t>t
max, then optimize and terminate and export current globally optimal solution p
gt (), wherein ε is the limits of error, t
maxfor maximum iteration time.Otherwise return step 3.
Step 7: assignment neural network also carries out neuron network simulation, prediction office building load.
Each solution of population correspond to one group of weights of neural network, threshold value and hidden layer number, so its globally optimal solution just correspond to one group of best initial weights of neural network, threshold value and hidden layer number.The global optimum assignment exported just is obtained the optimal neural network model of office building load prediction to neural network.Optimization model is with reference to shown in accompanying drawing 2.
The 2 optimum office building load prediction neural network models obtained with reference to the accompanying drawings, to predict the office building load value in T moment, directly can input its characteristic variable in a model: the load value in T-1 moment, the load value in T-2 moment, the load value in T-3 moment, the load value in T-24 moment, Zhou Bianliang, outdoor medial temperature, outdoor medial humidity, air-conditioner host 1 opening, air-conditioner host 2 opening, air-conditioner host 3 opening, chilled water temperature, automatically the predicted load in office building T moment can be exported by model, realize the high precision load prediction of office building.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from the technology of the present invention principle; can also make some improvement and distortion, these improve and distortion also should be considered as protection scope of the present invention.
Claims (6)
1., based on an office building load forecasting method for PSO Neural Network, it is characterized in that, comprise following steps:
Step 1: establish the input feature vector variable of office building load prediction neural network model and export object vector;
Step 2: initialization population disaggregation x
i=(a
1i, a
2i, b
1i, b
2i, c
i), wherein a
1iand b
1irepresent the weights and threshold between neural network input layer and hidden layer respectively, a
2iand b
2irepresent the weights and threshold between neural network hidden layer and output layer respectively, c
irepresent the number of neural network hidden layer, respectively initialization is carried out to it;
Step 3: the fitness value calculating each particle, target function value f (x) namely under current particulate;
Step 4: the more local optimum position p of new particle
i(t) and global optimum position p
g(t);
Step 5: the speed v upgrading particulate
ijand position x (t+1)
ij(t+1);
Step 6: judge termination condition;
If not yet meet termination condition, then return step 3; If meet termination condition, then export current optimal location p
g(t);
Step 7: assignment neural network also carries out neuron network simulation, prediction office building load.
2. the office building load forecasting method based on PSO Neural Network according to claim 1, is characterized in that, in step 1, the input feature vector variable of neural network model comprises: the load value in T-1 moment, the load value in T-2 moment, the load value in T-3 moment, the load value in T-24 moment, Zhou Bianliang, outdoor medial temperature, outdoor medial humidity, air-conditioner host 1 opening, air-conditioner host 2 opening, air-conditioner host 3 opening, chilled water temperature;
Export the predicted load that object vector is the T moment.
3. the office building load forecasting method based on PSO Neural Network according to claim 1, is characterized in that, in step 3, and each disaggregation x of population
i(t), t=(1,2,3 ..., t
max) all correspond to one group of weights of neural network, threshold value and hidden layer number, these values be assigned to neural network respectively and carry out network training, being calculated the fitness value f (x of particle by the inverse of the training error calculating neural network
i(t)).
4. the office building load forecasting method based on PSO Neural Network according to claim 1, is characterized in that, in step 4, more the local of new particle and the step of global optimum position are:
For each particle, by its adaptive value and the local optimum position p lived through
it the adaptive value of () compares, the local optimum location updating of particle is with reference to following formula:
Wherein, f (x
i(t+1)) be the adaptive value in particle t+1 generation, f (p
i(t)) for particle t is for the adaptive value of the desired positions lived through;
Global optimum position p in colony
gt () is p
g(t) ∈ { p
0(t), p
1(t) ..., p
s(t) }, and
F (P
g(t))=min{f (P
0(t)), f (P
1(t)) ..., f (P
s(t)) }; Global optimum's location updating formula is as follows:
5. the office building load forecasting method based on PSO Neural Network according to claim 1, is characterized in that, in step 5, upgrades the speed v of particulate
ijand position x (t+1)
ij(t+1) formula is as follows:
v
ij(t+1)＝ω·v
ij(t)+d
1r
1j(t)(p
ij(t)-x
ij(t))+d
2r
2j(t)(p
gj(t)-x
ij(t))
x
ij(t+1)＝x
ij(t)+v
ij(t+1)
Wherein, ω is inertia weight, d
1and d
2for aceleration pulse, r
1j(t) and r
2jt () is two separate random numbers between 0 and 1, t refers to iterations; v
ij(t) and x
ijt () represents speed and the position in t generation respectively, v
ijand x (t+1)
ij(t+1) speed and the position in t+1 generation is represented respectively.
6. the office building load forecasting method based on PSO Neural Network according to claim 1, is characterized in that,
In step 6, termination condition is:
If f is (p
g(t))-f (p
g(t+1))≤ε, or iterations t > is t
max, then optimize and terminate and export current globally optimal solution p
gt (), wherein ε is the limits of error, t
maxfor maximum iteration time;
Otherwise return step 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410675567.9A CN104331737A (en) | 2014-11-21 | 2014-11-21 | Office building load prediction method based on particle swarm neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410675567.9A CN104331737A (en) | 2014-11-21 | 2014-11-21 | Office building load prediction method based on particle swarm neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104331737A true CN104331737A (en) | 2015-02-04 |
Family
ID=52406458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410675567.9A Pending CN104331737A (en) | 2014-11-21 | 2014-11-21 | Office building load prediction method based on particle swarm neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104331737A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096761A (en) * | 2016-06-01 | 2016-11-09 | 新奥泛能网络科技股份有限公司 | A kind of building load Forecasting Methodology based on neutral net and device |
| CN106440185A (en) * | 2016-08-15 | 2017-02-22 | 深圳市纬度节能服务有限公司 | Method and device for energy saving of air-conditioning water system |
| CN107276114A (en) * | 2017-05-03 | 2017-10-20 | 国家电网公司 | A kind of power distribution network light storage generating active frequency fluctuation rejection coefficient Forecasting Methodology |
| CN107301478A (en) * | 2017-06-26 | 2017-10-27 | 广东电网有限责任公司珠海供电局 | A kind of cable run short-term load forecasting method |
| CN107590565A (en) * | 2017-09-08 | 2018-01-16 | 北京首钢自动化信息技术有限公司 | A kind of method and device for building building energy consumption forecast model |
| CN111240197A (en) * | 2020-01-10 | 2020-06-05 | 中国建筑科学研究院有限公司 | Energy efficiency deviation rectifying control method and device for electromechanical system of public building |
| CN111256294A (en) * | 2020-01-17 | 2020-06-09 | 深圳市得益节能科技股份有限公司 | Model prediction-based optimization control method for combined operation of water chilling unit |
| CN111649457A (en) * | 2020-05-13 | 2020-09-11 | 中国科学院广州能源研究所 | Dynamic predictive machine learning type air conditioner energy-saving control method |
| CN113188243A (en) * | 2021-04-08 | 2021-07-30 | 山东师范大学 | Comprehensive prediction method and system for air conditioner energy consumption |
| CN114251753A (en) * | 2021-12-29 | 2022-03-29 | 西安建筑科技大学 | Ice storage air conditioner cold load demand prediction distribution method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005122517A (en) * | 2003-10-17 | 2005-05-12 | Fuji Electric Holdings Co Ltd | Energy demand prediction method, energy demand prediction device and energy demand prediction program and recording medium |
| CN102980272A (en) * | 2012-12-08 | 2013-03-20 | 珠海派诺科技股份有限公司 | Air conditioner system energy saving optimization method based on load prediction |
| CN103729695A (en) * | 2014-01-06 | 2014-04-16 | 国家电网公司 | Short-term power load forecasting method based on particle swarm and BP neural network |
| CN104008427A (en) * | 2014-05-16 | 2014-08-27 | 华南理工大学 | Central air conditioner cooling load prediction method based on BP neural network |
-
2014
- 2014-11-21 CN CN201410675567.9A patent/CN104331737A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005122517A (en) * | 2003-10-17 | 2005-05-12 | Fuji Electric Holdings Co Ltd | Energy demand prediction method, energy demand prediction device and energy demand prediction program and recording medium |
| CN102980272A (en) * | 2012-12-08 | 2013-03-20 | 珠海派诺科技股份有限公司 | Air conditioner system energy saving optimization method based on load prediction |
| CN103729695A (en) * | 2014-01-06 | 2014-04-16 | 国家电网公司 | Short-term power load forecasting method based on particle swarm and BP neural network |
| CN104008427A (en) * | 2014-05-16 | 2014-08-27 | 华南理工大学 | Central air conditioner cooling load prediction method based on BP neural network |
Non-Patent Citations (1)
| Title |
|---|
| 尹新: "群智能算法与电力负荷预测研究", 《中国优秀博士学位论文全文数据库（工程科技II辑）》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096761A (en) * | 2016-06-01 | 2016-11-09 | 新奥泛能网络科技股份有限公司 | A kind of building load Forecasting Methodology based on neutral net and device |
| CN106440185A (en) * | 2016-08-15 | 2017-02-22 | 深圳市纬度节能服务有限公司 | Method and device for energy saving of air-conditioning water system |
| CN107276114A (en) * | 2017-05-03 | 2017-10-20 | 国家电网公司 | A kind of power distribution network light storage generating active frequency fluctuation rejection coefficient Forecasting Methodology |
| CN107301478A (en) * | 2017-06-26 | 2017-10-27 | 广东电网有限责任公司珠海供电局 | A kind of cable run short-term load forecasting method |
| CN107590565A (en) * | 2017-09-08 | 2018-01-16 | 北京首钢自动化信息技术有限公司 | A kind of method and device for building building energy consumption forecast model |
| CN111240197A (en) * | 2020-01-10 | 2020-06-05 | 中国建筑科学研究院有限公司 | Energy efficiency deviation rectifying control method and device for electromechanical system of public building |
| CN111256294A (en) * | 2020-01-17 | 2020-06-09 | 深圳市得益节能科技股份有限公司 | Model prediction-based optimization control method for combined operation of water chilling unit |
| CN111649457A (en) * | 2020-05-13 | 2020-09-11 | 中国科学院广州能源研究所 | Dynamic predictive machine learning type air conditioner energy-saving control method |
| CN113188243A (en) * | 2021-04-08 | 2021-07-30 | 山东师范大学 | Comprehensive prediction method and system for air conditioner energy consumption |
| CN114251753A (en) * | 2021-12-29 | 2022-03-29 | 西安建筑科技大学 | Ice storage air conditioner cold load demand prediction distribution method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104331737A (en) | Office building load prediction method based on particle swarm neural network | |
| Lissa et al. | Deep reinforcement learning for home energy management system control | |
| Borgstein et al. | Evaluating energy performance in non-domestic buildings: A review | |
| Yuce et al. | ANN–GA smart appliance scheduling for optimised energy management in the domestic sector | |
| Yan et al. | Reviews on uncertainty analysis of wind power forecasting | |
| Daut et al. | Building electrical energy consumption forecasting analysis using conventional and artificial intelligence methods: A review | |
| CN112614009B (en) | Power grid energy management method and system based on deep expectation Q-learning | |
| CN103218675B (en) | A kind of based on the short-term load forecasting method of cluster and moving window | |
| WO2023004838A1 (en) | Wind power output interval prediction method | |
| Liu et al. | Automated control of transactive hvacs in energy distribution systems | |
| CN104484715A (en) | Neural network and particle swarm optimization algorithm-based building energy consumption predicting method | |
| Kumar et al. | A novel hybrid model based on particle swarm optimisation and extreme learning machine for short-term temperature prediction using ambient sensors | |
| CN107704875A (en) | Based on the building load Forecasting Methodology and device for improving IHCMAC neutral nets | |
| CN104376389A (en) | Master-slave type micro-grid power load prediction system and master-slave type micro-grid power load prediction method based on load balancing | |
| CN103679263A (en) | Thunder and lightning approach forecasting method based on particle swarm support vector machine | |
| Paul et al. | IoT and machine learning based prediction of smart building indoor temperature | |
| CN104037761B (en) | AGC power multi-objective random optimization distribution method | |
| CN104850891A (en) | Intelligent optimal recursive neural network method of time series prediction | |
| CN102509027B (en) | Wind powder combined predication method based on cross entropy theory | |
| Ozoh et al. | A comparative analysis of techniques for forecasting electricity consumption | |
| Capizzi et al. | An advanced neural network based solution to enforce dispatch continuity in smart grids | |
| CN112149890A (en) | Comprehensive energy load prediction method and system based on user energy label | |
| Waseem et al. | Optimal BRA based electric demand prediction strategy considering instance‐based learning of the forecast factors | |
| CN112418495A (en) | Building energy consumption prediction method based on longicorn stigma optimization algorithm and neural network | |
| Dong et al. | Short-term building cooling load prediction model based on DwdAdam-ILSTM algorithm: A case study of a commercial building |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150204 |