CN111563610B - Building electric load comprehensive prediction method and system based on LSTM neural network - Google Patents
Building electric load comprehensive prediction method and system based on LSTM neural network Download PDFInfo
- Publication number
- CN111563610B CN111563610B CN202010237534.1A CN202010237534A CN111563610B CN 111563610 B CN111563610 B CN 111563610B CN 202010237534 A CN202010237534 A CN 202010237534A CN 111563610 B CN111563610 B CN 111563610B
- Authority
- CN
- China
- Prior art keywords
- building
- data
- load
- neural network
- lstm 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.)
- Active
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000012549 training Methods 0.000 claims abstract description 54
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000003062 neural network model Methods 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000010606 normalization Methods 0.000 claims abstract description 8
- 238000009434 installation Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 abstract description 4
- 210000004027 cell Anatomy 0.000 description 4
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000005251 gamma ray Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000717 retained effect Effects 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q50/08—Construction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Biophysics (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a building electric load comprehensive prediction method and system based on an LSTM neural network, which are used for acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing; establishing an electric load prediction model of the LSTM neural network, selecting data similar to typical days as training samples, wherein the training data comprise weather factors of the training days, building type data and load data, and training by taking the minimum error of the electric load as a target in the training process to obtain LSTM neural network model parameters; and (3) inputting building data of the building to be tested into the trained electric load prediction model of the LSTM neural network to obtain a typical daily load curve, a month load curve and a year load curve corresponding to the building. The building load prediction method based on the LSTM neural network comprehensively considers different characteristics of the building and the load fluctuation change condition to realize high-precision load prediction of the building, and has the functions of high precision and easy realization.
Description
Technical Field
The invention belongs to the technical field of load prediction, and particularly relates to a building electric load comprehensive prediction method and system based on an LSTM neural network.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The comprehensive intelligent energy is a distributed energy, whether the economic operation of the energy depends on whether the system configuration is optimized or not, the calculation of electric, thermal and cold loads is the basis of the system optimization configuration, and the closeness of the design load and the actual operation load directly determines the rationality of the system configuration and the economical efficiency of the operation.
The electric load prediction of the building searches the influence of the electric load change rule of the building on the new building load according to the historical data of the electric load, weather, building information and the like of the similar building, and seeks the correlation between the electric load demand and various factors, so that the electric load of the new building is scientifically predicted, a corresponding typical daily, monthly and annual load curve is provided, and a foundation is provided for electric load planning of the building.
The research of the combined cooling, heating and power system in China starts earlier, and the inventor discovers in the research that no clear method is provided for configuring the capacity of corresponding equipment in a building, and the main centralized method for predicting the electric load of the building is as follows:
1) Classical computing methods based on building structures;
2) A time-by-time load factor method based on software simulation;
3) A time-by-time energy load sharing proportion method based on historical data.
The above method plays a role in building load planning, but is difficult to provide relatively accurate prediction accuracy against uncertainty of complex and changeable building types and historical data.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a building electric load comprehensive prediction method based on an LSTM neural network, which comprehensively considers different characteristics of a building and the load fluctuation change condition to realize high-precision load prediction of the building, and has the functions of high precision and easy realization.
To achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
a building electric load comprehensive prediction method based on LSTM neural network comprises the following steps:
acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing;
establishing an electric load prediction model of the LSTM neural network, determining the number of input nodes of the neural network model according to the number of input variables, and simultaneously determining the number of nodes of an hidden layer, wherein the output variables of the neural network are electric load data;
selecting data similar to typical days as training samples, wherein the training data comprise weather factors, building type data and load data of the training days, and training by taking the error of electric load as a minimum target in the training process to obtain LSTM neural network model parameters;
and after building data of the building to be tested are input into the trained electric load prediction model of the LSTM neural network, a typical daily load curve, a month load curve and a year load curve corresponding to the building are output for planning of the building power installation.
On the other hand, in order to achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
an LSTM neural network based building electrical load comprehensive prediction system, comprising:
the data acquisition module acquires data from a load database, a historical weather database and a building data database of the historical building and sends the data to the data processing module for data processing;
and a data processing module: acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing;
LSTM predictive model training module: establishing an electric load prediction model of the LSTM neural network, determining the number of input nodes of the neural network model according to the number of input variables, and simultaneously determining the number of nodes of an hidden layer, wherein the output variables of the neural network are electric load data;
selecting data similar to typical days as training samples, wherein the training data comprise weather factors, building type data and load data of the training days, and training by taking the error of electric load as a minimum target in the training process to obtain LSTM neural network model parameters;
load prediction module: and (3) inputting building data of the building to be tested into a trained electric load prediction model of the LSTM neural network, and obtaining a typical daily load curve, a month load curve and a year load curve corresponding to the building, wherein the typical daily load curve, the month load curve and the year load curve are used for planning a building power installation machine.
The one or more of the above technical solutions have the following beneficial effects:
according to the invention, building load data of different areas are trained and fitted by using the LSTM neural network, so that an accurate typical daily load curve, month load curve and year load curve of a typical building are obtained, and the installation planning of a building power supply is guided. The building load prediction method based on the LSTM neural network comprehensively considers different characteristics of the building and the load fluctuation change condition to realize high-precision load prediction of the building, and has the functions of high precision and easy realization.
In comprehensive energy planning, building electric load planning is important, and electric load conditions of different geographic positions and different types of buildings are not only different, but also various influencing factors such as building area, climate conditions of areas where the buildings are located, building people flow, work and rest time, weather factors and the like need to be considered.
According to daily load curve requirements, similar typical days are selected as training samples to conduct daily load prediction, and prediction accuracy is improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is an internal block diagram of a neural network model according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a prediction process according to an embodiment of the present invention.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
Referring to fig. 2, this embodiment discloses a comprehensive prediction method for building electric load based on LSTM neural network, in which building electric load planning is important in comprehensive energy planning, electric load conditions of different geographic locations and different types of buildings are not only different, but also various influencing factors such as building area, climate conditions of the area where the building is located, building people flow, work and rest time, weather factors and the like need to be considered.
The electricity consumption data includes: building type, building area, functional partition, geographic position, climate condition and other inherent information, building people flow, work and rest time and corresponding meteorological data (including temperature, humidity and the like), and historical load information comprises an electric load curve and an electric quantity curve. And training and fitting building load data of different areas by utilizing the LSTM neural network to obtain an accurate typical daily load curve, a month load curve and an annual load curve of a typical building, and guiding the installation planning of a building power supply.
The method specifically comprises the following steps:
and collecting the collected load data, weather parameters and building data, carrying out normalization processing, and taking the processed data as an input variable of a prediction model so as to facilitate training of the LSTM neural network. The load data includes typical daily load data, typical monthly load data, and model annual load data. Typical daily load data are mainly the number of electrical loads at 96 points a day (daily load curve); typical month load data are month load data (month load curve) of 30 days for one month; annual load data is load data (annual load curve) of 12 months a year. The weather parameters mainly comprise temperature (highest temperature, lowest temperature, average temperature), humidity and weather information (cloudy, sunny, cloudy) of the load day. Building attribute data including building type, building area, building flow, and building work and rest time. These data are used as input data for training the neural network, and because of the different dimensions of the data, normalization processing is required for these data. The output of the neural network is a daily load power curve, a monthly load power curve and a yearly load power curve of the predicted day.
According to daily load curve requirements, selecting similar typical days as training samples to conduct daily load prediction.
According to the similarity principle, the data of a plurality of similarity days are used as training samples, the training data comprise weather factors of the training days, building type data and historical load data, the training is carried out, and the output variable is a building electric load curve. An electric load prediction model of the LSTM neural network is established, the model comprises an input layer, a hidden layer and an output layer, the number of input nodes of the neural network model is determined according to the number of input variables of the input layer, the number of nodes of the hidden layer is determined at the same time, and the variables of the output layer of the neural network are electric load data.
And training by taking the minimum error of the actual electric load data and the predicted electric load data as a target in the training process to obtain LSTM neural network model parameters.
After the LSTM neural network prediction model of a specific building is obtained, building area, building type, building area and building people stream data of the building are input, and a typical daily load curve, a month load curve and a year load curve corresponding to the building are obtained.
The internal structure of the neural network model is shown in fig. 1.
1) When normalized typical daily data is input and the neural network starts training, under the control of a sigmoid function, the hidden layer state at the last moment outputs h t-1 And current input x t To generate an f of 0 to 1 t Value f t Determining whether to let last time learn information C t-1 Through the device. I.e. cell state C at the last moment t-1 How much is saved to the current time C t 。
f t ＝σ(W f ·[h t-1 ,x t ]+b f ) (1)
Sigma is a sigmoid function, there are three sigmoid functions in the LSTM block, the output of which ranges from 0 to 1, which act as soft switches to decide which signals should pass through these gates, f t A value of 0 represents "no pass through", f t A value of 1 represents "pass-through permitted". W (W) f And b f Respectively corresponding weight coefficient matrixes and bias items.
2) Updating cell state values, i.e. determining input x at the current time t How much is saved to state element c t 。
i＝σ(W i ·[h t-1 ,x t ]+b i ) (2)
Tanh is a hyperbolic tangent activation function and i represents the output value of the input gate. The equations (2) and (3) respectively obtain an output state value determined by the sigmoid function of the input gate and an output state generated by the tanh function, the data leaves useful information through the input gate, useless information is removed, and the new neuron state is the combination of the original cell state under the action of the forgetting gate and updated information:
f t is the output of the forgetting gate and controls the state of the cell of the upper layer, namely C t-1 The degree to which the user is left behind,is a two-output multiplication operation of the input gate, indicating how much new information is retained.
3) Model output, control unit state c t How much output is to the current output h of LSTM t . First, an initial output o is obtained through a sigmoid layer t Then c is performed using the tanh function t Scaling the value to between-1 and 1, and multiplying the value with the output obtained by sigmoid pair by pair to obtain the output state value h of the model t I.e. the state value of the load prediction.
o t ＝σ(W o ·[h t-1 ,x t ]+b o ) (5)
h t ＝o t *tanh(C t ) (6)
The model comprehensively considers influencing factors of the electric load of the building: including building type (bui_t), building Area (bui_ Squ), building function partition, building geographic location (bui_area), building stream (bui_flow), work and rest time (bui_timetab), weather temperature (bui_temp), weather humidity (bui_humi), weather information (bui_cli). The model can comprehensively consider various factors influencing the electric load of the building, and takes the factors as input variables to train the model.
The model normalizes the input variables. Because the input variables have different dimensions, and particularly, some variables have larger differences, the input variables are normalized before training,sigma is [0,1 ]]The adjustment parameters among the two can unify the variables with different dimensions after normalizationAnd processing, and adjusting the normalized expansion degree by using the adjustment parameters.
The choice of similar days mainly takes into account the time interval omega 1 Type difference omega 2 And weather factor difference omega 3 The combination of the similar day characteristics depends on the three factors.
Wherein the time interval omega 1 The method comprises the following steps:
wherein Δd is the number of days apart, LD is the number of days of the similar day to be selected is the same, λ is the time difference, and N is the time interval.
Day type difference omega 2 : and selecting the data of the predicted daily load for association analysis, wherein the association is defined as:
|x j (k)-x i (k) I is x j And x i The absolute value of the kth point of the sequence,is the first-order minimum, expressed in curve x i Find each point x j Minimum difference of->For the second level minimum, this represents the minimum difference for all curves found on the basis of the minimum difference found on the curves. />Similarly, the relevance coefficient of each point is integrated, and the relevance theta of the whole curve is calculated ij The method comprises the following steps:
day type difference omega 2
ω 2 ＝1-θ ij (10)
The weather factor is characterized by: gamma ray 0 ＝{γ 01 ,γ 02 ,γ 03 ,γ 04 ,γ 05 ,γ 06 -6 parameters representing the influence of electrical load, the correlation coefficients representing temperature, humidity, rainfall, weather (sunny, cloudy, rainy), weather respectively being:
need to be equal to r j Make adjustments (due to negative values that occur)
Weather difference value omega between day to be selected and day to be predicted 3
ω 3 ＝1-r j (13)
And combining the factors to obtain a difference value between the candidate day and the predicted day, wherein the difference value is as follows:
||α-β||＝K 1 ω 1 +K 2 ω 2 +K 3 ω 3 (14)
wherein K is 1 +K 2 +K 3 ＝1
Similarity is as follows
φ i ＝1-||α-β|| (15)
The prediction model considers building types as follows: the electric loads of the same building have similarity, and the similar building data is used for training when the data is trained.
The obtained building load model can give a typical daily load curve, a typical month load curve (four seasons) and a model year load curve of a building of a given type according to the type, building area, building work and rest time and geographical location of the building.
When the prediction model considers the position of the building, the load divides the area of the building into 7 areas, which are respectively: a severe cold region of class I, a cold region of class II, a hot summer and cold winter region of class III, a hot summer and hot winter region of class IV, a mild region of class V, a severe cold region of class VI and a cold region of class VII.
The training input data also includes historical electrical load data for a class of buildings, including typical daily load data, monthly load data, and annual load data. The load relation among different areas of the similar building is analyzed, and the four-season typical daily load curve, the month load curve and the year load curve of the similar building with a new given area are predicted on the basis of continuously accumulating data.
Various factors influencing the building electric load are comprehensively considered in the load prediction process, and meanwhile, comprehensive judgment is carried out through accumulation of data from the electric load, the building and meteorological factors.
Example two
Based on the same inventive concept, the object of the present embodiment is to provide a computing device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the steps of a building electrical load comprehensive prediction method based on LSTM neural network in the first embodiment.
Example III
Based on the same inventive concept, an object of the present embodiment is to provide a computer-readable storage medium.
A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of a building electrical load comprehensive prediction method based on an LSTM neural network in embodiment one.
Example III
Based on the same inventive concept, the embodiment aims to provide a building electric load comprehensive prediction system based on an LSTM neural network.
An LSTM neural network based building electrical load comprehensive prediction system, comprising:
and a data processing module: acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing;
LSTM predictive model training module: establishing an electric load prediction model of the LSTM neural network, determining the number of input nodes of the neural network model according to the number of input variables, and simultaneously determining the number of nodes of an hidden layer, wherein the output variables of the neural network are electric load data;
selecting data similar to typical days as training samples, wherein the training data comprise weather factors, building type data and load data of the training days, and training by taking the error of electric load as a minimum target in the training process to obtain LSTM neural network model parameters;
load prediction module: and (3) inputting building data of the building to be tested into a trained electric load prediction model of the LSTM neural network, and obtaining a typical daily load curve, a month load curve and a year load curve corresponding to the building, wherein the typical daily load curve, the month load curve and the year load curve are used for planning a building power installation machine.
The system also comprises a data acquisition module, wherein the data acquisition module acquires data from a load database, a historical meteorological database and a building data database of the historical building and sends the data to the data processing module for data processing.
The steps involved in the devices of the second, third and fourth embodiments correspond to those of the first embodiment of the method, and the detailed description of the embodiments can be found in the related description section of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (8)
1. The building electric load comprehensive prediction method based on the LSTM neural network is characterized by comprising the following steps of:
acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing;
establishing an electric load prediction model of the LSTM neural network, determining the number of input nodes of the neural network model according to the number of input variables, and simultaneously determining the number of nodes of an hidden layer, wherein the output variables of the neural network are electric load data;
selecting data similar to typical days as training samples, wherein the training data comprise weather factors, building type data and load data of the training days, and training by taking the error of electric load as a minimum target in the training process to obtain LSTM neural network model parameters;
after building data of a building to be tested are input into a trained electric load prediction model of the LSTM neural network, a typical daily load curve, a month load curve and a year load curve corresponding to the building are output and are used for planning a building power installation machine;
the selection of the similar typical days mainly considers three factors of time interval, type difference and weather factor difference;
the building includes, but is not limited to: hospitals, office buildings, schools, and markets only use similar building data for training when training data.
2. The integrated prediction method for building electric load based on LSTM neural network as claimed in claim 1, wherein after obtaining LSTM neural network prediction model of building, building area, building type, building area and building people stream data of building are input, and typical daily load curve, month load curve and year load curve corresponding to building are obtained.
3. The integrated prediction method for building electric load based on LSTM neural network as claimed in claim 1, wherein the electric load prediction model of corresponding LSTM neural network is built for different buildings respectively, model training is performed by using corresponding training data, electric load prediction models of different building LSTM neural networks are obtained, and stored, when a certain type of building load is predicted, the electric load prediction model of corresponding LSTM neural network is selected for prediction.
4. The method for comprehensive prediction of building electrical load based on LSTM neural network as claimed in claim 1, wherein when the electrical load prediction model of LSTM neural network is outputted, an initial output is obtained through a sigmoid layer, then a tanh function is used to scale the state value of the control unit to between-1 and 1, and then the output obtained through the sigmoid layer is multiplied pair by pair, so that the output of the model is obtained.
5. The integrated prediction method for building electric load based on LSTM neural network as claimed in any one of claims 1-4, wherein the electric load prediction model of LSTM neural network comprehensively considers the influence factors of the electric load of the building: the model comprehensively considers various factors influencing the electric load of the building and takes the factors as input variables to train the model.
6. An LSTM neural network-based building electrical load comprehensive prediction system, comprising:
the data acquisition module acquires data from a load database, a historical weather database and a building data database of the historical building and sends the data to the data processing module for data processing;
and a data processing module: acquiring load data, weather parameters and building data of a typical building and carrying out normalization processing;
LSTM predictive model training module: establishing an electric load prediction model of the LSTM neural network, determining the number of input nodes of the neural network model according to the number of input variables, and simultaneously determining the number of nodes of an hidden layer, wherein the output variables of the neural network are electric load data;
selecting data similar to typical days as training samples, wherein the training data comprise weather factors, building type data and load data of the training days, and training by taking the error of electric load as a minimum target in the training process to obtain LSTM neural network model parameters;
load prediction module: after building data of a building to be tested are input into a trained electric load prediction model of an LSTM neural network, a typical daily load curve, a month load curve and a year load curve corresponding to the building are obtained and are used for planning a building power supply installation;
the selection of the similar typical days mainly considers three factors of time interval, type difference and weather factor difference;
the building includes, but is not limited to: hospitals, office buildings, schools, and markets only use similar building data for training when training data.
7. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of a building electrical load comprehensive prediction method based on an LSTM neural network as claimed in any one of claims 1 to 5.
8. A computing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of a building electrical load comprehensive prediction method based on LSTM neural network as claimed in any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010237534.1A CN111563610B (en) | 2020-03-30 | 2020-03-30 | Building electric load comprehensive prediction method and system based on LSTM neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010237534.1A CN111563610B (en) | 2020-03-30 | 2020-03-30 | Building electric load comprehensive prediction method and system based on LSTM neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111563610A CN111563610A (en) | 2020-08-21 |
| CN111563610B true CN111563610B (en) | 2023-07-18 |
Family
ID=72070330
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010237534.1A Active CN111563610B (en) | 2020-03-30 | 2020-03-30 | Building electric load comprehensive prediction method and system based on LSTM neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111563610B (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112116153B (en) * | 2020-09-18 | 2022-10-04 | 上海电力大学 | Park multivariate load joint prediction method coupling Copula and stacked LSTM network |
| CN112766535B (en) * | 2020-12-16 | 2023-04-07 | 国网山东省电力公司日照供电公司 | Building load prediction method and system considering load curve characteristics |
| CN112686442A (en) * | 2020-12-29 | 2021-04-20 | 博锐尚格科技股份有限公司 | Air conditioner tail end energy consumption prediction method and system based on operation diversity |
| CN112966868A (en) * | 2021-03-13 | 2021-06-15 | 山东大学 | Building load day-ahead prediction method and system |
| CN113221315B (en) * | 2021-03-23 | 2022-12-06 | 青岛理工大学 | Design and model selection method and system for building seawater source heat pump system unit |
| CN113091593A (en) * | 2021-03-31 | 2021-07-09 | 重庆文理学院 | Sensing system and method for low-strain scene of building material |
| CN113177366B (en) * | 2021-05-28 | 2024-02-02 | 华北电力大学 | Comprehensive energy system planning method and device and terminal equipment |
| CN113537571A (en) * | 2021-06-19 | 2021-10-22 | 复旦大学 | Construction energy consumption load prediction method and device based on CNN-LSTM hybrid network model |
| CN113610152B (en) * | 2021-08-06 | 2023-07-18 | 天津大学 | Load mode-based air conditioning system flexibility operation strategy formulation method |
| CN113515898A (en) * | 2021-08-09 | 2021-10-19 | 国网湖北省电力有限公司电力科学研究院 | Refined monitoring and modeling method for comprehensive energy utilization network of' electricity, cold and hot gas |
| CN113960925A (en) * | 2021-08-30 | 2022-01-21 | 中科苏州微电子产业技术研究院 | Building energy consumption control method and device based on artificial intelligence |
| CN114219144A (en) * | 2021-12-13 | 2022-03-22 | 广西电网有限责任公司北海供电局 | Main transformer load prediction method and system |
| CN114330902B (en) * | 2021-12-31 | 2024-10-18 | 河北工业大学 | Distribution line load prediction method based on decoupling mechanism |
| CN115018116A (en) * | 2022-04-18 | 2022-09-06 | 西安建筑科技大学 | Community building heating load prediction method based on system dynamics multi-subject modeling |
| CN115481788B (en) * | 2022-08-31 | 2023-08-25 | 北京建筑大学 | Phase change energy storage system load prediction method and system |
| CN116227725A (en) * | 2023-03-17 | 2023-06-06 | 广东热矩智能科技有限公司 | Load prediction method and device for building air conditioning system and electronic equipment |
| CN118172198A (en) * | 2024-03-13 | 2024-06-11 | 四川辰鳗科技有限公司 | Building energy load prediction method, system, electronic equipment and medium |
| CN118428550A (en) * | 2024-05-22 | 2024-08-02 | 北京云庐科技有限公司 | Prediction method, training method and prediction device for building energy consumption load |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063911A (en) * | 2018-08-03 | 2018-12-21 | 天津相和电气科技有限公司 | A kind of Load aggregation body regrouping prediction method based on gating cycle unit networks |
| CN110298501A (en) * | 2019-06-21 | 2019-10-01 | 河海大学常州校区 | Electric load prediction technique based on long Memory Neural Networks in short-term |
-
2020
- 2020-03-30 CN CN202010237534.1A patent/CN111563610B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063911A (en) * | 2018-08-03 | 2018-12-21 | 天津相和电气科技有限公司 | A kind of Load aggregation body regrouping prediction method based on gating cycle unit networks |
| CN110298501A (en) * | 2019-06-21 | 2019-10-01 | 河海大学常州校区 | Electric load prediction technique based on long Memory Neural Networks in short-term |
Non-Patent Citations (3)
| Title |
|---|
| "区域供冷系统能源站冷负荷预测及同时使用系数的确定";苏斌 等;《城市与建筑》;20141231;第13卷(第131期);第35-46页 * |
| "变压器经济运行实用化方法的研究";杨佳;《中国优秀博硕士学位论文全文数据库 (硕士) 工程科技Ⅱ辑》;20060615(第06期);第12-14页 * |
| "基于AＲIMA LSTM组合模型的楼宇短期负荷预测方法研究";李鹏辉 等;《上海电力学院学报》;20191231;第35卷(第6期);第573-579页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111563610A (en) | 2020-08-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111563610B (en) | Building electric load comprehensive prediction method and system based on LSTM neural network | |
| CN108846517B (en) | Integration method for predicating quantile probabilistic short-term power load | |
| CN113128793A (en) | Photovoltaic power combination prediction method and system based on multi-source data fusion | |
| CN110503256B (en) | Short-term load prediction method and system based on big data technology | |
| CN111260136A (en) | Building short-term load prediction method based on ARIMA-LSTM combined model | |
| CN112330077B (en) | Power load prediction method, power load prediction device, computer equipment and storage medium | |
| CN110380444B (en) | Capacity planning method for distributed wind power orderly access to power grid under multiple scenes based on variable structure Copula | |
| Sülo et al. | Energy efficient smart buildings: LSTM neural networks for time series prediction | |
| Selvi et al. | Investigation of performance of electric load power forecasting in multiple time horizons with new architecture realized in multivariate linear regression and feed-forward neural network techniques | |
| Al-Rousan et al. | Assessment of predicting hourly global solar radiation in Jordan based on Rules, Trees, Meta, Lazy and Function prediction methods | |
| CN113591368A (en) | Comprehensive energy system multi-energy load prediction method and system | |
| CN116128150B (en) | Photovoltaic intelligent output prediction method based on two-stage optimization | |
| CN111222700A (en) | Day-ahead electricity price probability prediction method based on dynamic network quantile model | |
| CN109978253A (en) | A kind of short-term load forecasting method based on incremental learning | |
| CN105160441A (en) | Real-time power load forecasting method based on integrated network of incremental transfinite vector regression machine | |
| CN112381315A (en) | LS-SVM intelligent platform area load prediction method and system based on PSO optimization | |
| JPH09215192A (en) | Method for predicting daily load curve | |
| CN115689146A (en) | Multi-energy intelligent building load scheduling method and system | |
| Khan et al. | Output power prediction of a photovoltaic module through artificial neural network | |
| Kumar et al. | Forecasting indoor temperature for smart buildings with ARIMA, SARIMAX, and LSTM: A fusion approach | |
| Chen et al. | Probabilistic Prediction of Photovoltaic Power Using Bayesian Neural Network-LSTM Model | |
| Agrawal et al. | Most influential parametrical and data needs for realistic wind speed prediction | |
| Fidalgo et al. | Forecasting Portugal global load with artificial neural networks | |
| CN113570106A (en) | Power load prediction method and device based on expressway neural network | |
| Yong et al. | Short-term building load forecasting based on similar day selection and LSTM network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |