CN113822482A - Method and device for establishing load prediction model of comprehensive energy system - Google Patents
Method and device for establishing load prediction model of comprehensive energy system Download PDFInfo
- Publication number
- CN113822482A CN113822482A CN202111119840.6A CN202111119840A CN113822482A CN 113822482 A CN113822482 A CN 113822482A CN 202111119840 A CN202111119840 A CN 202111119840A CN 113822482 A CN113822482 A CN 113822482A
- Authority
- CN
- China
- Prior art keywords
- load
- energy system
- representing
- comprehensive energy
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 69
- 238000004458 analytical method Methods 0.000 claims abstract description 28
- 230000007246 mechanism Effects 0.000 claims abstract description 20
- 238000007781 pre-processing Methods 0.000 claims abstract description 17
- 238000000556 factor analysis Methods 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims description 31
- 230000006870 function Effects 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 16
- 238000003860 storage Methods 0.000 claims description 14
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 13
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 238000004140 cleaning Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 238000010408 sweeping Methods 0.000 claims description 4
- 238000013519 translation Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims 1
- 230000008569 process Effects 0.000 description 18
- 238000010586 diagram Methods 0.000 description 14
- 230000006872 improvement Effects 0.000 description 10
- 238000013527 convolutional neural network Methods 0.000 description 8
- 238000010219 correlation analysis Methods 0.000 description 6
- 230000008878 coupling Effects 0.000 description 6
- 238000010168 coupling process Methods 0.000 description 6
- 238000005859 coupling reaction Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 1
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 229920001296 polysiloxane Polymers 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
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
- 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
- G06N3/047—Probabilistic or stochastic 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
- G06N3/08—Learning methods
-
- 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Power Engineering (AREA)
- Public Health (AREA)
- Probability & Statistics with Applications (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a device for establishing a load prediction model of an integrated energy system, wherein the method comprises the following steps: preprocessing the collected load historical data of the comprehensive energy system; carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result; and establishing a comprehensive energy system multi-energy load prediction model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load prediction model is used for predicting the comprehensive energy system load.
Description
Technical Field
The invention relates to the technical field of electric power big data analysis, in particular to a method and a device for establishing a comprehensive energy system load prediction model.
Background
Under the background of energy Internet and low-carbon electric power, the comprehensive energy system becomes an important carrier for energy conservation and emission reduction. With the increasing coupling degree of various energy sources in the comprehensive energy system, the large-scale access of renewable energy sources and the gradual marketization of energy source production and consumption, higher requirements on the accuracy, the real-time performance and the reliability of the load are provided. The comprehensive system load prediction can be divided into a cross-regional layer, a regional layer and a user layer from top to bottom. Due to the fact that requirements of users on electric power and cold and heat loads are different from energy preferences, fluctuation among various energy loads is strong, and coupling relations are complex, research on comprehensive energy system user-level load prediction has important significance in achieving scheduling which gives consideration to economic benefits and user satisfaction at the same time.
However, in the research and practice process of the prior art, the inventor of the present invention finds that most of the current research focuses on establishing a comprehensive prediction model by combining machine learning algorithms such as a BP neural network, support vector regression, random forest, gradient boosting decision tree, etc., and although the coupling relationship among multi-energy loads is considered, the invention does not relate to the diversity and rationality of a data structure in the multi-source information feature extraction process and the importance of different factors. Therefore, it is highly desirable to select a method and a system for load prediction of an integrated energy system that can integrate the characteristics of different prediction models.
Disclosure of Invention
The invention aims to provide a method and a device for establishing a load prediction model of an integrated energy system, and aims to solve the problems in the prior art.
The invention provides a method for establishing a load prediction model of an integrated energy system, which comprises the following steps:
preprocessing the collected load historical data of the comprehensive energy system;
carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result;
and establishing a comprehensive energy system multi-energy load prediction model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load prediction model is used for predicting the comprehensive energy system load.
The invention provides a comprehensive energy system load prediction model establishing device, which comprises:
the preprocessing module is used for preprocessing the collected comprehensive energy system load historical data;
the influence factor analysis module is used for carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result;
and the model establishing and predicting module is used for establishing a comprehensive energy system multi-energy load predicting model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load predicting model is used for predicting the comprehensive energy system load.
The embodiment of the invention also provides a device for establishing the comprehensive energy system load prediction model, which comprises the following steps: the system comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the computer program realizes the steps of the comprehensive energy system load prediction model building method when being executed by the processor.
The embodiment of the invention also provides a computer readable storage medium, wherein an implementation program for information transmission is stored on the computer readable storage medium, and when the implementation program is executed by a processor, the steps of the method for establishing the comprehensive energy system load prediction model are implemented.
By adopting the embodiment of the invention, the influence factors of the user-level load of the comprehensive energy system are analyzed by using the Pearson correlation coefficient from the three dimensions of the multi-energy coupling correlation time correlation and the external factor correlation, and the accuracy of the prediction of the multi-energy load of the comprehensive energy system can be effectively improved by the user-level load prediction method with the attention mechanism based on the CNN-BilSTM. .
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a method for building a load prediction model of an integrated energy system according to an embodiment of the invention;
FIG. 2 is a flowchart of an example of a method for building a load prediction model of an integrated energy system according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an integrated energy system multi-energy load prediction model according to an embodiment of the invention;
FIG. 4 is a histogram of historical data time correlation analysis of different load data according to an embodiment of the present invention;
FIG. 5 is a thermodynamic diagram illustrating correlation analysis and external environmental factor impact analysis of multi-energy loads under different load data according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an apparatus for modeling load prediction of an integrated energy system according to a first embodiment of the present invention;
fig. 7 is a schematic diagram of an apparatus for building an integrated energy system load prediction model according to a second embodiment of the present invention.
Detailed Description
In order to make up for the defects of the prior art, the embodiment of the invention provides a comprehensive energy system load prediction method and system based on a CNN-BilSTM attention mechanism model, and a comprehensive energy system user-level load prediction model is established by analyzing multidimensional factors of comprehensive energy system load prediction.
The technical solutions of the present invention will be described clearly and completely with reference to the following embodiments, and it should be understood that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", and the like, indicate orientations and positional relationships based on those shown in the drawings, and are used only for convenience of description and simplicity of description, and do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be considered as limiting the present invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise. Furthermore, the terms "mounted," "connected," and "connected" are to be construed broadly and may, for example, be fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Method embodiment
According to an embodiment of the present invention, a method for building a load prediction model of an integrated energy system is provided, fig. 1 is a flowchart of the method for building the load prediction model of the integrated energy system according to the embodiment of the present invention, as shown in fig. 1, the method for building the load prediction model of the integrated energy system according to the embodiment of the present invention specifically includes:
101, preprocessing collected comprehensive energy system load historical data; specifically, data cleaning and normalization processing are carried out on collected comprehensive energy system load historical data.
That is, the data preprocessing specifically includes the following steps:
(1) and performing data cleaning on the user load data.
(2) And normalizing the user load data.
102, carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result; step 102 specifically includes:
step 1021, analyzing the time correlation of the multiple energy loads of the integrated energy system, calculating the correlation between the load value of each type of load at the t-th moment and the load value of each type of load at the t-d moment by adopting the Pearson correlation coefficient shown in formula 1, setting a proper threshold value, and setting a correlation index rho according to the correlation index rhoXYTo determine the time step for each load input and output:
where ρ isXYRepresenting the load value at the t-th moment and the load value at the t-d moment of each type of loadCorrelation between XjRepresenting a sequence of load values at time t,representing the mean value of the series of load values at time t, YjRepresenting the sequence of load values at time t-d,the average value of the load value sequences at the t-d moments is shown, and N represents the lengths of the load value sequences at the t-d moments and the t-d moments;
step 1022, analyzing the correlation between the multi-energy load of the integrated energy system and the external environmental factors, and calculating by using pearson correlation coefficient shown in formula 1, where the external environmental influencing factors specifically include: setting proper threshold value according to air temperature and air pressure and according to correlation index rhoXYAnd selecting external environment factors with larger relevance as a part of model input.
And 103, establishing a comprehensive energy system multi-energy load prediction model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load prediction model is used for predicting the comprehensive energy system load. Step 103 specifically comprises:
according to the analysis result, local features are extracted through a CNN network model, time information is extracted through a BilSTM network model, weights are distributed through an attention mechanism model, optimization of the weights is carried out through an Adam algorithm, and a comprehensive energy system multi-energy load prediction model is established. The method specifically comprises the following steps:
extracting local dependencies between variables in the time dimension through CNN networks: wherein the convolutional layer of the CNN network consists of a plurality of cores, sweeping the kth filter through the input matrix, as represented:
hk=ReLU(Wk*X+bk) Formula 2;
where denotes the convolution operation, output hkA vector representing the output, the output of the convolutional layer being of a size dcA matrix of x T, wherein dcRepresenting the number of filters and T representing the time dimension. ReLU activation function ReLU (x) max (x,0), WkRepresenting the slope of the kth filter, bkDenotes the intercept of the kth filter, X is the input matrix, X ∈ RD×TD represents a multivariate variable dimension, and T represents a time dimension;
given an input sequence X ═ X1,x2,…,xT) Wherein x ist∈RnN is the dimension of the factor, and the hidden state at time t is ht∈RmAnd m is the dimension of the hidden state, and the input gate, the forgetting gate, the output gate, the memory unit updating and the output calculation in the LSTM model are carried out according to the formulas 3-7:
it=sigmoid(Wi[ht-1;xt]+bi) Formula 3;
ot=sigmoid(Wo[ht-1;xt]+bo) Equation 5;
st=ft⊙st-1+it⊙tanh(Ws[ht-1;xt]+bs) Equation 6;
ht=ot⊙tanh(st) Equation 7;
wherein [ ht-1;xt]∈Rm+nRepresenting a previous hidden state ht-1And the current input xtW is connected in seriesi,Wf,Wo,Ws∈Rm *(m+n)Representing a weight matrix, bi,bf,bo,bs∈RmRepresents an offset vector parameter, <' > represents an element-by-element multiplication, Rm+nRepresenting a matrix of dimension m + n, Rm*(m+n)Representing a matrix of dimensions m x (m + n), RmA matrix of dimension m is represented;
the observations are predicted by processing the BilSTM network model in two directions: forward LSTM units generate the preceding information, i.e. produce hidden statesWhereinGenerating subsequent information to the LSTM unit, i.e. producing hidden statesWhereinConcatenating the two preceding and succeeding information to produce a joint representation as shown in equation 8:
wherein,is the last hidden state of the forward LSTM unit,is the first hidden state of the backward LSTM unit;
the importance of the hidden states is distinguished by assigning different attention weights to the hidden states: given hiRepresenting the hidden state sequence of the BilSTM layer, h is expressed according to equation 9iTranslation to attention weight of target by fully connected layer:
ui=tan(W*hi+ b) equation 9;
wherein b represents the intercept and W represents the slope;
generating probability vector p of weights based on softmax function according to equation 9i:
Wherein u isvRepresents the vth attention weight, and M represents the number of attention weights;
assigning the generated attention weight to the corresponding hidden layer state hiH is performed according to equation 11iThe weighted average of (a) is calculated to obtain an attention scoring function s as follows:
s=∑ipi*hiequation 11;
and iteratively updating the weight according to the training data through an Adam algorithm, and optimizing the weight.
The technical solutions of the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The embodiment of the invention provides a comprehensive energy system load prediction method based on a CNN-BilSTM attention mechanism model, which specifically comprises the following steps of:
step 1(S101), preprocessing the collected comprehensive energy system load data.
Step 2(S102), analyzing multidimensional influence factors of multi-energy load prediction of the integrated energy system, which may specifically adopt the following processing:
step 2.1, analyzing the time correlation of the multi-energy load of the comprehensive energy system, calculating the correlation between the load value of each type of load at the t moment and the load value of each type of load at the t-d moment by adopting a Pearson correlation coefficient described in the following formula, setting a proper threshold value, and setting a correlation index rho according to the correlation index rhoXYTo determine the time step for each load input and output:
step 2.3, analyzing the correlation between the multi-energy load of the comprehensive energy system and external environmental factors, calculating by adopting the Pearson correlation coefficient shown in the formula, and preprocessing the multi-energy loadThe integrated energy system load historical data is used for carrying out correlation analysis between multi-energy loads and external environment influence factors, wherein the external environment influence factors specifically comprise: setting proper threshold value according to air temperature and air pressure and according to correlation index rhoXYAnd selecting external environment factors with larger relevance as a part of model input.
That is, the multidimensional influence factor analysis process
The specific analysis process is as follows:
analyzing influence factors of multi-energy load prediction from three aspects of time correlation analysis of historical data, correlation analysis among multi-energy loads and influence analysis of external environment factors, and displaying the influence factors through a histogram and a heat map respectively.
(1) The correlation between the load data at time t and the load data of the preceding k hours (k is 6 in this embodiment) was analyzed by a pearson correlation coefficient. The historical data time dependency analysis of the different load data is shown in fig. 3. As can be seen from the figure, the three types of load data have the same trend. The correlation gradually decreases with the increase of the time interval, and the correlation change of the electrical load is most significant. This is consistent with the most variable nature of the electrical load in practical applications.
(2) The multi-energy load correlation analysis and the external environment factor influence analysis of different load data by pearson correlation coefficient analysis are shown in fig. 4. In the experiment, the correlations between the power and the cooling/heating load and the pressure and temperature of the meteorological elements were analyzed by using a thermodynamic diagram. The darker the color, the stronger the correlation. As can be seen from fig. 5, there is a certain coupling relationship between the electrical load, the cold load and the thermal load. The air temperature has the greatest effect on the heat load. The influence of air temperature on cold and heat load is greater than air pressure.
Step 3(S103), establishing a comprehensive energy system multi-energy load prediction model, which can specifically adopt the following steps:
and 3.1, extracting local features through a CNN network model. CNN networks can extract local dependencies between variables and short-term patterns in the time dimension. Let the input matrix be X, where X ∈ RD×T. The convolutional layer is composed of a plurality of cores,the time dimension is ω and the variable dimension is D. The kth filter sweeps through the input matrix and can be represented as:
hk=ReLU(Wk*X+bk)
where denotes the convolution operation, output hkWill be a vector. The output of the convolutional layer is a size dc×TcIn which d iscIndicating the number of filters, Tc=T-ω+ 1。
And 3.2, extracting time information through the BilSTM. Recurrent Neural Networks (RNNs) are often used to solve the problem of time series prediction. The LSTM is an improvement of the RNN, can solve the problem of long-distance dependence of the RNN in the training process, and avoids the phenomena of gradient explosion or gradient dispersion. Each LSTM cell has a state s from a time ttThe memory unit consists of three sigmoid gates: input door itForgetting door ftAnd an output gate ot. Given an input sequence X ═ X1,x2…,xT) Wherein x ist∈RnN is the number of factors, and the hidden state at time t is ht∈RmAnd m is a hidden state. The computational expression in the LSTM model is then as follows:
an input gate:
it=sigmoid(Wi[ht-1;xt]+bi)
forget the door:
ft=sigmoid(Wf[ht-1;xt]+bf)
an output gate:
ot=sigmoid(Wo[ht-1;xt]+bo)
updating the memory unit:
st=ft⊙st-1+it⊙tanh(Ws[ht-1;xt]+bs)
and (3) outputting:
ht-ot⊙tanh(st)
wherein [ ht-1;xt]∈Rm+nIs a previous hidden state ht-1And the current input xtIn series. Wi,Wf,Wo,Ws∈Rm(m+n)As a weight matrix, bi,bf,bo,bs∈RmAre bias vector parameters. An element-by-element multiplication.
The BilSTM network allows processing in both directions to predict observations. The forward LSTM unit may process the previous information and the backward LSTM unit may generate the subsequent information. Generating hidden states in view of forward LSTM cellsWhereinGenerating hidden states to LSTM cellsWhereinThe results of the two series are concatenated to produce a joint representation:
whereinIs the last hidden state of the forward LSTM unit,is the first hidden state of the backward LSTM unit. The BilSTM can determine sequence trends in two directions, so that the learning long-term dependence is improved, and the prediction accuracy of the model is improved.
Step 3.3, weights are assigned by the attention mechanism. For use in a care layerRepresenting the correlation between the input sequence and the output result. By distributing different attention weights to the hidden states, the importance of the hidden states is distinguished, and the accuracy of prediction is improved. Given hiRepresenting the hidden state sequence, h, of the BilsTM layeriThe attention weight of the target is converted through the full connection layer. The formula is as follows:
ui=tan(W*hi+b)
the attention weight is then probabilistic, a probability vector piGenerated by the softmax function. The calculation process is as follows:
the generated attention weight is assigned to the corresponding hidden layer state hiWeighted average hiThe calculation is as follows:
s=Σipi*hi
and 3.4, optimizing by adopting an Adam algorithm. Adam is a first order optimization algorithm that can replace the traditional random gradient descent process. It may iteratively update the weights of the neural network based on the training data. Adam combines the optimal performance of AdaGrad and RMSProp algorithms, and can provide an optimization method to solve the problems of sparse gradient and noise.
That is, the load prediction process
The specific process is as follows:
(1) prediction process
The data set selected in this example is from an ASU campus integrated energy system. The data set includes 10198 groups of data including hourly power, thermal load data collected from Tanpey school zone 2015 on 1 month, 1 day 01:00:00-2016 on 2 months, 29 days 23:00:00, and local weather data. And selecting the data in the first 8185 hours as a training set, and using the rest data as a test set. According to the result of the multidimensional influence factor analysis, the input of the prediction model is the electricity, cold and heat load data and the air temperature and air pressure data of the past 4 hours, and the output is the electricity, cold and heat load data of the future 3 hours. In the parametric aspect of the prediction model, the batch size is 80, the epochs is 200, the initial learning rate is 10-3, and the Adam optimizer is selected.
(2) Predictive performance assessment
For evaluating the prediction performance, the Root Mean Square Error (RMSE), the average absolute percentage error (MAPE) and R2 are selected to comprehensively evaluate the prediction effects of single load and comprehensive load.
The root mean square error calculation formula is as follows:
wherein,to predict value, yiM is the actual value and the number of predicted values. The range of RMSE is [0, + ∞). When the predicted value is completely consistent with the actual value, RMSE is equal to 0. The larger the error, the larger the value.
The average absolute percentage error calculation formula is as follows:
the formula for R2 is as follows:
the decision coefficient represents the quality of the fit by a change in the data. From the above formula, the normal value range of the coefficient is determined to be [0, 1 ]. The closer to 1, the stronger the explanatory power of the equation variables for y, the better the model fits to the data.
In this example, the prediction results of the present method (denoted as CBLA) and four prediction models LSTM, GRU, BiLSTM, and BiLSTM (denoted as BLA) based on the attention mechanism were compared and analyzed, and the analysis results are shown in table 1.
TABLE 1
In summary, the embodiment of the invention provides a method for predicting the load of a comprehensive energy system based on a CNN-BiLSTM attention mechanism model, which analyzes the influence factors of the user-level load of the comprehensive energy system from three dimensions of time correlation, multi-energy coupling correlation and external factor correlation by using pearson correlation coefficients. In order to fully mine the time-space correlation of data and improve the prediction precision, a user-level load prediction method with an attention mechanism based on CNN-BilSTM is provided. The embodiment of the invention can effectively improve the accuracy of the multi-energy load prediction of the comprehensive energy system.
Apparatus embodiment one
According to an embodiment of the present invention, there is provided an apparatus for building an integrated energy system load prediction model, fig. 6 is a schematic diagram of the apparatus for building an integrated energy system load prediction model according to an embodiment of the present invention, and as shown in fig. 6, the apparatus for building an integrated energy system load prediction model according to an embodiment of the present invention specifically includes: the system comprises a packet preprocessing module, an analysis module and an establishment module, so that a multi-energy load prediction result can be obtained. Specifically, the method comprises the following steps:
the preprocessing module 60 is used for preprocessing the collected load historical data of the comprehensive energy system; the preprocessing module 60 is specifically configured to:
carrying out data cleaning and normalization processing on the collected comprehensive energy system load historical data;
the influence factor analysis module 62 is configured to perform multidimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result; the influence factor analysis module 62 is specifically configured to:
analyzing the time correlation of the multi-energy load of the comprehensive energy system, calculating the correlation between the load value of each type of load at the t moment and the load value at the t-d moment by adopting the Pearson correlation coefficient shown in formula 1, setting a proper threshold value, and setting a correlation index rho according to the correlation index rhoXYIs largeSmall to determine the time step for each load input and output:
where ρ isXYRepresenting the correlation between the load value at the time t and the load value at the time t-d for each type of load, XjRepresenting a sequence of load values at time t,representing the mean value of the series of load values at time t, YjRepresenting the sequence of load values at time t-d,the average value of the load value sequences at the t-d moments is shown, and N represents the lengths of the load value sequences at the t-d moments and the t-d moments;
analyzing the correlation between the multi-energy load of the comprehensive energy system and external environment factors, and calculating by using a Pearson correlation coefficient shown in formula 1, wherein the external environment influencing factors specifically include: setting proper threshold value according to air temperature and air pressure and according to correlation index rhoXYSelecting external environment factors with larger relevance as a part of model input;
and the establishing and predicting model module 64 is used for establishing a comprehensive energy system multi-energy load predicting model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load predicting model is used for predicting the comprehensive energy system load. The model for establishing a prediction model module 64 is specifically configured to:
according to the analysis result, local features are extracted through a CNN network model, time information is extracted through a BilSTM network model, weights are distributed through an attention mechanism model, optimization of the weights is carried out through an Adam algorithm, and a comprehensive energy system multi-energy load prediction model is established. Specifically, the method comprises the following steps:
extracting local dependencies between variables in the time dimension through CNN networks: wherein the convolutional layer of the CNN network consists of a plurality of cores, sweeping the kth filter through the input matrix, as represented:
hk=ReLU(Wk*X+bk) Formula 2;
where denotes the convolution operation, output hkA vector representing the output, the output of the convolutional layer being of a size dcA matrix of x T, wherein dcRepresenting the number of filters and T representing the time dimension. ReLU activation function ReLU (x) max (x,0), WkRepresenting the slope of the kth filter, bkDenotes the intercept of the kth filter, X is the input matrix, X ∈ RD×TD represents a multivariate variable dimension, and T represents a time dimension;
given an input sequence X ═ X1,x2,...,xT) Wherein x ist∈RnN is the dimension of the factor, and the hidden state at time t is ht∈RmAnd m is the dimension of the hidden state, and the input gate, the forgetting gate, the output gate, the memory unit updating and the output calculation in the LSTM model are carried out according to the formulas 3-7:
it=sigmoid(Wi[ht-1;xt]+bi) Formula 3;
ft=sigmoid(Wf[ht-1;xt]+bf) Formula 4;
ot=sigmoid(Wo[ht-1;xt]+bo) Equation 5;
st=ft⊙st-1+it⊙tanh(Ws[ht-1;xt]+bs) Equation 6;
ht=ot⊙tanh(st) Equation 7;
wherein [ ht-1;xt]∈Rm+nRepresenting a previous hidden state ht-1And the current input xtW is connected in seriesi,Wf,Wo,Ws∈Rm *(m+n)Representing a weight matrix, bi,bf,bo,bs∈RmRepresents an offset vector parameter, <' > represents an element-by-element multiplication, Rm+nRepresenting a matrix of dimension m + n, Rm*(m+n)Representing a matrix of dimensions m x (m + n), RmA matrix of dimension m is represented;
the observations are predicted by processing the BilSTM network model in two directions: forward LSTM units generate the preceding information, i.e. produce hidden statesWhereinGenerating subsequent information to the LSTM unit, i.e. producing hidden statesWhereinConcatenating the two preceding and succeeding information to produce a joint representation as shown in equation 8:
wherein,is the last hidden state of the forward LSTM unit,is the first hidden state of the backward LSTM unit;
the importance of the hidden states is distinguished by assigning different attention weights to the hidden states: given hiRepresenting the hidden state sequence of the BilSTM layer, h is expressed according to equation 9iThrough full connectionLayer-by-layer conversion to attention weight of target:
ui=tan(W*hi+ b) equation 9;
wherein b represents the intercept and W represents the slope;
generating probability vector p of weights based on softmax function according to equation 9i:
Wherein u isvRepresents the vth attention weight, and M represents the number of attention weights;
assigning the generated attention weight to the corresponding hidden layer state hiH is performed according to equation 11iThe weighted average of (a) is calculated to obtain an attention scoring function s as follows:
s=∑ipi*hiequation 11;
and iteratively updating the weight according to the training data through an Adam algorithm, and optimizing the weight.
The embodiment of the present invention is an apparatus embodiment corresponding to the above method embodiment, and specific operations of each module may be understood with reference to the description of the method embodiment, which is not described herein again.
Device embodiment II
The embodiment of the present invention provides a device for establishing a comprehensive energy system load prediction model, as shown in fig. 7, including: a memory 70, a processor 72 and a computer program stored on the memory 70 and executable on the processor 72, which computer program when executed by the processor 72 performs the steps as described in the method embodiments.
Device embodiment III
An embodiment of the present invention provides a computer-readable storage medium, on which an implementation program for information transmission is stored, and the program, when executed by a processor 72, implements the steps as described in the method embodiment.
The computer-readable storage medium of this embodiment includes, but is not limited to: ROM, RAM, magnetic or optical disks, and the like.
The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
In the 30 s of the 20 th century, improvements in a technology could clearly be distinguished between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Hardware Description Language), traffic, pl (core universal Programming Language), HDCal (jhdware Description Language), lang, Lola, HDL, laspam, hardward Description Language (vhr Description Language), vhal (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.
The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.
The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functions of the units may be implemented in the same software and/or hardware or in multiple software and/or hardware when implementing the embodiments of the present description.
One skilled in the art will recognize that one or more embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The description has been presented with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the description. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
One or more embodiments of the present description may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only an example of this document and is not intended to limit this document. Various modifications and changes may occur to those skilled in the art from this document. Any modifications, equivalents, improvements, etc. which come within the spirit and principle of the disclosure are intended to be included within the scope of the claims of this document.
Claims (10)
1. A method for building a comprehensive energy system load prediction model is characterized by comprising the following steps:
preprocessing the collected load historical data of the comprehensive energy system;
carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result;
and establishing a comprehensive energy system multi-energy load prediction model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load prediction model is used for predicting the comprehensive energy system load.
2. The method according to claim 1, wherein preprocessing the collected integrated energy system load history data specifically comprises:
and carrying out data cleaning and normalization processing on the collected comprehensive energy system load historical data.
3. The method according to claim 1, wherein the multi-dimensional influence factor analysis of the comprehensive energy system multi-energy load prediction is performed on the preprocessed comprehensive energy system load historical data to obtain an analysis result, and the method specifically comprises:
analyzing the time correlation of the multi-energy load of the comprehensive energy system, calculating the correlation between the load value of each type of load at the t moment and the load value at the t-d moment by adopting the Pearson correlation coefficient shown in formula 1, setting a proper threshold value, and setting a correlation index rho according to the correlation index rhoXYTo determine the time step for each load input and output:
where ρ isXYRepresenting the correlation between the load value at the time t and the load value at the time t-d for each type of load, XjRepresenting a sequence of load values at time t,is shown asMean value of the series of load values at time t, YjRepresenting the sequence of load values at time t-d,the average value of the load value sequences at the t-d moments is shown, and N represents the lengths of the load value sequences at the t-d moments and the t-d moments;
analyzing the correlation between the multi-energy load of the comprehensive energy system and external environment factors, and calculating by using a Pearson correlation coefficient shown in formula 1, wherein the external environment influencing factors specifically include: setting proper threshold value according to air temperature and air pressure and according to correlation index rhoXYAnd selecting external environment factors with larger relevance as a part of model input.
4. The method of claim 1, wherein the step of establishing the multi-energy load prediction model of the integrated energy system through the CNN-BiLSTM attention mechanism network according to the analysis result specifically comprises:
according to the analysis result, local features are extracted through a CNN network model, time information is extracted through a BilSTM network model, weights are distributed through an attention mechanism model, optimization of the weights is carried out through an Adam algorithm, and a comprehensive energy system multi-energy load prediction model is established.
5. The method of claim 4, wherein extracting local features through a CNN network model, extracting time information through a BilSTM network model, assigning weights through an attention mechanism model, and optimizing through an Adam algorithm specifically comprises:
extracting local dependencies between variables in the time dimension through CNN networks: wherein the convolutional layer of the CNN network consists of a plurality of cores, sweeping the kth filter through the input matrix, as represented:
hk=ReLU(Wk*X+bk) Formula 2;
where denotes the convolution operation, output hkA vector representing the output, the output of the convolutional layer being of a sizeIs dcA matrix of x T, wherein dcRepresenting the number of filters and T representing the time dimension. ReLU activation function ReLU (x) max (x,0), WkRepresenting the slope of the kth filter, bkDenotes the intercept of the kth filter, X is the input matrix, X ∈ RD×TD represents a multivariate variable dimension, and T represents a time dimension;
given an input sequence X ═ X1,x2,…,xT) Wherein x ist∈RnN is the dimension of the factor, and the hidden state at time t is ht∈RmAnd m is the dimension of the hidden state, and the input gate, the forgetting gate, the output gate, the memory unit updating and the output calculation in the LSTM model are carried out according to the formulas 3-7:
it=sigmoid(Wi[ht-1;xt]+bi) Formula 3;
ft=sigmoid(Wf[ht-1;xt]+bf) Formula 4;
ot=sigmoid(Wo[ht-1;xt]+bo) Equation 5;
st=ft⊙st-1+it⊙tanh(Ws[ht-1;xt]+bs) Equation 6;
ht=ot⊙tanh(st) Equation 7;
wherein [ ht-1;xt]∈Rm+nRepresenting a previous hidden state ht-1And the current input xtW is connected in seriesi,Wf,Wo,Ws∈Rm*(m+n)Representing a weight matrix, bi,bf,bo,bs∈RmRepresents an offset vector parameter, <' > represents an element-by-element multiplication, Rm+nRepresenting a matrix of dimension m + n, Rm*(m+n)Representing a matrix of dimensions m x (m + n), RmA matrix of dimension m is represented;
the observations are predicted by processing the BilSTM network model in two directions: forward LSTM Unit Generation ForwardSurface information, i.e. producing hidden statesWhereinGenerating subsequent information to the LSTM unit, i.e. producing hidden statesWhereinConcatenating the two preceding and succeeding information to produce a joint representation as shown in equation 8:
wherein,is the last hidden state of the forward LSTM unit,is the first hidden state of the backward LSTM unit;
the importance of the hidden states is distinguished by assigning different attention weights to the hidden states: given hiRepresenting the hidden state sequence of the BilSTM layer, h is expressed according to equation 9iTranslation to attention weight of target by fully connected layer:
ui=tan(W*hi+ b) equation 9;
wherein b represents the intercept and W represents the slope;
generating probability vector p of weights based on softmax function according to equation 9t:
Wherein u isvRepresents the vth attention weight, and M represents the number of attention weights;
assigning the generated attention weight to the corresponding hidden layer state hiH is performed according to equation 11iThe weighted average of (a) is calculated to obtain an attention scoring function s as follows:
s=∑lpl*hlequation 11;
and iteratively updating the weight according to the training data through an Adam algorithm, and optimizing the weight.
6. An integrated energy system load prediction model building device is characterized by comprising the following components:
the preprocessing module is used for preprocessing the collected comprehensive energy system load historical data;
the influence factor analysis module is used for carrying out multi-dimensional influence factor analysis of comprehensive energy system multi-energy load prediction on the preprocessed comprehensive energy system load historical data to obtain an analysis result;
and the model establishing and predicting module is used for establishing a comprehensive energy system multi-energy load predicting model through a CNN-BilSTM attention mechanism network according to the analysis result, wherein the comprehensive energy system multi-energy load predicting model is used for predicting the comprehensive energy system load.
7. The apparatus of claim 6,
the preprocessing module is specifically configured to:
carrying out data cleaning and normalization processing on the collected comprehensive energy system load historical data;
the influence factor analysis module is specifically configured to:
analyzing the time correlation of the multi-energy load of the comprehensive energy system, and calculating each type of negative load by adopting the Pearson correlation coefficient shown in the formula 1Setting a proper threshold value according to the correlation between the load value at the t-th moment and the load value at the t-d moment, and according to the correlation index rhoXYTo determine the time step for each load input and output:
where ρ isXYRepresenting the correlation between the load value at the time t and the load value at the time t-d for each type of load, XjRepresenting a sequence of load values at time t,representing the mean value of the series of load values at time t, YjRepresenting the sequence of load values at time t-d,the average value of the load value sequences at the t-d moments is shown, and N represents the lengths of the load value sequences at the t-d moments and the t-d moments;
analyzing the correlation between the multi-energy load of the comprehensive energy system and external environment factors, and calculating by using a Pearson correlation coefficient shown in formula 1, wherein the external environment influencing factors specifically include: setting proper threshold value according to air temperature and air pressure and according to correlation index rhoXYSelecting external environment factors with larger relevance as a part of model input;
the model for establishing a predictive model is specifically configured to:
according to the analysis result, local features are extracted through a CNN network model, time information is extracted through a BilSTM network model, weights are distributed through an attention mechanism model, optimization of the weights is carried out through an Adam algorithm, and a comprehensive energy system multi-energy load prediction model is established.
8. The apparatus of claim 7, wherein the establishing module is specifically configured to:
extracting local dependencies between variables in the time dimension through CNN networks: wherein the convolutional layer of the CNN network consists of a plurality of cores, sweeping the kth filter through the input matrix, as represented:
hk-PoLU(Wk*X+bk) Formula 2;
where denotes the convolution operation, output hkA vector representing the output, the output of the convolutional layer being of a size dcA matrix of x T, wherein dcRepresenting the number of filters and T representing the time dimension. ReLU activation function ReLU (x) max (x,0), WkRepresenting the slope of the kth filter, bkDenotes the intercept of the kth filter, X is the input matrix, X ∈ RD×TD represents a multivariate variable dimension, and T represents a time dimension;
given an input sequence X ═ X1,x2,…,xT) Wherein x ist∈RnN is the dimension of the factor, and the hidden state at time t is ht∈RmAnd m is the dimension of the hidden state, and the input gate, the forgetting gate, the output gate, the memory unit updating and the output calculation in the LSTM model are carried out according to the formulas 3-7:
it=sigmoid(Wi[ht-1;xt]+bl) Formula 3;
ft=sigmoid(Wf[ht-1;xt]+bf) Formula 4;
Ot=slgmold(Wo[ht-1;xt]+bo) Equation 5;
st=ft⊙st-1+it⊙tanh(Ws[ht-1;xt]+bs) Equation 6;
ht=ot⊙tanh(st) Equation 7;
wherein [ ht-1;xt]∈Rm+nRepresenting a previous hidden state ht-1And the current input xtW is connected in seriesi,Wf,Wo,Ws∈Rm*(m+n)Representing a weight matrix, bi,bf,bo,bs∈RmRepresents an offset vector parameter, <' > represents an element-by-element multiplication, Rm+nRepresenting a matrix of dimension m + n, Rm*(m+n)Representing a matrix of dimensions m x (m + n), RmA matrix of dimension m is represented;
the observations are predicted by processing the BilSTM network model in two directions: forward LSTM units generate the preceding information, i.e. produce hidden statesWhereinGenerating subsequent information to the LSTM unit, i.e. producing hidden statesWhereinConcatenating the two preceding and succeeding information to produce a joint representation as shown in equation 8:
wherein,is the last hidden state of the forward LSTM unit,is the first hidden state of the backward LSTM unit;
the importance of the hidden states is distinguished by assigning different attention weights to the hidden states: given hiRepresenting the hidden state sequence of the BilSTM layer, h is expressed according to equation 9iTranslation to attention weight of target by fully connected layer:
ui=tan(W*hi+ b) equation 9;
wherein b represents the intercept and W represents the slope;
generating probability vector p of weights based on softmax function according to equation 9i:
Wherein u isvRepresents the vth attention weight, and M represents the number of attention weights;
assigning the generated attention weight to the corresponding hidden layer state hiH is performed according to equation 11iThe weighted average of (a) is calculated to obtain an attention scoring function s as follows:
s=∑ipi*hiequation 11;
and iteratively updating the weight according to the training data through an Adam algorithm, and optimizing the weight.
9. An integrated energy system load prediction model building device is characterized by comprising the following components: memory, a processor and a computer program stored on the memory and executable on the processor, the computer program when executed by the processor implementing the steps of the integrated energy system load prediction model building method according to any one of claims 1 to 5.
10. A computer-readable storage medium, having stored thereon a program for implementing information transfer, which when executed by a processor implements the steps of the integrated energy system load prediction model building method according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111119840.6A CN113822482B (en) | 2021-09-24 | 2021-09-24 | Method and device for establishing comprehensive energy system load prediction model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111119840.6A CN113822482B (en) | 2021-09-24 | 2021-09-24 | Method and device for establishing comprehensive energy system load prediction model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113822482A true CN113822482A (en) | 2021-12-21 |
CN113822482B CN113822482B (en) | 2024-07-02 |
Family
ID=78921278
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111119840.6A Active CN113822482B (en) | 2021-09-24 | 2021-09-24 | Method and device for establishing comprehensive energy system load prediction model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113822482B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114677022A (en) * | 2022-03-31 | 2022-06-28 | 南通电力设计院有限公司 | Distributed management method and system for multi-element fusion energy |
CN114997488A (en) * | 2022-05-27 | 2022-09-02 | 中国石油大学(北京) | Comprehensive energy system load prediction method and device |
CN115330040A (en) * | 2022-08-05 | 2022-11-11 | 江苏润和软件股份有限公司 | Deep learning-based comprehensive energy distributed wind power generation prediction method and system |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109886498A (en) * | 2019-03-01 | 2019-06-14 | 北京邮电大学 | A kind of EMD-GRU Short-Term Load Forecasting Method based on feature selecting |
US20190347550A1 (en) * | 2018-05-14 | 2019-11-14 | Samsung Electronics Co., Ltd. | Method and apparatus with neural network parameter quantization |
CN111260147A (en) * | 2020-02-07 | 2020-06-09 | 河北工程大学 | Power generation capacity prediction method and device and terminal equipment |
CN112035745A (en) * | 2020-09-01 | 2020-12-04 | 重庆大学 | Recommendation algorithm based on counterstudy and bidirectional long-short term memory network |
CN112465385A (en) * | 2020-12-09 | 2021-03-09 | 武汉理工大学 | Demand response potential analysis method applying intelligent electric meter data |
CN112529283A (en) * | 2020-12-04 | 2021-03-19 | 天津天大求实电力新技术股份有限公司 | Comprehensive energy system short-term load prediction method based on attention mechanism |
-
2021
- 2021-09-24 CN CN202111119840.6A patent/CN113822482B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190347550A1 (en) * | 2018-05-14 | 2019-11-14 | Samsung Electronics Co., Ltd. | Method and apparatus with neural network parameter quantization |
CN109886498A (en) * | 2019-03-01 | 2019-06-14 | 北京邮电大学 | A kind of EMD-GRU Short-Term Load Forecasting Method based on feature selecting |
CN111260147A (en) * | 2020-02-07 | 2020-06-09 | 河北工程大学 | Power generation capacity prediction method and device and terminal equipment |
CN112035745A (en) * | 2020-09-01 | 2020-12-04 | 重庆大学 | Recommendation algorithm based on counterstudy and bidirectional long-short term memory network |
CN112529283A (en) * | 2020-12-04 | 2021-03-19 | 天津天大求实电力新技术股份有限公司 | Comprehensive energy system short-term load prediction method based on attention mechanism |
CN112465385A (en) * | 2020-12-09 | 2021-03-09 | 武汉理工大学 | Demand response potential analysis method applying intelligent electric meter data |
Non-Patent Citations (3)
Title |
---|
LEONARDO INTESA ET AL: "SPEED:Open-source framework to accelerate speech recognition on embedded GPUs", 《IEEE 2017 EUROMICRO CONFERENCE ON DIGITAL SYSTEM DESIGN(DSD)》, pages 94 - 101 * |
乔博鑫;章坚民;: "基于历史负荷的配电网量测数据关联清洗方法研究", 信息与电脑(理论版), no. 08, pages 75 - 76 * |
唐伟斌;李涛;邹品晶;曾建平;向仲卿;胡斯佳;: "基于皮尔逊相关系数的电网夏季空调负荷预测研究", 电力需求侧管理, no. 06, pages 11 - 17 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114677022A (en) * | 2022-03-31 | 2022-06-28 | 南通电力设计院有限公司 | Distributed management method and system for multi-element fusion energy |
CN114677022B (en) * | 2022-03-31 | 2023-09-15 | 南通电力设计院有限公司 | Multi-element fusion energy distributed management method and system |
CN114997488A (en) * | 2022-05-27 | 2022-09-02 | 中国石油大学(北京) | Comprehensive energy system load prediction method and device |
CN115330040A (en) * | 2022-08-05 | 2022-11-11 | 江苏润和软件股份有限公司 | Deep learning-based comprehensive energy distributed wind power generation prediction method and system |
Also Published As
Publication number | Publication date |
---|---|
CN113822482B (en) | 2024-07-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | A multiple time series-based recurrent neural network for short-term load forecasting | |
Ma et al. | A hybrid attention-based deep learning approach for wind power prediction | |
CN113822482A (en) | Method and device for establishing load prediction model of comprehensive energy system | |
Chen et al. | A weighted LS-SVM based learning system for time series forecasting | |
Yu et al. | A novel elastic net-based NGBMC (1, n) model with multi-objective optimization for nonlinear time series forecasting | |
CN112529283A (en) | Comprehensive energy system short-term load prediction method based on attention mechanism | |
Lin et al. | Temporal convolutional attention neural networks for time series forecasting | |
Islam et al. | LSTM-based electrical load forecasting for Chattogram city of Bangladesh | |
CN111027672A (en) | Time sequence prediction method based on interactive multi-scale recurrent neural network | |
Girimurugan et al. | Application of deep learning to the prediction of solar irradiance through missing data | |
Elhariri et al. | H-ahead multivariate microclimate forecasting system based on deep learning | |
Liu et al. | Incremental Bayesian broad learning system and its industrial application | |
CN110738363B (en) | Photovoltaic power generation power prediction method | |
CN115270640A (en) | Configuration optimization method for distributed photovoltaic power generation system of microgrid and terminal equipment | |
Syu et al. | Energy grid management system with anomaly detection and Q-learning decision modules | |
Tang et al. | Forecasting annual foreign tourist arrivals to China by incorporating firefly algorithm into fractional non-homogenous discrete Grey model | |
Sang et al. | Ensembles of gradient boosting recurrent neural network for time series data prediction | |
CN116993185B (en) | Time sequence prediction method, device, equipment and storage medium | |
Wang et al. | A transformer-based multi-entity load forecasting method for integrated energy systems | |
CN112330044A (en) | Support vector regression model based on iterative aggregation grid search algorithm | |
Roh et al. | Tfe-net: time and feature focus embedding network for multivariate-to-multivariate time series forecasting | |
Aksoy et al. | Design of Graphical User Interface for Artificial Intelligence-Based Energy Management System for Microgrids. | |
Zhao et al. | Photovoltaic maximum power point tracking based on IWD-SVM | |
Lee et al. | Application of temperature prediction model based on LSTNet in telecommunication Room | |
CN116526582B (en) | Combined dispatching method and system for electric power unit based on artificial intelligence combined driving |
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 |