CN115796382A - Regional heating load prediction method, device, equipment and storage medium - Google Patents
Regional heating load prediction method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN115796382A CN115796382A CN202211636175.2A CN202211636175A CN115796382A CN 115796382 A CN115796382 A CN 115796382A CN 202211636175 A CN202211636175 A CN 202211636175A CN 115796382 A CN115796382 A CN 115796382A
- Authority
- CN
- China
- Prior art keywords
- heating load
- data
- meteorological
- historical
- time sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a regional heating load prediction method, a regional heating load prediction device, regional heating load prediction equipment and a storage medium, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period; after the heating load time sequence data and the meteorological time sequence data are preprocessed, inputting a pre-trained regional heating load prediction model to obtain predicted heating load data of a target region in a future set time period; the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region. This application can obtain the space-time relation between heating load volume and the meteorological change through fusing meteorological data, compares with traditional heating load prediction, can improve regional heating load prediction's accuracy nature.
Description
Technical Field
The application relates to the field of artificial intelligence, in particular to a method and a device for predicting a regional heating load, electronic equipment and a storage medium.
Background
In China, the operation energy consumption of buildings accounts for about 30% of the total energy consumption, wherein the heating of northern cities consumes about 21% of the total energy consumption of buildings. Therefore, urban heating needs to be carried out with full force to implement a feasible energy-saving and emission-reduction scheme to match the achievement of the targets. The intelligent regional heating system is an important way for realizing green energy-saving and comfortable heating in the future, and is beneficial to improving the energy utilization efficiency and reducing the pollution emission.
The traditional heating prediction technology mainly counts the heating load data of all users in an area to perform direct prediction. However, this method has the following disadvantages: the heating demands and the use time periods of different regions are different, and the traditional method can only predict the total demand but cannot realize the demand distribution to a certain region; a large number of researches show that the heating load is in positive correlation with the meteorological change, and the traditional method cannot accurately predict the heating load change caused by the meteorological change.
It is therefore desirable to provide a method for accurately predicting district heating load.
Disclosure of Invention
The application aims to provide a regional heating load prediction method, a regional heating load prediction device, electronic equipment and a storage medium, so that the accuracy of regional heating load prediction is improved.
In a first aspect, an embodiment of the present application provides a district heating load prediction method, including:
acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period;
after the heating load time sequence data and the meteorological time sequence data are preprocessed, inputting a pre-trained regional heating load prediction model to obtain predicted heating load data of a target region in a future set time period; the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
In some embodiments of the present application, the district heating load prediction model is pre-trained in the following manner:
acquiring historical heating load data and historical meteorological data in a target area, and counting the historical heating load data and the historical meteorological data into time sequences in the same time interval to obtain a historical heating load time sequence data sample and a historical meteorological time sequence data sample of the target area;
after the historical heating load time sequence data samples and the historical meteorological time sequence data samples are preprocessed, the historical heating load time sequence data samples and the historical meteorological time sequence data samples are respectively converted into heating load supervised data and meteorological supervised data;
and setting an objective function, inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
In some embodiments of the present application, said converting said historical heating load time series data samples and historical weather time series data samples into heating load supervised data and weather supervised data, respectively, comprises:
taking the former variables in the historical heating load time sequence data sample as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain heating load supervised data;
and taking the former variables in the historical meteorological time series data samples as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain meteorological supervision data.
In some embodiments of the present application, the preprocessing includes maximum processing, minimum processing, data averaging, and normalization processing.
In some embodiments of the present application, the time series neural network comprises a weather timing representation component, a heating load timing representation component, and an autoregressive representation component;
the weather timing representing means includes: a first convolution representation layer and a first autoregressive representation layer;
the heating load time sequence expression means includes: a second convolution representation layer, a graph attention representation layer, a GRU neural network, a linear layer and a second autoregressive representation layer;
inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition, wherein the method comprises the following steps:
and inputting the heating load supervised data into the heating load time sequence representation component, inputting the weather supervised data into the weather time sequence representation component, training, adjusting parameters in the time sequence neural network, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
In some embodiments of the present application, the objective function is set to be a mean square error loss function.
In a second aspect, an embodiment of the present application provides a district heating load prediction device, including:
the acquisition module is used for acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period;
the prediction module is used for inputting a pre-trained regional heating load prediction model after preprocessing the heating load time sequence data and the meteorological time sequence data to obtain predicted heating load data of a target region in a future set time period;
the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
In some embodiments of the present application, the apparatus further comprises: the model training module is used for pre-training the district heating load prediction model according to the following modes:
acquiring historical heating load data and historical meteorological data in a target area, and counting the historical heating load data and the historical meteorological data into time sequences in the same time interval to obtain a historical heating load time sequence data sample and a historical meteorological time sequence data sample of the target area;
after the historical heating load time sequence data samples and the historical meteorological time sequence data samples are preprocessed, the historical heating load time sequence data samples and the historical meteorological time sequence data samples are respectively converted into heating load supervised data and meteorological supervised data;
and setting an objective function, inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
In some embodiments of the present application, the model training module is specifically configured to:
taking the former variables in the historical heating load time sequence data sample as input variables of the model, and taking the latter variables as the output of the model for polymerization to obtain heating load supervised data;
and taking the former variables in the historical meteorological time sequence data samples as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain meteorological supervised data.
In some embodiments of the present application, the preprocessing includes maximum processing, minimum processing, data averaging, and normalization processing.
In some embodiments of the present application, the time series neural network comprises a weather timing representation component, a heating load timing representation component, and an autoregressive representation component;
the weather timing representing means includes: a first convolution representation layer and a first autoregressive representation layer;
the heating load time sequence expression means includes: a second convolution representation layer, a graph attention representation layer, a GRU neural network, a linear layer and a second autoregressive representation layer;
the model training module is specifically configured to:
and inputting the heating load supervised data into the heating load time sequence representation component, inputting the weather supervised data into the weather time sequence representation component, training, adjusting parameters in the time sequence neural network, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
In some embodiments of the present application, the objective function is set to be a mean square error loss function.
In a third aspect, the present application provides an electronic device, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing when executing the computer program to implement the method according to the first aspect.
In a fourth aspect, the present application provides a computer readable storage medium having computer readable instructions stored thereon, the computer readable instructions being executable by a processor to implement the method according to the first aspect.
Compared with the prior art, the regional heating load prediction method provided by the application obtains the heating load time sequence data and the meteorological time sequence data of the target region in the historical time period; after the heating load time sequence data and the meteorological time sequence data are preprocessed, inputting a pre-trained regional heating load prediction model to obtain predicted heating load data of a target region in a future set time period; the regional heating load prediction model is obtained by training a time series neural network through a historical heating load time series data sample and a historical meteorological time series data sample of a target region, so that when the heating load data of the target region is predicted, the heating load and the meteorological phenomena of the target region at different moments in a historical time period are considered at the same time, the space-time relation between the heating load and the meteorological changes can be obtained by fusing the meteorological data, and compared with the traditional heating load prediction, the accuracy of regional heating load prediction can be improved.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the application. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
FIG. 1 is a flow chart illustrating a district heating load prediction method according to an embodiment of the present application;
FIG. 2 is a flowchart illustrating a district heating load prediction model training method according to an embodiment of the present disclosure;
FIG. 3 is a schematic diagram illustrating a prediction process of a district heating load prediction model according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a district heating load prediction device according to an embodiment of the present application;
fig. 5 shows a schematic diagram of an electronic device provided in an embodiment of the present application.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
It is to be noted that, unless otherwise specified, technical or scientific terms used herein shall have the ordinary meaning as understood by those skilled in the art to which this application belongs.
In addition, the terms "first" and "second", etc. are used to distinguish different objects, and are not used to describe a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Referring to fig. 1, fig. 1 is a flowchart of a district heating load prediction method according to an embodiment of the present application, including the following steps S101 to S102:
s101, heating load time sequence data and meteorological time sequence data of a target area in a historical time period are obtained.
The heating load time-series data is time-series data in which a plurality of heating load data of a target area are configured at set time intervals, and the weather time-series data is time-series data in which a plurality of weather data are configured at set time intervals.
Specifically, the time interval of the plurality of meteorological data and the time interval of the plurality of heating load data need to be kept consistent. The time interval may be set to a specific time interval of days, weeks, months, etc. Meteorological data includes temperature, barometric pressure, precipitation, etc.
And S102, after the heating load time sequence data and the meteorological time sequence data are preprocessed, inputting a pre-trained regional heating load prediction model to obtain predicted heating load data of a target region in a future set time period.
The regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
Illustratively, the operation of preprocessing includes: maximum processing, minimum processing, data averaging and normalization processing.
After the heating load time sequence data and the meteorological time sequence data are preprocessed, a pre-trained regional heating load prediction model is input, and predicted heating load data of a target region in a future set time period are obtained.
How to pre-train the district heating load prediction model will be described below, specifically, the district heating load prediction model may be pre-trained in the following manner, as shown in fig. 2, and includes steps S201 to S203:
s201, obtaining historical heating load data and historical meteorological data in a target area, and counting the historical heating load data and the historical meteorological data into time sequences in the same time interval to obtain historical heating load time sequence data samples and historical meteorological time sequence data samples of the target area.
The method comprises the steps of collecting historical meteorological data and historical heating load data of a target area, setting a time interval, and counting the historical heating load data and the historical meteorological data into a time sequence in the same time interval to obtain a historical heating load time sequence data sample and a historical meteorological time sequence data sample.
Specifically, to ensure that the time dimensions of the meteorological factors and the heating load factors are consistent, a time interval length T is first determined, and T may be set to be a specific time interval of days, weeks, months, and the like. Will { t be processed through preliminary data preprocessing operations such as maximum processing, minimum processing, data averaging, and the like 1 ,t 2 ,…,t n And integrating data in the time interval into statistical data at a specific calibration time, and traversing data samples to generate an initial input sequence.
At { t 1 ,t 2 ,…,t n And (4) counting heating load data of the area and various meteorological data in the area, such as temperature, air pressure, precipitation and the like in the time interval. With e T ＝{e 1 ,e 2 ,…,e T Represents a time-series sequence of a particular set of meteorological data variables e over T time intervals. By using
An input set of outer vectors representing the sequences of M meteorological features at time t. By x T ＝{x 1 ,x 2 ,…,x T Indicates the time sequence of the historical heating load variable x of a specific area within T time intervals.
The input set of heating load data for D zones at time t is shown.
S202, after the historical heating load time sequence data samples and the historical meteorological time sequence data samples are preprocessed, the historical heating load time sequence data samples and the historical meteorological time sequence data samples are respectively converted into heating load supervised data and meteorological supervised data.
Specifically, because meteorological data and heating load data have obvious data range difference, the application normalizes the data by using a Min-Max function and compresses the data to an interval [0,1] so as to accelerate the training speed of the model.
The formula for normalization and denormalization is as follows:
z＝z′(max(z)-min(z))+min(z)；
wherein the content of the first and second substances,
represents a set of samples, N represents the number of samples observed, z' represents the normalized data, and min (-) and max (-) represent the minimum and maximum values of the input vector, respectively. And applying an inverse normalization formula to the predicted value output by the reduction model to be the predicted heating load value.
Specifically, in S202, the converting the historical heating load time series data sample and the historical meteorological time series data sample into heating load supervised data and meteorological supervised data respectively includes:
taking the former variables in the historical heating load time sequence data sample as input variables of the model, and taking the latter variables as the output of the model for polymerization to obtain heating load supervised data;
and taking the former variables in the historical meteorological time sequence data samples as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain meteorological supervised data.
Specifically, in order to learn the time sequence characteristics from the time sequence, the application converts the time sequence data into supervised data. Given a time series of data, the first variables are used as input variables of the model and the last variables are aggregated as output of the model in the conversion process.
By collecting various meteorological data and heating load data of each family in the area and converting the meteorological data and the heating load data into supervised data, the supervised data is used for training a neural network, and finally, the trained model is used for providing more accurate prediction.
S203, setting a target function, inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
Specifically, the objective function may be set as a Mean Squared Error (MSE) loss function.
Specifically, the time series neural network comprises a weather time sequence representation component, a heating load time sequence representation component and an autoregressive representation component; the weather timing representing means includes: a first convolution representation layer and a first autoregressive representation layer; the heating load time sequence expression means includes: a second convolution representation layer, a graph attention representation layer, a GRU neural network, a linear layer, and a second autoregressive representation layer.
Referring to fig. 3, the district heating load prediction model will be described in detail.
The weather timing representing part includes: a Convolutional (CNN) representation layer and an Autoregressive (AR) representation layer. The CNN is used for processing input meteorological data and capturing non-linear correlation characteristics among the meteorological data, and the autoregressive representation layer is used for fusing the non-linear characteristics. The formula for converting meteorological data is as follows:
R a ＝W a ·(∑ T W e ·(W c *E))+B a ；
wherein R is a Representing a non-linear representation of meteorological data, W a Representing an autoregressive weight matrix, B a Denotes the offset, W e Representing a convolutional network weight matrix, W c Represents a convolution kernel weight matrix, representing a convolution operation.
The heating load sequence display unit includes: convolution representation layer (CNN), graph attention representation layer, GRU (Gated current Unit) neural network, linear layer, autoregressive representation layer, not all layers shown in fig. 3. The CNN is used to capture the non-linear relationship between different heating time sequences and map the time sequence data into a graph structure, and the transformation process of the graph structure to generate the ith characteristic diagram through convolution is shown as the following formula:
wherein, the first and the second end of the pipe are connected with each other,
is the output of the convolutional layer, C is the number of feature maps,
the weights of the convolution kernels representing the ith feature map,
is the bias term. The above feature graph can be regarded as a graph structure with a node number of T and an edge number of C, and the symbol U e { U ∈ { U } 1 ,u 2 ,…,u T Is represented by, wherein
Representing the signature at the ith timestamp.
The graph attention force representation layer uses a graph attention force mechanism to capture dynamic changes between nodes at different times, and the conversion process between any two nodes can be represented by the following formula:
d j,k ＝σ(W d ·[W u ·u j ；W u ·u k ])；
wherein the content of the first and second substances,
representing the correlation matrix between node j and node k,
and
are all weight matrices, [;]represents a matrix splicing operation, where σ is the LeakyReLU activation function, with for each element x in σ (·):
where λ >1 is a constant.
And (3) scaling the output correlation matrix by using a Softmax normalization function to highlight key information in the output correlation matrix, wherein the conversion process is shown as the following formula:
where v ∈ U denotes a neighbor node of node j, α j,k Representing a normalized attention score of node k to node j. The input features are then transformed with the attention scores to obtain final output features for each node.
In order to more comprehensively and efficiently highlight the key timing characteristics, a multi-head attention mechanism is adopted to convert the correlation matrix, and the conversion process of the single-head attention mechanism is shown as follows:
a j,q ＝Sigmoid(∑ k∈U α j,k,q ·W a ·u k )；
wherein the content of the first and second substances,
representing the weighted features of node j and the qth attention mechanism, sigmoid (-) represents a Sigmoid function, with the following transformation for any element within the function:
and splicing the Q single-head attention weighted features together to obtain an attention weighted feature representation. The process is shown by the following formula:
A j ＝[a j,1 ；a j,2 ；…；a j,Q ]；
wherein A ∈ R B×T×(C*Q) Is the output of the graph convolution attention mechanism, A j ∈R B×1×(C*Q) Is a weighted characteristic representation of node j. The attention representation A is then transformed using a GRU neural network and a hidden representation of the learned timing characteristics is output
The hidden representation may be represented by a hidden representation of a previous update time instant
(initially 0) was calculated. This calculation is shown by the following equation:
wherein r is t ，u t ，c t Respectively generationTable reset gate, refresh gate, memory layer of GRU neural network at time t, <' > being the product of Adama, W r ，W u ，W c Is a weight matrix, b r ，b u ，b c Is an offset. Then the
Is transmitted to the linear layer to obtain the output R of the heating load time sequence representation component o The transformation process is as follows:
wherein the content of the first and second substances,
W o is a linear weight matrix, b o Is the offset coefficient.
Autoregressive representation the transformation process is shown by the following disclosure:
R l ＝∑ T W l ·X+b l ；
wherein R is l ∈R B×1×D Is the output of the autoregressive representation component, W l As a weight matrix, b l Is the bias factor.
The output of the final district heating load prediction model is obtained by adding the outputs of the three parts, and the conversion process is as follows:
O＝R a +R o +R l in which
Is the output of the model.
Setting model parameters such as an objective function, the size T of a time window, the quantity of characteristic graphs of CNN, a lambda value in an activation function sigma (-), a hidden state parameter H of a GRU neural network and the like, training the model, and reducing the model to output predicted regional heating load data by using an inverse normalization function. And adjusting the model parameters according to the result output by the model to obtain an optimal regional heating load prediction model.
In the application, the heating load supervision data is input into the heating load time sequence representation component, and meanwhile, the weather supervision data is input into the weather time sequence representation component for training, parameters in the time sequence neural network are adjusted, and the regional heating load prediction model is obtained after a preset training cut-off condition is reached. The model learns the space-time relation between the heating load and the meteorological change by fusing the meteorological data, captures the rules of the meteorological change and the heating load change, and is higher in accuracy compared with a traditional heating prediction model.
After the district heating load prediction model is obtained, the heating load time sequence data and the meteorological time sequence data of the target district are input into the heating load prediction model to predict the predicted heating load data of the target district in a set time period in the future, the district heating load prediction process can be shown in fig. 3, and after the prediction result is obtained, the prediction result can be provided for a heating system to make an accurate production plan so as to carry out accurate district heating.
The regional heating load prediction method provided by the embodiment of the application has the following beneficial effects:
according to the method, through fusion of meteorological data, the space-time relation between the heating load amount and the meteorological change is learned, the law of the meteorological change and the heating load amount change is captured, and compared with the traditional heating prediction, the accuracy is higher.
This application is through the time sequence characteristic of the heating load quantity time sequence that the graph attention force mechanism study that provides possessed different user habits, compares in traditional approach, and this application is stronger to heating load data's generalization ability.
In the above embodiment, a district heating load prediction method is provided, and the present application also provides a district heating load prediction apparatus 10. The district heating load prediction device according to the embodiment of the present application may implement the above-described district heating load prediction method, and the district heating load prediction device may be implemented by software, hardware, or a combination of software and hardware. For example, the district heating load prediction device may comprise integrated or separate functional modules or units to perform the corresponding steps in the methods described above.
Please refer to fig. 4, which includes:
the system comprises an acquisition module 101, a storage module and a processing module, wherein the acquisition module 101 is used for acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period;
the prediction module 102 is configured to input a pre-trained regional heating load prediction model after preprocessing the heating load time sequence data and the meteorological time sequence data, so as to obtain predicted heating load data of a target region in a future set time period;
the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
In some embodiments of the present application, the apparatus further comprises: the model training module is used for pre-training the district heating load prediction model according to the following modes:
acquiring historical heating load data and historical meteorological data in a target area, and counting the historical heating load data and the historical meteorological data into time sequences in the same time interval to obtain a historical heating load time sequence data sample and a historical meteorological time sequence data sample of the target area;
after the historical heating load time sequence data samples and the historical meteorological time sequence data samples are preprocessed, the historical heating load time sequence data samples and the historical meteorological time sequence data samples are respectively converted into heating load supervised data and meteorological supervised data;
and setting an objective function, inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
In some embodiments of the present application, the model training module is specifically configured to:
taking the former variables in the historical heating load time sequence data sample as input variables of the model, and taking the latter variables as the output of the model for polymerization to obtain heating load supervised data;
and taking the former variables in the historical meteorological time series data samples as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain meteorological supervision data.
In some embodiments of the present application, the preprocessing includes maximum processing, minimum processing, data averaging, and normalization processing.
In some embodiments of the present application, the time series neural network comprises a weather timing representation component, a heating load timing representation component, and an autoregressive representation component;
the weather timing representing means includes: a first convolution representation layer and a first autoregressive representation layer;
the heating load time sequence expression means includes: a second convolution representation layer, a graph attention representation layer, a GRU neural network, a linear layer, and a second autoregressive representation layer.
In some embodiments of the present application, the objective function is set to be a mean square error loss function.
The embodiment of the present application further provides an electronic device corresponding to the method provided in the foregoing embodiment, where the electronic device may be an electronic device for a client, such as a mobile phone, a notebook computer, a tablet computer, a desktop computer, and the like, to execute the method for predicting a district heating load.
Please refer to fig. 5, which illustrates a schematic diagram of an electronic device according to some embodiments of the present application. As shown in fig. 5, the electronic device 20 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, wherein the processor 200, the communication interface 203 and the memory 201 are connected through the bus 202; the memory 201 stores a computer program that can be executed on the processor 200, and the processor 200 executes the phishing mail tracing method provided by any one of the previous embodiments when executing the computer program.
The Memory 201 may include a high-speed Random Access Memory (RAM) and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used.
The processor 200 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 200. The Processor 200 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; but may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the method in combination with the hardware thereof.
The electronic device provided by the embodiment of the application and the district heating load prediction method provided by the embodiment of the application have the same inventive concept and have the same beneficial effects as the method adopted, operated or realized by the electronic device.
The present embodiment also provides a computer-readable storage medium corresponding to the district heating load prediction method provided in the foregoing embodiment, and having a computer program (i.e., a program product) stored thereon, where the computer program is executed by a processor to execute the district heating load prediction method provided in any of the foregoing embodiments.
It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, a phase change memory (PRAM), a Static Random Access Memory (SRAM), a Dynamic Random Access Memory (DRAM), other types of Random Access Memories (RAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a flash memory, or other optical and magnetic storage media, which are not described in detail herein.
The computer-readable storage medium provided by the above-mentioned embodiment of the present application has the same beneficial effects as the method adopted, operated or implemented by the application program stored in the computer-readable storage medium, based on the same inventive concept as the district heating load prediction method provided by the embodiment of the present application.
Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present disclosure, and the present disclosure should be construed as being covered by the claims and the specification.
Claims (10)
1. A district heating load prediction method, comprising:
acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period;
after the heating load time sequence data and the meteorological time sequence data are preprocessed, inputting a pre-trained regional heating load prediction model to obtain predicted heating load data of a target region in a future set time period;
the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
2. The method of claim 1, wherein the district heating load prediction model is pre-trained in the following manner:
acquiring historical heating load data and historical meteorological data in a target area, and counting the historical heating load data and the historical meteorological data into time sequences in the same time interval to obtain a historical heating load time sequence data sample and a historical meteorological time sequence data sample of the target area;
after the historical heating load time sequence data samples and the historical meteorological time sequence data samples are preprocessed, the historical heating load time sequence data samples and the historical meteorological time sequence data samples are respectively converted into heating load supervised data and meteorological supervised data;
and setting an objective function, inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
3. The method of claim 2, wherein said converting said historical heating load time series data samples and historical weather time series data samples into heating load supervised data and weather supervised data, respectively, comprises:
taking the former variables in the historical heating load time sequence data sample as input variables of the model, and taking the latter variables as the output of the model for polymerization to obtain heating load supervised data;
and taking the former variables in the historical meteorological time sequence data samples as input variables of the model, and taking the latter variables as the output of the model for aggregation to obtain meteorological supervised data.
4. The method of claim 1, wherein the pre-processing comprises maximum processing, minimum processing, data averaging, and normalization processing.
5. The method of claim 2, wherein the time series neural network comprises a weather timing representation component, a heating load timing representation component, and an autoregressive representation component;
the weather timing representing means includes: a first convolution representation layer and a first autoregressive representation layer;
the heating load time sequence expression means includes: a second convolution representation layer, a graph attention representation layer, a GRU neural network and a linear layer and a second autoregressive representation layer;
the step of inputting the heating load supervised data and the meteorological supervised data into the time series neural network for training, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition comprises the following steps:
and inputting the heating load supervised data into the heating load time sequence representation component, inputting the meteorological supervised data into the meteorological time sequence representation component at the same time, training, adjusting parameters in the time sequence neural network, and obtaining the regional heating load prediction model after reaching a preset training cut-off condition.
6. The method of claim 2, wherein the objective function is set to a mean square error loss function.
7. A district heating load prediction device, comprising:
the acquisition module is used for acquiring heating load time sequence data and meteorological time sequence data of a target area in a historical time period;
the prediction module is used for inputting a pre-trained regional heating load prediction model after preprocessing the heating load time sequence data and the meteorological time sequence data to obtain predicted heating load data of a target region in a future set time period;
the regional heating load prediction model is obtained by training a time series neural network through historical heating load time series data samples and historical meteorological time series data samples of a target region.
8. The apparatus of claim 7, wherein the pre-processing comprises maximum processing, minimum processing, data averaging, and normalization processing.
9. An electronic device, comprising: memory, processor and computer program stored on the memory and executable on the processor, characterized in that the processor, when executing the computer program, is adapted to carry out the method according to any of claims 1 to 6.
10. A computer-readable storage medium having computer-readable instructions stored thereon, the computer-readable instructions being executable by a processor to implement the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211636175.2A CN115796382A (en) | 2022-12-20 | 2022-12-20 | Regional heating load prediction method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211636175.2A CN115796382A (en) | 2022-12-20 | 2022-12-20 | Regional heating load prediction method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115796382A true CN115796382A (en) | 2023-03-14 |
Family
ID=85426161
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211636175.2A Pending CN115796382A (en) | 2022-12-20 | 2022-12-20 | Regional heating load prediction method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796382A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116293310A (en) * | 2023-04-07 | 2023-06-23 | 浙江锦华智能工程有限公司 | Security monitoring equipment |
-
2022
- 2022-12-20 CN CN202211636175.2A patent/CN115796382A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116293310A (en) * | 2023-04-07 | 2023-06-23 | 浙江锦华智能工程有限公司 | Security monitoring equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Short-term load forecasting for industrial customers based on TCN-LightGBM | |
| Imani | Electrical load-temperature CNN for residential load forecasting | |
| CN109802430B (en) | Wind power grid control method based on LSTM-Attention network | |
| Shen et al. | Wind speed prediction of unmanned sailboat based on CNN and LSTM hybrid neural network | |
| CN114626512B (en) | High-temperature disaster forecasting method based on directed graph neural network | |
| He et al. | A combined model for short-term wind power forecasting based on the analysis of numerical weather prediction data | |
| Chiang et al. | Hybrid time-series framework for daily-based PM 2.5 forecasting | |
| WO2019081782A1 (en) | Machine learning systems with memory based parameter adaptation for learning fast and slower | |
| CN115796382A (en) | Regional heating load prediction method, device, equipment and storage medium | |
| Lu et al. | A hybrid model based on convolutional neural network and long short-term memory for short-term load forecasting | |
| CN115271253B (en) | Method and device for constructing water-wind-solar power generation power prediction model and storage medium | |
| CN116451594B (en) | Training method and device of icing prediction model, prediction method and device and electronic equipment | |
| CN113469427A (en) | Convolution LSTM-based photovoltaic power station day-ahead power generation amount prediction method and system | |
| CN114065653A (en) | Construction method of power load prediction model and power load prediction method | |
| CN111863276B (en) | Hand-foot-and-mouth disease prediction method using fine-grained data, electronic device, and medium | |
| Vakitbilir et al. | Hybrid deep learning models for multivariate forecasting of global horizontal irradiation | |
| CN116722544B (en) | Distributed photovoltaic short-term prediction method and device, electronic equipment and storage medium | |
| CN117391466A (en) | Novel early warning method and system for contradictory dispute cases | |
| CN113222281A (en) | Power distribution network short-term load prediction method and device based on improved AlexNet-GRU model | |
| CN116576504A (en) | Interpretable region thermal load prediction method, interpretable region thermal load prediction device, interpretable region thermal load prediction equipment and storage medium | |
| Xu et al. | Residual autoencoder-LSTM for city region vehicle emission pollution prediction | |
| CN115481788B (en) | Phase change energy storage system load prediction method and system | |
| Du et al. | An innovative interpretable combined learning model for wind speed forecasting | |
| Senekane et al. | Weather Nowcasting Using Deep Learning Techniques | |
| CN115099497A (en) | CNN-LSTM-based real-time flood forecasting intelligent method |
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 |