CN117172138A - Urban traffic carbon emission prediction method and device based on deep learning - Google Patents
Urban traffic carbon emission prediction method and device based on deep learning Download PDFInfo
- Publication number
- CN117172138A CN117172138A CN202311449714.6A CN202311449714A CN117172138A CN 117172138 A CN117172138 A CN 117172138A CN 202311449714 A CN202311449714 A CN 202311449714A CN 117172138 A CN117172138 A CN 117172138A
- Authority
- CN
- China
- Prior art keywords
- urban traffic
- data
- carbon
- carbon emission
- traffic system
- 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
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 148
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 147
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000013135 deep learning Methods 0.000 title claims abstract description 36
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 230000004927 fusion Effects 0.000 claims abstract description 25
- 238000006073 displacement reaction Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000004088 simulation Methods 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 13
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims description 12
- 238000012544 monitoring process Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 9
- 238000001514 detection method Methods 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000000611 regression analysis Methods 0.000 claims description 4
- 238000013468 resource allocation Methods 0.000 claims description 4
- 230000003313 weakening effect Effects 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 9
- 230000006872 improvement Effects 0.000 abstract description 2
- 239000010410 layer Substances 0.000 description 29
- 238000010586 diagram Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000007405 data analysis Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000003915 air pollution Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000002441 reversible effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009466 transformation 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
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Traffic Control Systems (AREA)
Abstract
The application discloses a method and a device for predicting carbon emission of urban traffic based on deep learning, wherein the method comprises the following steps: acquiring basic data of an urban traffic system, and preprocessing the basic data; based on the preprocessed basic data, establishing a digital twin model of urban traffic corresponding to the urban traffic system; acquiring carbon displacement related data of the urban traffic system in real time, inputting the data into the digital twin model of the urban traffic system, and simulating the running state of the urban traffic system by using the digital twin model; and calculating the predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model. The method can fully utilize the multisource data of the urban traffic, improves the accuracy of the urban traffic carbon emission prediction through the pertinence improvement of the basic model, and can solve the problems of low data utilization rate and low prediction accuracy in the existing urban traffic carbon emission prediction technology.
Description
Technical Field
The application relates to the technical field of large data analysis in the transportation industry, in particular to a method and a device for predicting urban traffic carbon emission based on deep learning.
Background
In recent years, global climate change has attracted extensive attention from the world and human activities emit CO 2 Is one of the important causes of climate change. Cities serve as important areas of population aggregation, and due to the great demand for petrochemical energy, a large amount of carbon emissions are caused, while in the development process of modern metropolitan areas, carbon emissions from urban traffic occupy a significant part. Therefore, how to accurately evaluate and predict the total carbon emission in the future urban traffic field, and complete the balance development among economy, energy and environment on the basis of realizing green low-carbon urban traffic is a problem to be solved in the prior art.
Aiming at the problem of how to realize the prediction of the carbon emission of urban traffic, the prior art has the following defects: 1) Urban geographic information and urban vehicle information are disordered, the entity types are numerous, the characteristic difference is large, the result is not satisfactory when a single data model is adopted for mining, and the accuracy is low. 2) Although the research of the deep learning algorithm is very mature, the prediction model applied to the aspects of traffic flow prediction and air pollution control is still not perfect and too simplified, and the complexity and diversity of the urban traffic system cannot be truly reflected.
Therefore, it is necessary to provide a method and a device for predicting carbon emission of urban traffic based on deep learning, which can fully utilize multi-source data of urban traffic, and improve the accuracy of urban traffic carbon emission prediction through pertinence improvement of a basic model, so as to solve the problems of low data utilization rate and low prediction accuracy in the existing urban traffic carbon emission prediction technology.
Disclosure of Invention
The application provides a method and a device for predicting carbon emission of urban traffic based on deep learning, which are used for solving the technical problems of inaccurate carbon emission prediction results, insufficient data mining depth and low data utilization rate caused by single data and excessively simplified prediction model, which are not suitable for application scenes of carbon emission prediction of urban traffic in the prior art.
In order to solve the above problems, the present application provides a method for predicting carbon emission of urban traffic based on deep learning, comprising:
acquiring basic data of an urban traffic system, and preprocessing the basic data;
based on the preprocessed basic data, establishing a digital twin model of urban traffic corresponding to the urban traffic system;
acquiring carbon displacement related data of the urban traffic system in real time, inputting the data into the digital twin model of the urban traffic system, and simulating the running state of the urban traffic system by using the digital twin model;
and calculating the predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model.
Further, the urban traffic digital twin model comprises a data input module, a simulation module, a prediction optimization module and a detection feedback module;
the data input module is used for acquiring the carbon discharge related data of the urban traffic system in real time;
the simulation module is used for combining the scene modeling of the urban traffic and the carbon displacement related data to obtain a digital twin model which is used for simulating the running state and performance of the urban traffic system;
the prediction optimization module is used for optimizing the urban traffic system and predicting the efficiency and the safety of the urban traffic system;
the detection feedback module is used for monitoring and feeding back the urban traffic system in real time and responding and processing the emergency.
Further, the preset carbon fusion algorithm model comprises a Resnet network, a stacked LSTM unit, a full-connection layer and an output layer which are sequentially connected;
the Resnet network comprises a plurality of residual blocks for weakening the connection between adjacent network layers;
the stacked LSTM unit is used for extracting the carbon loss characteristic information of the carbon discharge related data, establishing a time sequence model and predicting the future carbon emission according to historical traffic data;
the full connection layer is used for completing regression analysis of data;
and the output layer is used for calculating and obtaining the predicted value of the carbon emission.
Further, each residual block of the Resnet network contains two 3×3 convolutional layers of the same number of output channels; each 3 x 3 convolution layer is followed by a batch normalization layer and a ReLu activation function; by skipping two convolution operations in the residual block across the layer data path, the input is directly preceded by the final ReLu activation function.
Further, the method further comprises:
and optimizing the resource allocation of the urban traffic by utilizing the urban traffic digital twin model according to the predicted value of the urban traffic carbon emission.
Further, preprocessing the basic data, including: data cleaning, data interpolation, abnormal data processing, data integration and data dimension reduction.
Further, the base data includes average temperature, population density, city infrastructure, traffic network, and air quality.
Further, the carbon displacement related data includes: the sensing data collected by various measuring devices from an extra-high voltage direct current transmission system of the urban traffic system.
The application also provides an urban traffic carbon emission prediction device based on deep learning, which comprises:
the data acquisition module is used for acquiring basic data of the urban traffic system and preprocessing the basic data;
the model building module is used for building a digital twin model of urban traffic corresponding to the urban traffic system based on the preprocessed basic data;
the simulation operation module is used for acquiring the carbon displacement related data of the urban traffic system in real time and inputting the data into the digital twin model of the urban traffic, and simulating the operation state of the urban traffic system by using the digital twin model;
and the carbon emission prediction module is used for calculating a predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model.
The application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for predicting the carbon emission of the urban traffic based on the deep learning according to any one of the technical schemes is realized.
Compared with the prior art, the application has the beneficial effects that: firstly, obtaining basic data of an urban traffic system, and preprocessing the basic data; secondly, establishing a corresponding urban traffic digital twin model based on urban traffic basic data, and determining a calculation mode of carbon emission loss of an urban traffic system; and finally, extracting the characteristics of the sample data by using a preset carbon fusion algorithm model, and calculating according to the actual monitoring value to obtain an accurate carbon emission predicted value of the urban traffic system. The method fully utilizes the urban related data collected from a plurality of sources, and utilizes the carbon fusion algorithm model constructed based on the LSTM and ResNet networks, thereby improving the accuracy of the carbon emission value of the urban traffic system and providing a theoretical tool for optimizing the carbon emission of the urban traffic. The method is suitable for early warning and daily evaluation of carbon emission in an urban traffic system, and has strong practicability.
Drawings
FIG. 1 is a schematic flow chart of an urban traffic carbon emission prediction method based on deep learning according to an embodiment of the application;
FIG. 2 is a schematic diagram of a digital twin model according to an embodiment of the present application;
FIG. 3 is a flowchart for constructing a map for urban traffic carbon emission prediction according to an embodiment of the present application;
FIG. 4 is a schematic diagram of a carbon fusion algorithm model based on Resnet and stacked LSTM according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of an LSTM network according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of a res net network residual block according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of an urban traffic carbon emission prediction device based on deep learning according to an embodiment of the present application.
Detailed Description
The following detailed description of preferred embodiments of the application is made in connection with the accompanying drawings, which form a part hereof, and together with the description of the embodiments of the application, are used to explain the principles of the application and are not intended to limit the scope of the application.
Before describing embodiments of the present application, related terms of the present application will be explained first.
Digital twinning: digital twinning is a technique that creates a digitized image of a physical system in the real world, and improves the performance and efficiency of the physical system by analyzing, simulating, and optimizing the digital image. In practice, digital twinning can be applied in a number of fields, such as industrial manufacturing, urban planning, healthcare, etc. In the field of urban planning, digital twinning can create a digital urban model including parameters in terms of urban infrastructure, traffic network, air quality, etc., thereby simulating the urban development process and optimizing the urban planning scheme. The digital twinning can also help city managers to predict traffic flow and control air pollution, and improve city operation efficiency and environment quality.
Deep learning: deep learning is a machine learning technology based on an artificial neural network, and can extract high-level abstract features from a large amount of data through multi-layer nonlinear transformation, and perform tasks such as classification, regression, clustering and the like. Since deep learning is widely used in the field of image recognition, it has been widely used in various fields.
ResNet (residual neural network): resNet is a deep convolutional neural network, and the introduction of ResNet can effectively solve the problems of gradient disappearance and precision reduction of the deep convolutional neural network. In urban traffic carbon emission prediction, resNet can be used for extracting characteristics in traffic data, and the expression capability of a model on complex traffic data is enhanced.
Stacked LSTM (long short term memory network): stacked LSTM is a deep learning technique that can be used for sequence prediction, typically by stacking multiple LSTM layers together to better capture long-term dependencies in a sequence. In urban traffic carbon emission prediction, stacked LSTM may be used to build a time series model to predict future carbon emissions based on historical traffic data.
The embodiment of the application provides a method for predicting carbon emission of urban traffic based on deep learning, and fig. 1 is a schematic flow chart of the method for predicting carbon emission of urban traffic based on deep learning, as shown in fig. 1, wherein the method comprises the following steps:
step S101: acquiring basic data of an urban traffic system, and preprocessing the basic data;
step S102: based on the preprocessed basic data, establishing a digital twin model of urban traffic corresponding to the urban traffic system;
step S103: acquiring carbon displacement related data of the urban traffic system in real time, inputting the data into the digital twin model of the urban traffic system, and simulating the running state of the urban traffic system by using the digital twin model;
step S104: and calculating the predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model.
According to the urban traffic carbon emission prediction method based on deep learning, firstly, basic data of an urban traffic system are obtained, and the basic data are preprocessed; secondly, establishing a digital twin model of urban traffic corresponding to the urban traffic system according to the basic data, and determining a calculation mode of carbon emission loss of the urban traffic system; and finally, extracting the characteristics of the sample data by using a preset carbon fusion algorithm model, and calculating to obtain an accurate carbon emission value of the urban traffic system. The method of the embodiment fully utilizes the urban related data collected from a plurality of sources, and utilizes the carbon fusion algorithm model constructed based on the LSTM and ResNet networks, thereby improving the accuracy of the carbon emission value of the urban traffic system and providing a theoretical tool for optimizing the carbon emission of the urban traffic. The method of the embodiment is suitable for early warning and daily evaluation of carbon emission in an urban traffic system, and has strong practicability.
In some embodiments, the base data includes average temperature, population density, city infrastructure, traffic network, and air quality. The corresponding digital twin model is established through the basic data of urban traffic, so that the urban development process is simulated, the urban planning scheme is optimized, traffic flow prediction and air pollution control are facilitated for urban managers, and urban operation efficiency and environment quality are improved.
In some embodiments, preprocessing the base data includes: data cleaning, data interpolation, abnormal data processing, data integration and data dimension reduction.
Specifically, the data preprocessing process comprises the following steps: urban related data is collected from a plurality of sources, and the collected data is subjected to cleaning operations including deduplication, filling in missing values, and processing outliers.
It should be noted that the cleansing operation performs data cleansing for different data features, including ensuring the uniqueness of the data by deleting duplicate items in the data set based on one or more features in the data record.
The specific method for filling the missing value is as follows: the interpolation algorithm is used to infer missing values present at certain locations in the dataset from surrounding data.
In order to exclude operations performed by the influence of abnormal values on the data analysis and modeling results, the abnormal values may be generated for various reasons, such as data input errors, equipment malfunction, and the like. If the occurrence of outliers is not due to the real data itself, we can reject the outliers or replace them with appropriate values using some statistical method to ensure the accuracy and reliability of the data.
After data deduplication, missing value interpolation and outlier elimination, integrating data from different sources, establishing a unified data storage structure and standard format, and converting the original data into a data format which can be used for analysis after storage, such as numerical data, typing data or time series data.
From the stored mass features, meaningful features are extracted from the data, such as average temperature, population density, traffic conditions, etc., according to the need.
Finally, carrying out normalization operation on all the characteristic data so that the data among different characteristics have comparability in a numerical range;
specifically, normalization operation is performed on all the characteristic data, and a calculation formula is as follows:
in the method, in the process of the application,for normalized value, ++>Original data->And->The maximum and minimum values of the measured data, respectively.
The high-dimensional feature is dimensionality reduced to facilitate better visualization and analysis.
Through the data preprocessing step, the data quality can be obviously improved, and the data reliability is ensured. And storing the preprocessed data into a database or a file system, so that subsequent data analysis and application are facilitated.
In some embodiments, the carbon displacement related data comprises: the sensing data collected by various measuring devices from an extra-high voltage direct current transmission system of the urban traffic system.
Specifically, the extra-high voltage direct current transmission system of the urban traffic system comprises: distributed power sources, power distribution equipment, alternating current-direct current transformers, converters, direct current transmission lines and the like.
Based on Digital Twin (DT) technology, the real-time updating and historical operation data of a physical mechanism model and a sensor are fully utilized, a simulation process of multiple physical quantities, multiple time scales and multiple probabilities (randomness) is integrated, and mapping is completed in a Digital Twin body, so that the full life cycle process of a corresponding physical entity and the complex operation state of an actual system are reflected.
In some embodiments, the urban traffic digital twin model comprises a data input module, a simulation module, a prediction optimization module and a detection feedback module;
the data input module is used for acquiring the carbon discharge related data of the urban traffic system in real time;
the simulation module is used for combining the scene modeling of the urban traffic and the carbon displacement related data to obtain a digital twin model which is used for simulating the running state and performance of the urban traffic system;
the prediction optimization module is used for optimizing the urban traffic system and predicting the efficiency and the safety of the urban traffic system;
the detection feedback module is used for monitoring and feeding back the urban traffic system in real time and responding and processing the emergency.
Specifically, the operation steps of the urban traffic digital twin model include:
various data in the physical space of the urban traffic system, including traffic flow, speed, density and the like, are collected in real time through equipment such as sensors and the like, and then preprocessing operations such as de-duplication, filling missing values, processing abnormal values and the like are carried out on the data.
The urban traffic scene modeling and the real-time data acquisition and processing are combined to establish a digital twin model for simulating the running state and performance of the urban traffic system.
The urban traffic system is simulated and optimized through the digital twin model, for example, the traffic flow, road conditions and the like are predicted, so that the efficiency and the safety of the urban traffic system are improved.
And carrying out real-time monitoring and feedback on the urban traffic system through real-time data acquisition and analysis results of the digital twin model, wherein the real-time monitoring and feedback comprises traffic information release, driving advice, road repair and the like, and response and processing of emergency.
As shown in fig. 2, fig. 2 is a schematic structural diagram of a digital twin model. The digital twin model of urban traffic is a digital twin technology-based urban traffic simulation model, which models an urban traffic system as a digital image, and predicts and optimizes traffic flow, road conditions and the like through real-time monitoring and data analysis. The precise mapping and real-time feedback mechanism between the physical world and the digital world are established through the digital twin technology, so that the interconnection, intercommunication and interoperation of the physical world and the digital world are realized, a new system for describing, diagnosing, predicting and deciding the physical world by using the virtual world is constructed, and finally the resource allocation efficiency of the physical world is optimized.
Compared with the traditional urban traffic simulation method, the digital twin model has higher precision and reliability, and can better reflect the complexity and time variability of an urban traffic system. Is an emerging product under the combined action of multiple backgrounds such as the increasing complexity of urban traffic systems, the increasing trend of data presentation blowout, the perfect development of digital twin technology and the like. Compared with an information physical system focusing on a real-time control entity and simulation software driven by a classical model, the information physical system focusing on data-driven real-time situation awareness and super-real-time virtual deduction are more focused, and the information physical system aims at providing references for marketing, regulation and control decisions.
As shown in fig. 3, fig. 3 is a map construction flow chart of urban traffic carbon emission prediction, after various measuring devices collect data from the physical space of an urban traffic system, the data are uploaded to a carbon metering data processing platform, and then feature extraction, fusion and application are performed by utilizing artificial intelligence and big data analysis technology through a preset carbon fusion algorithm model (Res-skLSTM model in fig. 3), so that virtual control, real reverse action and lean operation and maintenance are finally realized.
In some embodiments, the preset carbon fusion algorithm model includes a Resnet network, a stacked LSTM unit, a full connection layer and an output layer connected in sequence;
the Resnet network comprises a plurality of residual blocks for weakening the connection between adjacent network layers;
the stacked LSTM unit is used for extracting the carbon loss characteristic information of the carbon discharge related data, establishing a time sequence model and predicting the future carbon emission according to historical traffic data;
the full connection layer is used for completing regression analysis of data;
and the output layer is used for calculating and obtaining the predicted value of the carbon emission.
And excavating an internal rule and a coupling relation between the high-dimensional data by using a deep learning algorithm through a preset carbon fusion algorithm model. As shown in FIG. 4, FIG. 4 is a schematic diagram of a carbon fusion algorithm model based on Resnet and stacked LSTM.
It should be noted that, in the deep learning algorithm, the LSTM network adopts a gate structure to update the states of neurons in the network, and stores and controls information in a storage structure, so that the LSTM network has stronger modeling and learning capabilities for time series data, is more suitable for the scene where the present application is located, and the structure is shown in fig. 5. Because LSTM has strong feature extraction capability, the feature extraction of sample data can be accomplished by the LSTM layer when constructing the carbon metrology model.
But as the number of network layers increases, gradient extinction and explosion problems occur. To solve the above problem, some network layers may skip the connection of the next layer of neurons, i.e. the strong connection between adjacent network layers is weakened by interlayer connection. The neural network of this structure is called a residual network. Therefore, the model algorithm is improved and fused, and a Resnet network is added into the carbon fusion algorithm model.
In some embodiments, each residual block of the Resnet network contains two 3 x 3 convolutional layers of the same number of output channels; each 3 x 3 convolution layer is followed by a batch normalization layer and a ReLu activation function; by skipping two convolution operations in the residual block across the layer data path, the input is directly preceded by the final ReLu activation function. As shown in fig. 6, fig. 6 is a schematic diagram of the structure of a residual block of the res net network.
The ResNet network based on the residual structure can better solve the gradient disappearance problem caused by the continuous deepening of the network layer number.
In addition, the accuracy and training efficiency of LSTM neural network models are affected by the number of hidden layers and the number of neurons. In order to shorten training time and ensure accuracy of results as much as possible while improving feature extraction capability of the model on carbon loss feature information, stacked LSTM units are adopted, the optimal number of the LSTM units is determined through experiments, and finally, a final regression part is completed by using a full connection layer.
The depth of the LSTM network is increased by adding the number of hidden layers, and meanwhile, the model training efficiency and the accuracy of the calculation result are improved.
Through the transformation of the basic deep learning network to the urban traffic system, the constructed carbon metering fusion algorithm model Res-skLSTM has good feature extraction capability and feature fusion capability, can be used for processing high-dimensional data in the urban traffic system, improves the accuracy of carbon metering, and has practicability and reliability more than a direct calculation method adopted in the prior art.
In order to quantitatively evaluate the accuracy of the carbon metering digital twin model and the superiority of the model relative to other deep learning models, the carbon metering model Res-skLSTM constructed in the prior art is calculated to obtain an accurate system carbon emission value, and the accurate system carbon emission value is respectively compared with the calculation results of Resnet and LSTM, so that the effectiveness and superiority of the improved model are verified. And measuring the deviation between the actual value and the calculated value of the system carbon loss measurement by using indexes such as mean square error, root mean square error, average absolute error and the like, wherein the definition is as follows:
1) Mean square error (Mean Squared Error, MSE)
2) Root mean square error (Root Mean Squared Error, RMSE)
3) Mean absolute error (Mean Absolute Error, MAE)
4) Average absolute percent error (Mean Absolute Percentage Error, MAPE)
5) R2 fraction (R2)
Wherein,and->Representing the carbon of the extra-high voltage direct current transmission system on the ith sample respectivelyThe true and calculated values of the loss-metering error, < >>Representing the mean of the true values of the samples, n being the test sample volume.
From the above accuracy index formula, other indexes are inversely proportional to the accuracy of the digital twin model except the R2 fraction, namely, the smaller the index is, the higher the model accuracy is, the closer R2 is to 1, and the calculated value is close to the distribution of the original data.
The final test results are shown in Table 1.
Table 1 results of comparison of carbon measurement test indicators for different models
As can be seen from Table 1, the Res-skLSTM model has lower error bias than the stacked LSTM model and Resnet model alone, while also being closer to the distribution of raw data on the R2 index.
Taking MSE as an example, the Res-skLSTM model has 30.51% lower error than the stacked LSTM model and 39.71% lower error than the Resnet model. From this, it can be derived that the Res-skLSTM model has the highest accuracy in system carbon metering. The validity of the Res-skLSTM model was verified.
In some embodiments, the method further comprises:
and optimizing the resource allocation of the urban traffic by utilizing the urban traffic digital twin model according to the predicted value of the urban traffic carbon emission.
The method comprises the steps of utilizing a deep learning algorithm to mine an internal rule and a coupling relation between high-dimensional data, utilizing a sensing measurement acquisition device to acquire carbon emission related data, utilizing an artificial intelligence and big data analysis technology to extract, fuse and apply characteristics, and utilizing a digital twin model to simulate and predict an optimization result so as to finally realize virtual control, real reverse action and lean operation and maintenance.
The embodiment of the application also provides an urban traffic carbon emission prediction device 700 based on deep learning, as shown in fig. 7, comprising:
the data acquisition module 701 is configured to acquire basic data of an urban traffic system, and perform preprocessing on the basic data;
the model building module 702 is configured to build a digital twin model of urban traffic corresponding to the urban traffic system based on the preprocessed base data;
the simulation running module 703 is configured to acquire data related to carbon displacement of the urban traffic system in real time, input the data into the digital twin model of urban traffic, and simulate the running state of the urban traffic system by using the digital twin model;
and the carbon emission prediction module 704 is configured to calculate a predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model.
The embodiment also provides a computer readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method for predicting urban traffic carbon emission based on deep learning according to any one of the above technical schemes is implemented.
According to the computer readable storage medium and the computing device provided in the above embodiments of the present application, the implementation of the specific description of the method for predicting carbon emission of urban traffic based on deep learning according to the present application may be referred to, and have similar advantages as the method for predicting carbon emission of urban traffic based on deep learning according to the present application, and will not be described herein.
The application discloses a method and a device for predicting carbon emission of urban traffic based on deep learning, which are characterized in that firstly, basic data of an urban traffic system are obtained, and the basic data are preprocessed; secondly, establishing a corresponding urban traffic digital twin model based on urban traffic basic data, and determining a calculation mode of carbon emission loss of an urban traffic system; and finally, extracting the characteristics of the sample data by using a preset carbon fusion algorithm model, and calculating according to the actual monitoring value to obtain an accurate carbon emission predicted value of the urban traffic system. The method fully utilizes the urban related data collected from a plurality of sources, and utilizes the carbon fusion algorithm model constructed based on the LSTM and ResNet networks, thereby improving the accuracy of the carbon emission value of the urban traffic system and providing a theoretical tool for optimizing the carbon emission of the urban traffic. The method and the device are suitable for early warning and daily evaluation of carbon emission in an urban traffic system, and have strong practicability.
The present application is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present application are intended to be included in the scope of the present application.
Claims (8)
1. The urban traffic carbon emission prediction method based on deep learning is characterized by comprising the following steps of:
acquiring basic data of an urban traffic system, and preprocessing the basic data;
based on the preprocessed basic data, establishing a digital twin model of urban traffic corresponding to the urban traffic system; the urban traffic digital twin model comprises a data input module, a simulation module, a prediction optimization module and a detection feedback module; the data input module is used for acquiring the carbon discharge related data of the urban traffic system in real time; the simulation module is used for combining the scene modeling of the urban traffic and the carbon displacement related data to obtain a digital twin model which is used for simulating the running state and performance of the urban traffic system; the prediction optimization module is used for optimizing the urban traffic system and predicting the efficiency and the safety of the urban traffic system; the detection feedback module is used for monitoring and feeding back the urban traffic system in real time and responding and processing the emergency;
acquiring carbon displacement related data of the urban traffic system in real time, inputting the data into the digital twin model of the urban traffic system, and simulating the running state of the urban traffic system by using the digital twin model;
calculating a predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model; the preset carbon fusion algorithm model comprises a Resnet network, a stacked LSTM unit, a full-connection layer and an output layer which are sequentially connected; the Resnet network comprises a plurality of residual blocks for weakening the connection between adjacent network layers; the stacked LSTM unit is used for extracting the carbon loss characteristic information of the carbon discharge related data, establishing a time sequence model and predicting the future carbon emission according to historical traffic data; the full connection layer is used for completing regression analysis of data; and the output layer is used for calculating and obtaining the predicted value of the carbon emission.
2. The deep learning-based urban traffic carbon emission prediction method according to claim 1, wherein each residual block of the Resnet network comprises two 3 x 3 convolutional layers of the same number of output channels; each 3 x 3 convolution layer is followed by a batch normalization layer and a ReLu activation function; by skipping two convolution operations in the residual block across the layer data path, the input is directly preceded by the final ReLu activation function.
3. The deep learning-based urban traffic carbon emission prediction method according to claim 1, characterized in that it further comprises:
and optimizing the resource allocation of the urban traffic by utilizing the urban traffic digital twin model according to the predicted value of the urban traffic carbon emission.
4. The deep learning-based urban traffic carbon emission prediction method according to claim 1, wherein preprocessing the basic data comprises: data cleaning, data interpolation, abnormal data processing, data integration and data dimension reduction.
5. The deep learning based urban traffic carbon emission prediction method according to claim 1, wherein the base data comprises average temperature, population density, urban infrastructure, traffic network and air quality.
6. The deep learning-based urban traffic carbon emission prediction method according to claim 1, wherein the carbon displacement-related data comprises: the sensing data collected by various measuring devices from an extra-high voltage direct current transmission system of the urban traffic system.
7. Urban traffic carbon emission prediction device based on deep learning, characterized by comprising:
the data acquisition module is used for acquiring basic data of the urban traffic system and preprocessing the basic data;
the model building module is used for building a digital twin model of urban traffic corresponding to the urban traffic system based on the preprocessed basic data; the urban traffic digital twin model comprises a data input module, a simulation module, a prediction optimization module and a detection feedback module; the data input module is used for acquiring the carbon discharge related data of the urban traffic system in real time; the simulation module is used for combining the scene modeling of the urban traffic and the carbon displacement related data to obtain a digital twin model which is used for simulating the running state and performance of the urban traffic system; the prediction optimization module is used for optimizing the urban traffic system and predicting the efficiency and the safety of the urban traffic system; the detection feedback module is used for monitoring and feeding back the urban traffic system in real time and responding and processing the emergency;
the simulation operation module is used for acquiring the carbon displacement related data of the urban traffic system in real time and inputting the data into the digital twin model of the urban traffic, and simulating the operation state of the urban traffic system by using the digital twin model; the preset carbon fusion algorithm model comprises a Resnet network, a stacked LSTM unit, a full-connection layer and an output layer which are sequentially connected; the Resnet network comprises a plurality of residual blocks for weakening the connection between adjacent network layers; the stacked LSTM unit is used for extracting the carbon loss characteristic information of the carbon discharge related data, establishing a time sequence model and predicting the future carbon emission according to historical traffic data; the full connection layer is used for completing regression analysis of data; the output layer is used for calculating and obtaining the predicted value of the carbon emission
And the carbon emission prediction module is used for calculating a predicted value of the urban traffic carbon emission according to the running state and the carbon emission related data by using a preset carbon fusion algorithm model.
8. A computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and when the computer program is executed by a processor, the computer program realizes the urban traffic carbon emission prediction method based on deep learning according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311449714.6A CN117172138B (en) | 2023-11-02 | 2023-11-02 | Urban traffic carbon emission prediction method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311449714.6A CN117172138B (en) | 2023-11-02 | 2023-11-02 | Urban traffic carbon emission prediction method and device based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117172138A true CN117172138A (en) | 2023-12-05 |
| CN117172138B CN117172138B (en) | 2024-02-02 |
Family
ID=88937964
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311449714.6A Active CN117172138B (en) | 2023-11-02 | 2023-11-02 | Urban traffic carbon emission prediction method and device based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117172138B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20140102095A (en) * | 2013-02-13 | 2014-08-21 | 청주대학교 산학협력단 | Land use/transportation/environment model simulation system of low carbon emission style in urban development zone |
| CN115545450A (en) * | 2022-09-27 | 2022-12-30 | 广东师大维智信息科技有限公司 | Carbon emission collaborative prediction method based on digital twinning |
| CN116798218A (en) * | 2022-10-31 | 2023-09-22 | 广东师大维智信息科技有限公司 | Urban low-carbon traffic big data detection method based on digital twinning |
| CN116843083A (en) * | 2023-07-21 | 2023-10-03 | 国网福建省电力有限公司营销服务中心 | Carbon emission prediction system and method based on hybrid neural network model |
-
2023
- 2023-11-02 CN CN202311449714.6A patent/CN117172138B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20140102095A (en) * | 2013-02-13 | 2014-08-21 | 청주대학교 산학협력단 | Land use/transportation/environment model simulation system of low carbon emission style in urban development zone |
| CN115545450A (en) * | 2022-09-27 | 2022-12-30 | 广东师大维智信息科技有限公司 | Carbon emission collaborative prediction method based on digital twinning |
| CN116798218A (en) * | 2022-10-31 | 2023-09-22 | 广东师大维智信息科技有限公司 | Urban low-carbon traffic big data detection method based on digital twinning |
| CN116843083A (en) * | 2023-07-21 | 2023-10-03 | 国网福建省电力有限公司营销服务中心 | Carbon emission prediction system and method based on hybrid neural network model |
Non-Patent Citations (7)
| Title |
|---|
| CHENG, X., ZHANG, W., WENZEL, A. ET AL.: "Stacked ResNet-LSTM and CORAL model for multi-site air quality prediction", 《NEURAL COMPUT & APPLIC》, vol. 34, pages 13849, XP037910387, DOI: 10.1007/s00521-022-07175-8 * |
| H. CHOI, S. RYU AND H. KIM: "Short-Term Load Forecasting based on ResNet and LSTM", 《2018 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS, CONTROL, AND COMPUTING TECHNOLOGIES FOR SMART GRIDS (SMARTGRIDCOMM), AALBORG, DENMARK》, pages 1 - 6 * |
| PARK J, YANG B: "GIS-Enabled Digital Twin System for Sustainable Evaluation of Carbon Emissions: A Case Study of Jeonju City, South Korea", 《SUSTAINABILITY》, vol. 12, no. 21, pages 9186 * |
| S. SONG, J. C. K. LAM, Y. HAN AND V. O. K. LI: "ResNet-LSTM for Real-Time PM2.5 and PM₁₀ Estimation Using Sequential Smartphone Images", 《IEEE》, vol. 8, pages 220069 - 220082, XP011825708, DOI: 10.1109/ACCESS.2020.3042278 * |
| X. CHEN, W. CHEN, V. DINAVAHI, Y. LIU AND J. FENG: "Short-Term Load Forecasting and Associated Weather Variables Prediction Using ResNet-LSTM Based Deep Learning", 《IEEE》, vol. 11, pages 5393 - 5405 * |
| XU ZHENG, TANGHAN JIANG, AND NAN ZHENG: "Developing and analyzing eco-driving strategies for on-road emission reduction in urban transport systems-A VR-enabled digital-twin approach", 《CHEMOSPHERE》, vol. 305, pages 203 - 15 * |
| 王琛等: "基于ResNet-LSTM 网络和注意力机制的综合能源系统多元负荷预测", 《电工技术学报》, vol. 37, no. 7, pages 1789 - 1799 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117172138B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220137612A1 (en) | Transformer fault diagnosis and positioning system based on digital twin | |
| CN115578015B (en) | Sewage treatment whole process supervision method, system and storage medium based on Internet of things | |
| Yu et al. | Prediction of highway tunnel pavement performance based on digital twin and multiple time series stacking | |
| CN112085261B (en) | Enterprise production status diagnosis method based on cloud fusion and digital twin technology | |
| Li et al. | Data-driven battery state of health estimation based on interval capacity for real-world electric vehicles | |
| CN110516820B (en) | BIM-based steel structure bridge informatization operation and maintenance system and processing method | |
| CN105574593B (en) | Track state static detection and control system and method based on cloud computing and big data | |
| CN106934720A (en) | Equipment insurance intelligent pricing method and system based on Internet of Things | |
| CN108304661A (en) | Diagnosis prediction method based on TDP models | |
| CN116681187B (en) | Enterprise carbon quota prediction method based on enterprise operation data | |
| CN104778622A (en) | Method and system for predicting TPS transaction event threshold value | |
| CN112688431A (en) | Power distribution network load overload visualization method and system based on big data | |
| CN111178585A (en) | Fault reporting amount prediction method based on multi-algorithm model fusion | |
| CN109299552A (en) | A kind of appraisal procedure and its assessment system of battery power status | |
| CN110716520B (en) | Flight control servo actuator reliability evaluation modeling method based on multi-source information fusion | |
| CN109902344B (en) | Device and system for predicting structural performance of middle-small span bridge group | |
| CN117435997B (en) | Elevator early warning and on-demand maintenance method based on transducer and time sequence compression mechanism | |
| CN113837473A (en) | Charging equipment fault rate analysis system and method based on BP neural network | |
| CN117391674A (en) | Reliability-based preventive maintenance optimization method and device for electrical equipment | |
| CN112418662A (en) | Power distribution network operation reliability analysis method using artificial neural network | |
| CN117172138B (en) | Urban traffic carbon emission prediction method and device based on deep learning | |
| CN117493798A (en) | Meteorological environment data analysis method and system | |
| Kovalyov | Design and development of a power system digital twin: A model-based approach | |
| CN114897262A (en) | Rail transit equipment fault prediction method based on deep learning | |
| Nowaczyk et al. | Towards autonomous knowledge creation from big data in smart cities |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |