Expressway service area carbon emission health state evaluation system and method thereof

Abstract

The invention relates to the technical field of carbon emission, in particular to an expressway service area carbon emission health status assessment system and method, an acquisition module acquires expressway service area carbon emission data in real time, and the data is stored monthly and annual; the data center module receives and preprocesses the data; the calculation module extracts spatial features by using a convolutional neural network, and performs classification and regression prediction through a long-short-term memory network; the state evaluation module performs classification evaluation on the carbon emission state based on a health index system; and the visualization module converts the evaluation result into a chart to realize data visualization. According to the system, real-time acquisition, processing, prediction, evaluation and display of carbon emission data are realized, a scientific basis is provided for carbon emission management, optimization of emission control strategies and promotion of development of the environmental protection cause are facilitated, and due to innovative application of the deep learning model, the system can adapt to characteristics and environmental changes of different service areas and has strong generalization ability.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission, in particular to a system and method for evaluating the health status of carbon emission in a highway service area. Background Art

[0002] With the rapid development of my country's economy and the continuous increase in the number of cars, the highway network is expanding. As an important part of highways, the carbon emission problem of service areas is becoming increasingly prominent.

[0003] Traditional highway service area carbon emission monitoring methods mainly rely on fixed sensors and manual inspections, which have many limitations. First, a single fixed sensor is difficult to fully capture the carbon emission characteristics in the complex environment of the service area, resulting in incomplete and inaccurate monitoring data. Second, the manual inspection method is inefficient and difficult to achieve real-time monitoring and rapid response. In addition, existing carbon emission assessment models are often too simple to effectively process multi-source heterogeneous data, and are difficult to adapt to the characteristics of different service areas and dynamically changing environments.

[0004] In recent years, with the development of the Internet of Things and artificial intelligence technologies, some researchers have tried to apply these new technologies to the field of carbon emission monitoring. However, these methods still have some problems. For example, although some systems have achieved multi-source data collection, they are not efficient in data transmission and processing, and it is difficult to meet the needs of real-time monitoring. Although other systems use machine learning algorithms for carbon emission prediction, the generalization ability of the model is insufficient, and it is difficult to adapt to the characteristics and seasonal changes of different service areas. In addition, the existing systems generally lack intelligent decision support functions, making it difficult to provide targeted optimization suggestions for service area managers.

[0005] In view of the shortcomings of existing technologies, it is urgent to develop an efficient, accurate and intelligent highway service area carbon emission health status assessment system. The system should be able to achieve real-time collection and fusion of multi-source data, use advanced artificial intelligence algorithms for accurate prediction and evaluation, and have adaptive optimization and intelligent decision support functions, so as to comprehensively improve the level of highway service area carbon emission management. Summary of the invention

[0006] The highway service area carbon emission health status assessment system and method provided by the present invention effectively solves the above technical problems by innovatively integrating multiple advanced technologies. The system adopts a modular design, including a carbon emission data acquisition module, a data center module, a data calculation module, a status assessment module and a visualization module. The modules work closely together to form a complete carbon emission monitoring, analysis and assessment closed loop.

[0007] The present invention proposes a highway service area carbon emission health status assessment system, comprising:

[0008] Carbon emission data collection module, used for:

[0009] Real-time collection of carbon emission data from highway service areas;

[0010] Save historical data at monthly and yearly granularity;

[0011] The carbon emission data center module is connected to the carbon emission data collection module for:

[0012] Receiving real-time data and historical data sent by the carbon emission data collection module;

[0013] Preprocessing the received data;

[0014] The carbon emission data calculation module is connected to the carbon emission data center module for:

[0015] Receiving preprocessed data sent by the carbon emission data center module;

[0016] Extracting spatial features of the preprocessed data using a convolutional neural network;

[0017] The extracted spatial features are input into the long short-term memory network for classification prediction and regression prediction;

[0018] The carbon emission data status assessment module is in communication with the carbon emission data calculation module and is used to:

[0019] Receiving the prediction result of the carbon emission data calculation module;

[0020] Based on the predetermined health status assessment indicator system, classify and assess the carbon emission status;

[0021] The carbon emission data visualization module is in communication with the carbon emission data status assessment module and is used to:

[0022] Receiving the evaluation result of the carbon emission data status evaluation module;

[0023] Generate corresponding visualization charts based on the evaluation results to provide data visualization display.

[0024] As a preference, it also includes:

[0025] The edge computing service layer module is connected to the carbon emission data collection module for:

[0026] Receiving real-time data collected by the carbon emission data collection module;

[0027] Preprocess, extract features, classify data and perform regression prediction on real-time data;

[0028] The calculation results are sent to the carbon emission data center module.

[0029] Preferably, the edge computing service layer module includes:

[0030] GPU computing unit, used for:

[0031] Distributed computing based on GPU;

[0032] Dynamically adjust computing resource allocation;

[0033] A hardware adapter unit is connected to the GPU computing unit for:

[0034] Automatic switching between GPU and CPU in extreme temperature environments;

[0035] Control the fan to cool down in real time.

[0036] Preferably, the carbon emission data collection module comprises:

[0037] Multi-source data acquisition unit for:

[0038] Collect traffic data, weather data and carbon dioxide concentration data;

[0039] The Internet of Things transmission unit is communicatively connected with the multi-source data acquisition unit and is used for:

[0040] The collected multi-source heterogeneous data are transmitted in real time to the carbon emission data center module through the Internet of Things technology.

[0041] Preferably, the carbon emission data calculation module further includes:

[0042] Multi-model fusion prediction unit, used for:

[0043] Combine time series prediction model, classification prediction model and regression prediction model to predict carbon emissions from multiple angles;

[0044] The error prediction unit is communicatively connected with the multi-model fusion prediction unit and is used for:

[0045] Constructing a set of prediction errors;

[0046] Long short-term memory networks are used to predict errors and achieve model self-correction.

[0047] As a preference, it also includes:

[0048] The adaptive multi-objective parameter optimization module is in communication with the carbon emission data calculation module and is used to:

[0049] Dynamically adjust model parameters for different highway service areas;

[0050] Balance the three optimization goals of prediction accuracy, precision and efficiency.

[0051] Preferably, the carbon emission data center module comprises:

[0052] Data pre-processing unit, used for:

[0053] De-noising the received data;

[0054] Perform feature extraction on the processed data;

[0055] The extracted feature data are smoothed and dimensionally reduced.

[0056] Preferably, the carbon emission data status assessment module comprises:

[0057] Multi-level evaluation unit for:

[0058] Rating the health status of carbon emissions;

[0059] Assign weight coefficients to different levels;

[0060] A SQLite data storage unit is communicatively connected to the multi-level evaluation unit, and is used for:

[0061] Use SQLite database to store evaluation results;

[0062] Provides efficient data storage and access mechanism.

[0063] As a preference, it also includes:

[0064] The intelligent vital sign monitoring strategy optimization module is connected to the carbon emission data status assessment module for:

[0065] Based on deep learning algorithms, real-time assessment of carbon emission status;

[0066] Based on the evaluation results, the monitoring strategy is adaptively adjusted.

[0067] The method for evaluating the carbon emission health status using the highway service area carbon emission health status evaluation system comprises the following steps:

[0068] S1: Collect real-time carbon emission data and historical data of highway service areas through the carbon emission data collection module;

[0069] S2: Transmit the collected data to the carbon emission data center module for preprocessing;

[0070] S3: Use the convolutional neural network in the carbon emission data calculation module to extract the spatial features of the preprocessed data, and input the extracted features into the long short-term memory network for classification prediction and regression prediction;

[0071] S4: The carbon emission data status assessment module classifies and assesses the prediction results based on a predetermined health status assessment indicator system;

[0072] S5: The carbon emission data visualization module generates corresponding visualization charts according to the evaluation results;

[0073] S6: The edge computing service layer module preprocesses, extracts features, classifies data, and performs regression prediction on the collected real-time data, and sends the results to the carbon emission data center module;

[0074] S7: Adaptive multi-objective parameter optimization module dynamically adjusts model parameters according to the characteristics of different highway service areas;

[0075] S8: The intelligent vital sign monitoring strategy optimization module evaluates the carbon emission status in real time based on the deep learning algorithm and makes adaptive adjustments to the monitoring strategy.

[0076] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0077] The system of the present invention has achieved technological breakthroughs and innovations in many aspects. First, through the fusion of multi-source heterogeneous data and the application of edge computing technology, the system has greatly improved the comprehensiveness and real-time performance of data collection. The carbon emission data collection module can not only collect real-time data, but also save historical data at different time granularities, providing rich data support for subsequent analysis. Secondly, the system cleverly combines convolutional neural networks and long short-term memory networks to effectively capture the spatiotemporal characteristics of carbon emission data and significantly improve the accuracy of predictions. The innovative application of this deep learning model enables the system to adapt to the characteristics and environmental changes of different service areas, and has strong generalization capabilities.

[0078] In addition, the system of the present invention also introduces an adaptive multi-objective parameter optimization module, which can dynamically adjust the model parameters according to the characteristics of different service areas to achieve the best balance between prediction accuracy, precision and efficiency. This adaptive optimization mechanism greatly enhances the flexibility and adaptability of the system, enabling it to maintain efficient operation in service areas of different sizes and types.

[0079] In practical applications, the system of the present invention has shown significant advantages. Through real-time and accurate carbon emission status assessment, the system can detect abnormal situations in a timely manner and provide managers with opportunities for rapid response. At the same time, the prediction model based on deep learning can warn of possible carbon emission peaks in advance and help service areas make plans. More importantly, the system's intelligent decision support function can generate targeted optimization suggestions based on the evaluation results, such as adjusting the energy use structure, optimizing equipment operating parameters, etc., so as to achieve a continuous reduction in carbon emissions.

[0080] In general, the highway service area carbon emission health status assessment system of the present invention realizes intelligent management of the entire process from data collection, processing, analysis to decision support through the organic combination of multiple technological innovations. This not only greatly improves the efficiency and accuracy of carbon emission monitoring, but also provides strong technical support for the low-carbon operation of service areas. With the widespread application of the system, it is expected to significantly promote the green development of highway service areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 It is a logic block diagram of the whole system of the present invention.

[0082] Figure 2 This is a logic block diagram of the carbon emission data acquisition module of the present invention.

[0083] Figure 3 It is a logic block diagram of the carbon emission data center module of the present invention.

[0084] Figure 4 This is a logic block diagram of the carbon emission data calculation module of the present invention.

[0085] Figure 5 It is a logic block diagram of the carbon emission data status assessment module of the present invention.

[0086] Figure 6 This is a logical block diagram of the edge computing service layer module of the present invention. DETAILED DESCRIPTION

[0087] See also Figure 1-6 The present invention provides a highway service area carbon emission health status assessment system and method. The system includes a carbon emission data collection module 1, a carbon emission data center module 2, a carbon emission data calculation module 3, a carbon emission data status assessment module 4 and a carbon emission data visualization module 5.

[0088] The carbon emission data collection module 1 is used to collect carbon emission data of highway service areas in real time, and save historical data in monthly and annual time granularity. Preferably, the present invention uses a variety of sensors to collect data such as traffic flow and energy consumption in the service area to achieve comprehensive monitoring of carbon emission data. For example, equipment such as vehicle flow sensors and energy consumption monitors can be used, and the collection frequency can be set to once every 5 minutes to ensure the real-time and accuracy of the data.

[0089] The carbon emission data center module 2 is communicatively connected to the carbon emission data collection module 1, and is used to receive real-time data and historical data sent by the carbon emission data collection module 1, and pre-process the received data. In one embodiment of the present invention, the pre-processing process includes steps such as data cleaning, outlier processing, and data standardization. For example, for outlier processing, a box plot method can be used to treat data exceeding Q3+1.5IQR or less than Q1-1.5IQR as outliers (where Q1 is the first quartile, Q3 is the third quartile, and IQR is the interquartile range), and replace it with the median.

[0090] The carbon emission data calculation module 3 is connected to the carbon emission data center module 2 for receiving the preprocessed data sent by the carbon emission data center module 2, extracting the spatial features of the preprocessed data using a convolutional neural network (CNN), and inputting the extracted spatial features into a long short-term memory network (LSTM) for classification prediction and regression prediction. The present invention adopts a combined model of CNN and LSTM, making full use of the advantages of CNN in spatial feature extraction and the advantages of LSTM in time series data processing.

[0091] The following are the specific implementations of CNN and LSTM:

[0092] 1.CNN part:

[0093] Assuming that the input data shape is (batch_size, time_steps, features), the number of convolution kernels in the convolution layer is N, and the convolution kernel size is (1, kernel_size), the convolution operation can be expressed as:

[0094] H t =ReLU(W*X t +b),

[0095] Among them, X t is the input data, W is the convolution kernel weight, b is the bias term, * represents the convolution operation, and ReLU is the activation function.

[0096] 2.LSTM part:

[0097] The core calculation formula of LSTM is as follows:

[0098] f t =σ(W f ·[h t-1 ,x t ]+b f ),

[0099] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0100]

[0101] o t =σ(W o ·[h t-1 ,x t ]+b o ),

[0102] h t =o t *tanh(C t ),

[0103] Among them, f t For the forget gate, i t is the input gate, is a candidate memory unit, C t is the memory unit, o t is the output gate, h t is the hidden state, σ is the sigmoid function, tanh is the hyperbolic tangent function, W and b are the weight matrix and bias vector respectively.

[0104] The carbon emission data status assessment module 4 is connected to the carbon emission data calculation module 3 for receiving the prediction results of the carbon emission data calculation module 3 and classifying and assessing the carbon emission status based on a predetermined health status assessment indicator system. The present invention has designed a complete health status assessment indicator system, including multiple dimensions such as carbon emission intensity and carbon emission growth rate. For example, the carbon emission intensity (carbon emissions per unit area) can be divided into the following levels: excellent (≤50kg / m 2 ), good (50-100kg / m 2 ), general (100-150kg / m 2 ), poor (150-200kg / m 2 ) and difference (>200kg / m 2 ).

[0105] The carbon emission data visualization module 5 is connected to the carbon emission data status assessment module 4 in communication, and is used to receive the assessment results of the carbon emission data status assessment module 4, and generate corresponding visualization charts according to the assessment results to provide data visualization display. The present invention uses a variety of visualization technologies, such as heat maps, line graphs, dashboards, etc., to intuitively display the health status of carbon emissions. For example, a heat map can be used to display the carbon emission intensity of different time periods and regions, and a line graph can be used to display the changing trend of carbon emissions.

[0106] The system of the present invention also includes an edge computing service layer module 6, which is in communication connection with the carbon emission data collection module 1. The edge computing service layer module 6 is used to receive the real-time data collected by the carbon emission data collection module 1, perform preprocessing, feature extraction, data classification and regression prediction on the real-time data, and send the calculation results to the carbon emission data center module 2. This edge computing architecture can significantly reduce data transmission delay and improve system response speed.

[0107] In a preferred embodiment of the present invention, the edge computing service layer module 6 includes a GPU computing unit 61 and a hardware adaptation unit 62. The GPU computing unit 61 is used to perform distributed computing based on the GPU and dynamically adjust the allocation of computing resources. For example, the NVIDIA CUDA architecture can be used to accelerate the data processing process by using the parallel computing power of the GPU. The hardware adaptation unit 62 is communicated with the GPU computing unit 61 to realize automatic switching of the GPU and the CPU in extreme temperature environments, and to control the fan for real-time cooling. This design improves the adaptability and reliability of the system under various environmental conditions. For example, when the ambient temperature exceeds 75°C, the system can automatically switch to CPU mode and start a high-speed cooling fan to ensure the normal operation of the equipment.

[0108] The highway service area carbon emission health status assessment system of the present invention realizes full-process management from data collection, processing, analysis to visualization through the collaborative work of multiple modules. The system adopts advanced deep learning algorithms and edge computing technologies to greatly improve the accuracy and real-time performance of carbon emission health status assessment. In this way, the present invention provides strong technical support for carbon emission management in highway service areas, which helps to achieve refined management and energy conservation and emission reduction goals. The system of the present invention further includes a carbon emission data acquisition module 1, which is composed of a multi-source data acquisition unit 11 and an Internet of Things transmission unit 12. The multi-source data acquisition unit 11 is used to collect traffic data, meteorological data and carbon dioxide concentration data. The Internet of Things transmission unit 12 is communicatively connected to the multi-source data acquisition unit 11, and is used to transmit the collected multi-source heterogeneous data to the carbon emission data center module 2 in real time through the Internet of Things technology.

[0109] In a preferred embodiment of the present invention, the multi-source data acquisition unit 11 uses a variety of sensors and detection equipment to work together. For example, traffic data can be collected through vehicle detectors, including information such as vehicle flow, speed and vehicle model; meteorological data can be collected through meteorological stations, including temperature, humidity, wind speed and rainfall; carbon dioxide concentration data can be collected through a dedicated CO2 sensor network. This multi-source data collection method can fully reflect the factors affecting carbon emissions in highway service areas and provide rich data support for subsequent analysis.

[0110] The IoT transmission unit 12 adopts advanced IoT technologies, such as NB-IoT or LoRa, to achieve low-power, wide-coverage data transmission. Preferably, the present invention adopts edge computing technology to perform preliminary processing and compression at the source of the data to reduce the transmission bandwidth requirement and improve the overall efficiency of the system. For example, a data cache and aggregation mechanism can be set at the edge node to transmit summary data every 5 minutes, which not only ensures the real-time nature of the data, but also reduces the network load.

[0111] The carbon emission data calculation module 3 also includes a multi-model fusion prediction unit 31 and an error prediction unit 32. The multi-model fusion prediction unit 31 is used to combine the time series prediction model, the classification prediction model and the regression prediction model to perform multi-dimensional prediction of carbon emissions. The error prediction unit 32 is connected to the multi-model fusion prediction unit 31 in communication to construct a prediction error set and use a long short-term memory network to predict the error to achieve model self-correction.

[0112] In an embodiment of the present invention, the multi-model fusion prediction unit 31 adopts an integrated learning method to comprehensively utilize the advantages of multiple prediction models. Specifically, the time series prediction model can use algorithms such as ARIMA or Prophet, the classification prediction model can use random forest or XGBoost, and the regression prediction model can use support vector regression (SVR) or gradient boosted regression tree (GBRT). The prediction results of these models are fused by weighted averaging or stacking to obtain more stable and accurate prediction results.

[0113] The introduction of the error prediction unit 32 is an innovation of the present invention. The unit constructs an error prediction model by analyzing historical prediction errors, thereby correcting future prediction results. The specific implementation is as follows:

[0114] 1. Error set construction:

[0115]

[0116] Among them, E t is the prediction error at time t, Y t is the actual value, is the predicted value.

[0117] 2.LSTM error prediction model:

[0118] The sliding window method is used to construct the input sequence of LSTM, which is set to 24 (assuming the data is at the hourly level).

[0119]

[0120] 3. Correction of prediction results:

[0121]

[0122] In this way, the system of the present invention can continuously learn and adapt to the pattern of prediction errors, significantly improving prediction accuracy.

[0123] The system of the present invention also includes an adaptive multi-objective parameter optimization module 7, which is in communication with the carbon emission data calculation module 3. The module is used to dynamically adjust the model parameters for different highway service areas and balance the three optimization objectives of prediction accuracy, precision and efficiency.

[0124] In a preferred embodiment of the present invention, the adaptive multi-objective parameter optimization module 7 uses a multi-objective genetic algorithm (NSGA-II) to achieve parameter optimization. The optimization objective function is defined as follows:

[0125] 1. Forecast accuracy: Use mean absolute percentage error (MAPE)

[0126]

[0127] 2. Prediction accuracy: Use root mean square error (RMSE)

[0128]

[0129] 3. Prediction efficiency: Use the model running time T,

[0130] The optimization problem can be expressed as:

[0131] minF(x)=(MAPE(x),RMSE(x),T(x)),

[0132] stx∈X

[0133] Among them, x is the model parameter vector and X is the parameter feasible domain.

[0134] In this way, the system of the present invention can automatically adjust the model parameters according to the characteristics of different highway service areas to achieve dynamic optimization of prediction performance. For example, for service areas with large traffic volume, the system may tend to select a more complex model structure to improve prediction accuracy; while for smaller service areas, a simpler model may be selected to ensure calculation efficiency.

[0135] The carbon emission data center module 2 includes a data preprocessing unit 21, which is used to perform denoising on the received data, extract features from the processed data, and perform smoothing and dimensionality reduction operations on the extracted feature data.

[0136] In one embodiment of the present invention, the processing flow of the data preprocessing unit 21 is as follows:

[0137] 1. Denoising: Use the wavelet transform method, select the appropriate wavelet basis function (such as Daubechies wavelet) and decomposition level, perform multi-scale decomposition on the original signal, and then remove high-frequency noise through the soft threshold method.

[0138] 2. Feature extraction: Extract features using a combination of time domain and frequency domain methods. Time domain features include statistics such as mean, standard deviation, kurtosis, and skewness; frequency domain features are obtained through fast Fourier transform (FFT), including the amplitude and phase information of the main frequency components.

[0139] 3. Smoothing: Use the exponentially weighted moving average (EWMA) method to smooth the data. The calculation formula of EWMA is as follows:

[0140] S t =αY t +(1-α)S t-1 ,

[0141] Among them, S t is the smoothed value, Y t is the original value, and α is the smoothing coefficient (0<α<1). Preferably, the present invention dynamically adjusts the α value according to the volatility of the data to achieve the best smoothing effect.

[0142] 4. Dimensionality reduction operation: Dimensionality reduction is performed using principal component analysis (PCA). PCA converts the original features into linearly independent new features (principal components) through orthogonal transformation. The present invention selects the principal component with a cumulative contribution rate of 95% as the feature after dimensionality reduction, which not only retains the main information of the data, but also significantly reduces the data dimension.

[0143] Through this series of preprocessing steps, the system of the present invention can effectively improve the data quality and lay a solid foundation for the subsequent carbon emission health status assessment. Preferably, the present invention also introduces a data quality monitoring mechanism to monitor the quality of the preprocessed data in real time. If an abnormality is detected (for example, the variance of certain features suddenly decreases significantly), the system will automatically alarm and start the manual intervention process to ensure the reliability of data processing. The carbon emission data status assessment module 4 of the present invention includes a multi-level assessment unit 41 and a SQLite data storage unit 42. The multi-level assessment unit 41 is used to grade the carbon emission health status and assign weight coefficients to different levels. The SQLite data storage unit 42 is communicatively connected to the multi-level assessment unit 41, and is used to store the assessment results using a SQLite database, and provides an efficient data storage and access mechanism.

[0144] In a preferred embodiment of the present invention, the multi-level evaluation unit 41 uses a fuzzy comprehensive evaluation method to evaluate the health status of carbon emissions. The specific steps are as follows:

[0145] 1. Establish the evaluation factor set U = {u 1 ,u 2 ,...,u n}, including carbon emission intensity, carbon emission growth rate, energy structure and other factors.

[0146] 2. Determine the review set V = {v 1 ,v 2 ,v 3 ,v 4 ,v 5}, corresponding to the five levels of excellent, good, average, poor and bad.

[0147] 3. Construct the fuzzy relationship matrix R:

[0148]

[0149] Among them, r ij It represents the membership of the i-th factor to the j-th comment.

[0150] 4. Determine the weight vector A = (a 1 ,a 2 ,...,a n ), satisfying ∑a i = 1. The present invention adopts the analytic hierarchy process (AHP) to determine the weight.

[0151] 5. Calculate the fuzzy comprehensive evaluation results:

[0152] B=A°R=(b 1 ,b 2 ,b 3 ,b 4,b 5 )

[0153] in, Represents a fuzzy composite operation.

[0154] 6. Determine the final evaluation level based on the principle of maximum membership.

[0155] Preferably, the system of the present invention will adjust the weights of the evaluation factors regularly (e.g., quarterly) to adapt to the carbon emission characteristics under different seasons and policy environments. For example, the weight of heating-related factors may be increased in winter, and the weight of energy structure factors may be increased after the implementation of new energy policies.

[0156] The SQLite data storage unit 42 uses a lightweight SQLite database, which has the characteristics of occupying less resources and fast response speed, and is very suitable for embedded systems. The present invention designs an optimized database architecture, which mainly includes the following table structure:

[0157] 1. Assessment results table (assessment_results):

[0158] id:INTEGERPRIMARYKEY

[0159] timestamp:DATETIME

[0160] service_area_id:INTEGER

[0161] overall_score:REAL

[0162] grade:TEXT

[0163] factor_scores: TEXT (JSON format stores the scores of each factor)

[0164] 2. Assessment factor table (assessment_factors):

[0165] id:INTEGERPRIMARYKEY

[0166] factor_name:TEXT

[0167] weight:REAL

[0168] update_time:DATETIME

[0169] Through this design, the system of the present invention can efficiently store and retrieve evaluation results, supporting rapid historical data analysis and trend identification. Preferably, the present invention also implements data compression and periodic archiving mechanisms to save storage space and maintain query performance.

[0170] The system of the present invention also includes an intelligent vital sign monitoring strategy optimization module 8, which is in communication connection with the carbon emission data status evaluation module 4. This module is used to evaluate the carbon emission status in real time based on the deep learning algorithm, and to adaptively adjust the monitoring strategy according to the evaluation results.

[0171] In one embodiment of the present invention, the intelligent vital sign monitoring strategy optimization module 8 adopts a deep reinforcement learning algorithm, specifically a double deep Q network (DoubleDQN), to achieve dynamic optimization of the monitoring strategy. The core idea is to model the carbon emission monitoring problem as a Markov decision process (MDP) and learn the optimal monitoring strategy through interaction with the environment.

[0172] The implementation steps are as follows:

[0173] 1. Define the state space S: including current carbon emission levels, historical monitoring data, environmental factors, etc.

[0174] 2. Define action space A: including adjusting monitoring frequency, switching monitoring equipment, starting specific monitoring programs, etc.

[0175] 3. Define the reward function R:

[0176] R=w 1 *accuracy+w 2 *efficiency-w 3 *cost,

[0177] Among them, accuracy is the monitoring accuracy, efficiency is the monitoring efficiency, cost is the monitoring cost, and w 1 、w 2 、w 3 is the weight coefficient.

[0178] Construct a dual-depth Q network: including the current network Q and the target network Q′, both of which adopt a multi-layer perceptron (MLP) structure.

[0179] 5. Experience replay: Use the experience replay buffer to store the transfer samples (s, a, r, s′).

[0180] 6. Training process:

[0181] Sample batches of data from the experience replay buffer;

[0182] Calculate the target Q value: y i =ri +γQ′(s′ i ,argmax a Q(s′ i ,a;θ);θ′);

[0183] Update the current network parameters:

[0184] Update the target network every C steps: θ′←θ;

[0185] In this way, the system of the present invention can automatically adjust the monitoring strategy according to the real-time carbon emission status and environmental changes, and achieve the best balance between monitoring accuracy, efficiency and cost. For example, when abnormal fluctuations in carbon emissions are detected, the system may automatically increase the monitoring frequency and start additional monitoring equipment; while when carbon emissions are stable, the monitoring frequency may be reduced to save resources.

[0186] Finally, the present invention also provides a method for evaluating the carbon emission health status using the above highway service area carbon emission health status evaluation system. The method includes the following steps:

[0187] S1: Collecting real-time carbon emission data and historical data of highway service areas through carbon emission data collection module 1;

[0188] S2: Transmit the collected data to the carbon emission data center module 2 for preprocessing;

[0189] S3: Use the convolutional neural network in the carbon emission data calculation module 3 to extract the spatial features of the preprocessed data, and input the extracted features into the long short-term memory network for classification prediction and regression prediction;

[0190] S4: The carbon emission data status assessment module 4 classifies and assesses the prediction results based on a predetermined health status assessment indicator system;

[0191] S5: The carbon emission data visualization module 5 generates corresponding visualization charts according to the evaluation results;

[0192] S6: The edge computing service layer module 6 performs preprocessing, feature extraction, data classification and regression prediction on the collected real-time data, and sends the results to the carbon emission data center module 2;

[0193] S7: Adaptive multi-objective parameter optimization module 7 dynamically adjusts model parameters according to the characteristics of different highway service areas;

[0194] S8: The intelligent vital sign monitoring strategy optimization module 8 evaluates the carbon emission status in real time based on the deep learning algorithm and makes adaptive adjustments to the monitoring strategy.

[0195] In a preferred embodiment of the present invention, the data collection frequency of step S1 is dynamically adjusted according to the size of the service area and the traffic flow. For example, for a large service area, the collection frequency can be set to once every 5 minutes; while for a small service area, it can be reduced to once every 15 minutes to balance data accuracy and system load.

[0196] In step S3, the specific structure of the convolutional neural network is: 3 convolutional layers (the convolution kernel sizes are 3x3, 5x5, and 7x7 respectively), each convolutional layer is followed by a maximum pooling layer and a batch normalization layer, and finally the features are output through a fully connected layer. The LSTM network adopts a bidirectional structure, the number of hidden layer neurons is 128, and the time step is set to 24 (corresponding to 24 hours of data).

[0197] The health status assessment indicator system of step S4 includes the following key indicators:

[0198] 1. Carbon emission intensity: Carbon emissions per unit area (kg / m 2 );

[0199] 2. Carbon emission growth rate: month-on-month growth rate (%);

[0200] 3. Energy structure: proportion of clean energy use (%);

[0201] 4. Peak load: the ratio of maximum hourly carbon emissions to daily average carbon emissions (%);

[0202] 5. Carbon neutrality progress: the deviation between actual carbon emissions and the carbon neutrality target (%);

[0203] Through this systematic and intelligent method, the present invention can comprehensively, accurately and in real time evaluate the carbon emission health status of highway service areas, and provide powerful decision-making support for service area managers.

[0204] In order to verify the effectiveness and superiority of the highway service area carbon emission health status assessment system and method of the present invention, the present invention selected three service areas of different sizes on a highway in Jiangsu Province as test objects and conducted a 6-month simulation test. The test period covered the spring and summer seasons, and fully considered the impact of climate change, holiday passenger flow fluctuations and other factors on carbon emissions.

[0205] Example 1 uses the complete system of the present invention, including core technologies such as multi-source data acquisition, edge computing, deep learning prediction, and adaptive parameter optimization. Comparative Example 1 uses the traditional fixed threshold monitoring method and evaluates only based on a single carbon emission sensor data. Comparative Example 2 uses a simple machine learning model (such as random forest), but does not include edge computing and adaptive optimization functions.

[0206] The test focuses on the following five core indicators: prediction accuracy, system response time, anomaly detection rate, energy consumption optimization effect and scalability. The testing standards and methods of these indicators are as follows:

[0207] 1. Forecast accuracy: measured by mean absolute percentage error (MAPE), calculated by comparing the predicted values ​​with the actual observed values.

[0208] 2. System response time: measures the average time from data collection to generating evaluation results.

[0209] 3. Anomaly detection rate: Artificially set some abnormal carbon emission scenarios and calculate the proportion of successful identification by the system.

[0210] 4. Energy consumption optimization effect: Compare the changes in average daily carbon emissions in the service area before and after the implementation of the system.

[0211] 5. Scalability: Test the system's adaptability to service areas of different sizes, including deployment time and performance stability.

[0212] The test results are shown in the following table:

[0213] index Example 1 Comparative Example 1 Comparative Example 2 Forecast accuracy (MAPE) 3.2 15.7 8.9 System response time 0.5 5.2 2.1 Anomaly detection rate 97.5 62.3 85.1 Energy consumption optimization effect -18.6 -5.2 -10.3 Scalability (deployment time) 2 days 7 days 4 days

[0214] It can be clearly seen from the test results that the system of the present invention is significantly superior to traditional methods and simple machine learning methods in all key indicators. It is particularly noteworthy that the prediction accuracy of the present invention reaches 96.8% (MAPE is 3.2%), which is a very outstanding result in a complex service area environment. This is mainly due to the fact that the deep learning model used in the present invention can effectively capture the spatiotemporal characteristics of carbon emissions, and the adaptive parameter optimization technology can dynamically adjust the model according to the characteristics of different service areas.

[0215] The significant improvement in system response time (only 0.5 seconds) fully demonstrates the advantages of edge computing technology. By preprocessing and preliminary analysis at the source of the data, the present invention significantly reduces the amount of data transmitted and the computing burden of the central server, thereby achieving near-real-time carbon emission status assessment. This is crucial for timely detection and handling of abnormal carbon emission situations.

[0216] In terms of anomaly detection, the system of the present invention performs well, with a detection rate of up to 97.5%. This is due to the application of multi-source data fusion and deep learning algorithms, which enable the system to identify subtle changes in carbon emission patterns. For example, in one test, the system successfully detected a slight but continuous increase in carbon emissions caused by a malfunction in the air conditioning system, an anomaly that was completely ignored in traditional methods.

[0217] Energy consumption optimization is one of the most significant advantages of the present invention. Through real-time monitoring and intelligent suggestions, the average daily carbon emissions of the service area have been reduced by 18.6%. This has not only brought considerable economic benefits, but also made an important contribution to the sustainable development of the service area. For example, the system recommends adjusting the operating parameters of the refrigeration equipment during off-peak hours, which alone saves about 5% of energy consumption.

[0218] In terms of scalability, the system of the present invention also shows obvious advantages. Thanks to modular design and adaptive optimization technology, the system can be deployed and debugged in service areas of different sizes in just 2 days, and the performance remains stable. This greatly reduces the cost and difficulty of system promotion.

[0219] The best implementation is the application in large service areas (average daily traffic volume exceeds 10,000 vehicles). In this complex environment, the system of the present invention gives full play to its powerful data processing and analysis capabilities. Especially during peak passenger flow periods such as the National Day holiday, the system predicts passenger flow and carbon emission trends and gives a series of optimization suggestions in advance, such as temporarily adding new energy vehicle charging piles and optimizing the layout of catering areas, which effectively reduces carbon emission intensity.

[0220] These test results fully demonstrate the innovation and practicality of the present invention in the field of carbon emission health status assessment of highway service areas. The system can not only accurately and in real time assess the carbon emission status, but also effectively help service areas reduce carbon emissions and promote green and low-carbon development through intelligent optimization suggestions. In the future, with the accumulation of more service area data and further optimization of the algorithm, the performance of this system is expected to be further improved.

[0221] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Highway service area carbon emission health status assessment system, characterized by: include: Carbon emission data collection module, used for: Real-time collection of carbon emission data from highway service areas; Save historical data at monthly and yearly granularity; The carbon emission data center module is connected to the carbon emission data collection module for: Receiving real-time data and historical data sent by the carbon emission data collection module; Preprocessing the received data; The carbon emission data calculation module is connected to the carbon emission data center module for: Receiving preprocessed data sent by the carbon emission data center module; Extracting spatial features of the preprocessed data using a convolutional neural network; The extracted spatial features are input into the long short-term memory network for classification prediction and regression prediction; The carbon emission data status assessment module is in communication with the carbon emission data calculation module and is used to: Receiving the prediction result of the carbon emission data calculation module; Based on the predetermined health status assessment indicator system, classify and assess the carbon emission status; The carbon emission data visualization module is in communication with the carbon emission data status assessment module and is used to: Receiving the evaluation result of the carbon emission data status evaluation module; Generate corresponding visualization charts based on the evaluation results to provide data visualization display.

2. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: Also includes: The edge computing service layer module is connected to the carbon emission data collection module for: Receiving real-time data collected by the carbon emission data collection module; Preprocess, extract features, classify data and perform regression prediction on real-time data; The calculation results are sent to the carbon emission data center module.

3. The highway service area carbon emission health status assessment system according to claim 2 is characterized in that: The edge computing service layer module includes: GPU computing unit, used for: Distributed computing based on GPU; Dynamically adjust computing resource allocation; A hardware adapter unit is connected to the GPU computing unit for: Automatic switching between GPU and CPU in extreme temperature environments; Control the fan to cool down in real time.

4. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: The carbon emission data collection module includes: Multi-source data acquisition unit for: Collect traffic data, weather data and carbon dioxide concentration data; The Internet of Things transmission unit is communicatively connected with the multi-source data acquisition unit and is used for: The collected multi-source heterogeneous data are transmitted in real time to the carbon emission data center module through the Internet of Things technology.

5. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: The carbon emission data calculation module also includes: Multi-model fusion prediction unit, used for: Combine time series prediction model, classification prediction model and regression prediction model to predict carbon emissions from multiple angles; The error prediction unit is communicatively connected with the multi-model fusion prediction unit and is used for: Constructing a set of prediction errors; Long short-term memory networks are used to predict errors and achieve model self-correction.

6. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: Also includes: The adaptive multi-objective parameter optimization module is in communication with the carbon emission data calculation module and is used to: Dynamically adjust model parameters for different highway service areas; Balance the three optimization goals of prediction accuracy, precision and efficiency.

7. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: The carbon emission data center module includes: Data pre-processing unit, used for: De-noising the received data; Perform feature extraction on the processed data; The extracted feature data are smoothed and dimensionally reduced.

8. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: The carbon emission data status assessment module includes: Multi-level evaluation unit for: Rating the health status of carbon emissions; Assign weight coefficients to different levels; A SQLite data storage unit is communicatively connected to the multi-level evaluation unit, and is used for: Use SQLite database to store evaluation results; Provides efficient data storage and access mechanism.

9. The highway service area carbon emission health status assessment system according to claim 1 is characterized in that: Also includes: The intelligent vital sign monitoring strategy optimization module is connected to the carbon emission data status assessment module for: Based on deep learning algorithms, real-time assessment of carbon emission status; Based on the evaluation results, the monitoring strategy is adaptively adjusted.

10. A method for evaluating the health status of carbon emissions in a highway service area using the highway service area carbon emissions health status evaluation system according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Collect real-time carbon emission data and historical data of highway service areas through the carbon emission data collection module; S2: Transmit the collected data to the carbon emission data center module for preprocessing; S3: Use the convolutional neural network in the carbon emission data calculation module to extract the spatial features of the preprocessed data, and input the extracted features into the long short-term memory network for classification prediction and regression prediction; S4: The carbon emission data status assessment module classifies and assesses the prediction results based on a predetermined health status assessment indicator system; S5: The carbon emission data visualization module generates corresponding visualization charts according to the evaluation results; S6: The edge computing service layer module preprocesses, extracts features, classifies data, and performs regression prediction on the collected real-time data, and sends the results to the carbon emission data center module; S7: Adaptive multi-objective parameter optimization module dynamically adjusts model parameters according to the characteristics of different highway service areas; S8: The intelligent vital sign monitoring strategy optimization module evaluates the carbon emission status in real time based on the deep learning algorithm and makes adaptive adjustments to the monitoring strategy.

Images