Intelligent carbon emission analysis system for highway tunnel lighting

Abstract

The invention discloses a carbon emission intelligent analysis system for highway tunnel illumination, and the system comprises a demand analysis module which is used for calculating the demand illumination brightness of a plurality of illumination devices in a tunnel according to the traffic flow, the natural illuminance outside the tunnel and the PM2.5 value in the tunnel which are collected in real time; the carbon effect analysis module is used for constructing a tunnel carbon emission model and performing uncertainty quantification on the tunnel carbon emission model by utilizing Monte Carlo simulation to obtain a carbon emission confidence interval; and the environment optimization module is used for optimizing equipment control parameters by utilizing a reinforcement learning framework when the upper limit of the carbon emission confidence interval is greater than a preset threshold value to obtain equipment optimization parameters, and controlling equipment by utilizing the equipment optimization parameters. The invention relates to the technical field of carbon emission optimization, and solves the technical problem that an existing carbon emission analysis method is insufficient in dynamic adaptability.

Description

Technical Field

[0001] The present invention belongs to the field of highway tunnels and relates to carbon emission optimization technology, specifically to an intelligent carbon emission analysis system for highway tunnel lighting. Background Art

[0002] As a vital component of transportation infrastructure, highway tunnels face increasing challenges with their lighting systems' energy consumption and carbon emissions. Tunnel lighting's carbon emissions include not only electricity consumption during operation but also the implicit emissions throughout the equipment's lifecycle, including production, transportation, and maintenance. Accurately calculating and optimizing tunnel lighting's carbon emissions is crucial for reducing the transportation sector's carbon footprint and promoting green transportation.

[0003] However, existing tunnel lighting carbon emission analysis methods often use fixed parameters for calculations, relying on preset lighting standards and equipment parameters. This makes it difficult to dynamically adjust lighting strategies based on real-time conditions. This results in lighting brightness failing to meet actual needs during peak traffic hours or in adverse weather conditions, which can increase the risk of traffic accidents. When natural light is sufficient, the brightness of artificial lighting cannot be reduced in a timely manner, resulting in serious energy waste and further increasing the carbon emission burden. Furthermore, due to a lack of comprehensive consideration of the embodied carbon emissions throughout the entire life cycle of equipment production, transportation, and maintenance, carbon emission assessments of tunnel lighting systems are insufficiently comprehensive and accurate, failing to provide a reliable basis for the development of energy-saving and emission-reduction measures. Furthermore, existing methods fail to fully consider the impact of lighting on driver and passenger comfort, nor do they establish an effective feedback mechanism to optimize lighting solutions, making it difficult to achieve intelligent and green management of lighting systems. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an intelligent carbon emission analysis system for highway tunnel lighting, which is used to solve the technical problem of insufficient dynamic adaptability of existing carbon emission analysis methods.

[0005] To achieve the above objectives, a first aspect of the present invention provides an intelligent carbon emissions analysis system for highway tunnel lighting, comprising: Demand analysis module: used to collect real-time traffic flow, natural light intensity outside the tunnel, PM inside the tunnel, etc. 2.5 Value, analyze and control the lighting brightness of several lighting equipment in the tunnel; Carbon Efficiency Analysis Module: used to build a tunnel carbon emission model, quantify the uncertainty of the tunnel carbon emission model using Monte Carlo simulation, and obtain the carbon emission placement confidence interval; Environmental optimization module: used to optimize the equipment control parameters using the reinforcement learning framework when the upper limit of the carbon emission placement confidence interval is greater than the preset threshold, obtain the equipment optimization parameters, and use the equipment optimization parameters to control the equipment; wherein, the equipment includes lighting equipment and ventilation equipment.

[0006] Furthermore, the analysis and control of the brightness of a plurality of lighting devices in the tunnel includes: A1, divide the tunnel into several segments according to the preset distance, and collect the traffic flow V(t) and PM in the tunnel in real time. 2.5 Value PM(t), natural light illumination outside the tunnel L ext (t); where t represents the current time; A2, based on PM in the tunnel 2.5 The visibility C(t) is calculated from the value PM(t) using the following formula: ; A3, based on the traffic volume V(t), visibility C(t), and natural light intensity L outside the tunnel ext (t), calculate the basic lighting brightness I of several segments base (t), the calculation formula is: ; Among them, α, β, γ represent dynamic learning parameters, which are obtained by online updating of the neural network model, V max Indicates the maximum traffic capacity of the tunnel, L base (x) represents the segmented reference illuminance; A4, setting segment correction coefficients S(x) for several segments, calculating segment lighting brightness based on the segment correction coefficients, and controlling several lighting devices based on the segment lighting brightness; wherein x represents a segment index.

[0007] Furthermore, the calculating of the segmented lighting brightness according to the segmented correction coefficient includes: A41, according to the formula I'(t)=I base (t)×S(x) is used to calculate the required illumination I'(t) of several segments; A42: Determine whether the required illumination of several segments I'(t) is greater than or equal to the segment reference illumination L base (x); If yes, the required illuminance I'(t) is marked as the segmented lighting brightness I(t); if no, the segmented reference illuminance L base (x) is marked as the segment illumination brightness I(t).

[0008] By collecting real-time traffic flow, PM 2.5 The system uses data from the user's location and natural light intensity data, combined with visibility and dynamic learning parameters, to achieve adaptive adjustment of lighting brightness. The system also uses a neural network to update parameters online, ensuring that the lighting brightness meets safety requirements while maximizing the use of natural light and reducing energy consumption. This solves the energy waste and lack of adaptability of traditional fixed illumination modes, achieving a balance between energy conservation and safety.

[0009] Furthermore, the dynamic learning parameters are obtained by performing online updating of a neural network model, including: A31, based on the data collection cycle, uses the sliding window method to obtain the time series data within the most recent preset time period to obtain the input features; The time series data includes traffic volume V, visibility C, and natural light intensity L outside the tunnel. ext , traffic flow gradient , and the traffic flow gradient is expressed by the formula get, Indicates a preset time interval; A32, performs standardization preprocessing on the input features to obtain preprocessed data; A33 constructs a recurrent neural network based on a deep learning algorithm and sets a dynamic loss function, online incremental training parameters, and initial dynamic parameters α0, β0, and γ0 of the recurrent neural network to obtain an online updated model; wherein the online incremental training parameters include an optimizer, a number of iterations, and a gradient norm threshold; A34, according to the update cycle, inputs the preprocessed data into the online update model to update the parameters and obtain the dynamic parameter increment 、 、 ; A35, update the formula according to the parameters , get the dynamic learning parameters; among them, Indicates the learning rate for parameter update.

[0010] Furthermore, the process of constructing the tunnel carbon emission model includes: B1-1, collect the rated power P of lighting equipment rated , rated illumination I rated , unit production carbon emissions M and design life L, and obtain the grid carbon emission factor λ(t) based on the State Grid real-time data interface; B1-2, calculate the real-time power P(t) of the lighting equipment based on the rated power. The calculation formula is: ; B1-3, calculate the equivalent operating time T(t) of the lighting equipment based on the rated illuminance. The calculation formula is: ;in, represents the integral variable, t0 represents the starting time of carbon emission monitoring; B1-4, construct the tunnel carbon emission model as follows: ; Where i represents the lighting device index, n represents the number of lighting devices, The implicit carbon emission allocation coefficient of the i-th lighting equipment is expressed by the formula Calculated, Indicates the real-time carbon emissions in the tunnel, Indicates the installation time.

[0011] A comprehensive model that includes real-time operating energy consumption and implicit carbon emissions is constructed, in which the real-time power and equivalent operating time can accurately reflect the actual energy consumption of the equipment. By correlating the rated parameters with the real-time operating conditions, a refined calculation of the carbon emissions of the entire life cycle of tunnel lighting is achieved, providing a basis for carbon emission analysis and optimization.

[0012] Furthermore, the use of Monte Carlo simulation to quantify uncertainty in the tunnel carbon emission model includes: B2-1, define the input parameters and parameter distribution as follows: vehicle flow V(t) follows Poisson distribution, visibility C(t) follows normal distribution, and grid carbon emission factor λ(t) follows uniform distribution; B2-2, setting the number of iterations of the Monte Carlo model, and in each iteration, obtaining the simulated traffic flow value, simulated visibility value, and simulated grid carbon emission factor value by sampling to obtain the sampled simulation value; B2-3, substituting the sampled simulation values ​​of each iteration into the tunnel carbon emission model to obtain simulated carbon emission values, and storing them in an array to obtain a carbon emission simulation data set; B2-4, obtain the simulated mean AE and simulated standard deviation σ based on the carbon emission simulation data set E , calculate E using the formula CO2 ∈[AE-1.96σ E ,AE+1.96σ E ]Get the carbon emission confidence interval E CO2 Among them, AE+1.96σ E Place an upper bound on the confidence interval for carbon emissions.

[0013] By defining parameter distributions for traffic volume, visibility, and power grid factors and conducting multiple Monte Carlo simulations, we generate confidence intervals for carbon emissions, quantifying the impact of fluctuations in input parameters on carbon emissions and providing decision makers with a risk assessment range. Determining confidence intervals using simulated means and standard deviations addresses the inability of traditional models to reflect uncertainty and enhances the robustness of decision-making.

[0014] Furthermore, the optimization of the device control parameters using the reinforcement learning framework includes: C1, set the state space of the reinforcement learning framework to: traffic flow gradient , predicted visibility , grid carbon emission factor λ(t) and remaining carbon budget E re (t); where the remaining carbon budget E re The calculation formula of (t) is: re(t) = E quota -E total (t), E quota Indicates the preset carbon emission limit, t represents the preset time interval.

[0015] C2, set the action space of the reinforcement learning framework to: the dimming level l of the lighting equipment i , fan speed f of ventilation equipment j And the equipment maintenance instruction m, the action vector a(t)=[l i ∈{l1,l2,…,l p},f j ∈{f1,f2,…,f q},m∈{0,1}]; where p represents the number of dimming level parameters, q represents the number of fan speed parameters, and the equipment maintenance instruction is used to send a signal indicating whether the lighting equipment should be replaced. When m=1, it indicates sending a lighting equipment replacement signal, and when m=0, it indicates maintaining the current state of the lighting equipment. C3, set the reward function of the reinforcement learning framework to: ; Among them, R1 represents the carbon emission reward function, R2 represents the comfort reward function, R3 represents the safety penalty function, w j Represents the weight coefficient, which is determined based on historical experience; C4, collects real-time state space data to obtain the state vector, inputs the state vector into the reinforcement learning framework for environmental interaction and experience storage learning, and obtains the device optimization parameters; The device optimization parameters are the output of the reinforcement learning framework, including the dimming level of the lighting equipment, the fan speed of the ventilation equipment, and the equipment maintenance instructions.

[0016] The reinforcement learning framework can dynamically perceive the real-time changes in the tunnel's internal and external environment by collecting real-time traffic flow gradients, predicted visibility, grid carbon emission factors, and remaining carbon budget data. 2.5 When a sudden increase in light value causes visibility to decrease, the system can quickly adjust lighting brightness and ventilation equipment parameters through the updated strategy network to ensure safety while optimizing carbon emissions.

[0017] Furthermore, the step of inputting the state vector into the reinforcement learning framework for environment interaction and experience storage learning includes: C41, using the multi-objective reinforcement learning algorithm to initialize the reinforcement learning framework, including initializing the parameters of the reinforcement learning framework's policy network, value network, and target network, and setting hyperparameters such as the learning rate, discount factor, and soft update coefficient; C42 creates an experience replay buffer to store the agent's experience of interacting with the environment. The agent represents the system that controls the lighting and ventilation equipment in the tunnel. It generates action vectors based on the state vector through the decision network and observes the feedback from the environment. C43, inputs the state vector s(t) into the reinforcement learning framework and generates the action vector a(t) according to the policy network; C44, executes action a(t), obtains the next state vector s(t+1), the calculation result of the reward function r(t), and the mark d(t+1) indicating whether to terminate learning; C45, stores the experience tuple (s(t), a(t), r(t), s(t+1), d(t+1)) into the experience replay buffer; C46, repeat C43 to C45 until the number of experience tuples in the experience replay buffer reaches the batch size threshold, randomly sampling a batch of experience tuples from the experience replay buffer to obtain experience sampling; C47, uses experience sampling to update the reinforcement learning framework and calculates the average reward value of the last n rounds and the average reward value of the first n rounds ,judge and Is the difference less than the preset change threshold? If yes, terminate the learning and obtain the output result; if not, repeat C43 to C46.

[0018] By collecting real-time status data and interacting with the environment, reinforcement learning can accumulate experience from actual operation and gradually optimize control strategies. When making equipment maintenance decisions, the system can analyze historical data to learn the optimal time to replace lighting equipment, balancing equipment lifespan and energy costs.

[0019] Furthermore, the predicted visibility is obtained by predicting a visibility prediction model constructed based on a machine learning algorithm; The visibility prediction model is constructed as follows: D1, collect historical meteorological data, traffic volume and visibility, and arrange them in chronological order to obtain time series data; wherein, the meteorological data includes air humidity, temperature, wind speed and air pressure, and the visibility C(t) is measured by PM in the tunnel. 2.5 The value PM(t) is obtained, and the calculation formula is: ; D2, build a time series model based on machine learning algorithms and define the model's loss function and training parameters; D3, preprocess the time series data and divide the preprocessed data into training sets, validation sets, and test sets according to the preset ratio; D4: Input the training set and validation set into the model for iterative training and parameter update. Use the test set to evaluate whether the accuracy of the updated model is greater than the preset accuracy threshold. If yes, save the updated model and output it as a visibility prediction model for visibility prediction. If no, adjust the model's loss function and training parameters, and repeat D4. The input of the visibility prediction model includes meteorological data and traffic flow, and the output is predicted visibility.

[0020] Furthermore, the reward function of the reinforcement learning framework includes: The carbon emission reward function R1 is: ; The comfort reward function R2 is: ; The security penalty function is R3: ;in, Represents an indicative function.

[0021] Compared with the prior art, the present invention has the following beneficial effects: The present invention collects traffic flow, PM 2.5 The system uses data such as road conditions and natural light intensity, and combines it with a neural network to dynamically update learning parameters online to accurately calculate required illumination. It also optimizes equipment parameters through a reinforcement learning framework, minimizing energy consumption and carbon emissions while ensuring driving comfort and safety. For example, when visibility is predicted to decrease, the system automatically increases lighting brightness and adjusts ventilation equipment. It also uses a multi-objective reward function to balance energy conservation and safety requirements. The present invention integrates real-time operating energy consumption and equipment implicit carbon emissions, and quantifies the uncertainty of input parameters through Monte Carlo simulation to generate carbon emission placement confidence intervals, providing decision makers with a basis for risk assessment, enabling the system to predict carbon exceeding risks in advance and adjust strategies; in addition, the reinforcement learning framework continuously optimizes the control strategy through an online learning mechanism to achieve autonomous updating of equipment parameters. For example, it can adjust the dimming level according to the real-time changes in the carbon emission factor of the power grid, ensuring long-term optimality and improving the green and intelligent level of tunnel operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0023] Figure 1 A schematic diagram of the technical process of a carbon emission intelligent analysis system for highway tunnel lighting provided by the present invention; Figure 2 This is a schematic diagram of the workflow of the demand analysis module in the carbon emission intelligent analysis system for highway tunnel lighting provided by the present invention. DETAILED DESCRIPTION

[0024] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] See also Figure 1-Figure 2 The first embodiment of the present invention provides a carbon emission intelligent analysis system for highway tunnel lighting, comprising: Demand analysis module: used to collect real-time traffic flow, natural light intensity outside the tunnel, PM inside the tunnel, etc. 2.5 Value, analyze and control the lighting brightness of several lighting equipment in the tunnel; Carbon Efficiency Analysis Module: used to build a tunnel carbon emission model, quantify the uncertainty of the tunnel carbon emission model using Monte Carlo simulation, and obtain the carbon emission placement confidence interval; Environmental optimization module: When the upper limit of the carbon emission placement confidence interval is greater than a preset threshold, the reinforcement learning framework is used to optimize the equipment control parameters, obtain the equipment optimization parameters, and use the equipment optimization parameters to control the equipment; the equipment includes lighting equipment and ventilation equipment.

[0026] It should be noted that the demand analysis module, carbon efficiency analysis module and environment optimization module of the present invention are communicatively connected.

[0027] Based on the aforementioned technical modules, this invention achieves intelligent, low-carbon management of highway tunnel lighting systems. The demand analysis module collects data in real time and calculates required illumination, providing a basis for lighting equipment control. The carbon efficiency analysis module quantifies carbon emissions, providing a decision-making basis for the environmental optimization module. The environmental optimization module utilizes a reinforcement learning framework to optimize equipment control parameters based on carbon emissions and other status information, achieving energy conservation, emission reduction, and multi-objective balance. The visibility prediction model provides predictive information for the entire system, assisting the reinforcement learning framework in making more accurate decisions.

[0028] In addition to vehicle exhaust, carbon emissions from highway tunnels also include emissions from the energy consumption of lighting systems, such as energy consumption during the production and operation of lighting equipment, as well as carbon emissions caused by related power grid transmission losses. Therefore, to effectively reduce carbon emissions, the demand analysis module collects key data such as real-time vehicle flow, natural light levels outside the tunnel, and haze concentrations inside the tunnel. This model constructs a dynamic lighting demand model, accurately analyzing the required illumination levels of various lighting devices within the tunnel under different environmental and traffic conditions. This ensures optimal energy utilization, avoids energy waste caused by excessive lighting, and reduces carbon emissions from the lighting system at the source.

[0029] Specifically, in the demand analysis module, since the natural light intensity obtained at different locations inside the tunnel is different, the tunnel is first divided into several segments according to the preset distance, and then the traffic flow V(t) and PM in the tunnel of several segments are collected in real time. 2.5 Value PM(t), natural light illumination outside the tunnel L ext (t) to reflect the real-time changes in the internal and external environment of the tunnel; Then, according to the haze concentration PM in the tunnel 2.5 The visibility C(t) is calculated from the value PM(t), and the calculation formula is: , to consider the impact of haze on visibility, providing key parameters for subsequent accurate calculation of lighting needs; Then, according to the traffic volume V(t), visibility C(t), and natural light intensity L outside the tunnel ext (t), construct the dynamic lighting model as: , and calculate the basic lighting brightness I of several segments in real time base (t); where V max Indicates the maximum traffic capacity of the tunnel, obtained from the engineering design documents; L base (x) represents the segmented benchmark illumination, which is determined by the "Highway Tunnel Design Regulations"; α, β, and γ represent dynamic learning parameters, which are obtained by online updating of the neural network model; Then set the segment reference illumination L base (x) and the segment correction coefficient S(x), x represents the segment index; then according to the formula I'(t)=I base (t)×S(x) is used to calculate the required illumination I'(t) of several segments; At the same time, it is also necessary to determine whether the required illumination of several segments I'(t) is greater than or equal to the segment reference illumination L base (x); If yes, then use the required illuminance I'(t) to control the lighting equipment; if no, use the segmented reference illuminance L base (x) Control lighting equipment.

[0030] In one embodiment, the dynamic learning parameters in the dynamic lighting model are obtained by online updating of the neural network model, which may specifically include the following steps: A31, based on the data collection cycle, uses the sliding window method to obtain the time series data within the most recent preset time period to obtain the input features; Among them, the time series data includes traffic volume V, visibility C, and natural light illumination outside the tunnel L ext , traffic flow gradient , and the traffic flow gradient is expressed by the formula get, Indicates a preset time interval; A32, performs standardization preprocessing on the input features to obtain preprocessed data; A33 constructs a recurrent neural network based on a deep learning algorithm and sets the dynamic loss function, online incremental training parameters, and initial dynamic parameters α0, β0, and γ0 of the recurrent neural network to obtain an online updated model. The online incremental training parameters include the optimizer, number of iterations, and gradient norm threshold. A34, according to the update cycle, inputs the preprocessed data into the online update model to update the parameters and obtain the dynamic parameter increment 、 、 ; A35, update the formula according to the parameters , get the dynamic learning parameters; among them, Indicates the learning rate for parameter updates, which is determined based on historical experience and is usually 0.01.

[0031] The dynamic learning parameters obtained through online updates enable the dynamic demand lighting model to continuously adapt to changes in the environment and traffic conditions, and continuously optimize the calculation of lighting brightness, thereby minimizing the energy consumption and carbon emissions of the lighting system while meeting tunnel lighting needs.

[0032] Assume that there is a 1000-meter-long twin-hole tunnel divided into the following four sections, and the collected data for each section is shown in Table 1: Table 1 is a sample table for segmented data collection:

[0033] Since the middle section does not receive natural light, in actual applications, the lighting equipment in the middle section can be directly set according to the benchmark illuminance value. For the entrance and exit sections, the optimal illuminance is calculated using a dynamic lighting model and a segmented correction coefficient: At a certain moment, the collected traffic flow data is: V(t) = 800 vehicles / h; the maximum traffic capacity of the tunnel is V max=1200 vehicles / h; the natural light intensity outside the tunnel is L ext (t) = 5000 lux; initial values ​​of dynamic parameters are: α = 1.2, β = 0.8, γ = 0.5; (1) Entry section calculation process Visibility: C(t)=1 / (1+0.02×30)≈0.625; Basic lighting brightness: I base (t)=1.2×(800 / 1200)^0.8×e^[0.5×(1-0.625)]×(5000 / 80)≈65.8lux; Required illumination: I'(t)=65.81×[max(1.0,5000 / 1500)]≈219lux; (2) Calculation process of the export section Visibility: C(t)=1 / (1+0.02×40)≈0.556; Basic lighting brightness: I base (t)=1.2×(800 / 1200)^0.8×e^[0.5×(1-0.556)]×(5000 / 90)≈58.7lux; Required illumination: I'(t)=58.7×[min(1.5,5000 / 1000)]=88.1lux<90lux, with 90lux as the segmented lighting brightness I(t) for control.

[0034] In this embodiment, by collecting real-time traffic flow, PM 2.5 Based on data such as concentration and external light intensity, a dynamic lighting model is constructed and combined with segmented correction coefficients to precisely adjust the illumination in each tunnel zone, effectively addressing the impact of complex environments such as smog and alternating light and dark on visibility. A dynamic correction coefficient at the entrance and a natural light transition mechanism at the exit mitigate the "black hole" and "white hole" effects experienced by drivers entering and exiting the tunnel, ensuring a smooth visual adaptation process. Furthermore, a baseline illumination threshold is set, forcing the use of a fixed baseline value when the dynamically calculated illumination falls below a safety threshold, striking a balance between energy conservation and safety. A traffic flow-sensitive formula links lighting brightness with traffic density, automatically increasing illumination in high-traffic scenarios to enhance responsiveness. Furthermore, a neural network is introduced to update dynamic parameters online, continuously optimizing the model's adaptability to environmental changes. Finally, through multi-parameter fusion and full-tunnel segmented collaborative design, this solution addresses the entire safety process, from strong light adaptation at the entrance to stable lighting in the middle section and then to the natural light transition at the exit, forming a multi-dimensional collaborative protection system. While reducing carbon emissions, this solution significantly improves tunnel driving safety through five core mechanisms: dynamic perception, precise control, safety thresholds, traffic linkage, and adaptive learning.

[0035] In order to grasp the carbon emissions of highway tunnel lighting in real time and provide a scientific basis for tunnel operation and management, the carbon efficiency analysis module collects data such as the rated power, rated illumination, unit production carbon emissions, design life, and grid carbon emission factors of lighting equipment to construct a tunnel carbon emission model that includes real-time operating energy consumption and implicit carbon emissions. At the same time, Monte Carlo simulation is used to perform multiple sampling calculations to obtain carbon emission placement confidence intervals to identify potential carbon exceeding risks in advance.

[0036] Specifically, in the carbon efficiency analysis module, building a tunnel carbon emission model may include the following steps: First, collect the key parameters of the lighting equipment, including the rated power P rated , rated illumination I rated , unit production carbon emissions M and design life L, and use the State Grid real-time data interface to obtain the grid carbon emission factor λ(t); Based on the collected data, the formula and Calculate the real-time power P(t) and equivalent operating time T(t) of lighting equipment respectively to reflect the actual energy consumption of the equipment at different times; Based on the above calculation results, the tunnel carbon emission model is constructed as follows: ; Where i represents the lighting device index, n represents the number of lighting devices, The implicit carbon emission allocation coefficient of the i-th lighting equipment is expressed by the formula Calculated, Indicates the real-time carbon emissions in the tunnel, The model comprehensively considers the real-time energy consumption and embodied carbon emissions of the equipment, and comprehensively displays the carbon emissions of the tunnel lighting system throughout its entire life cycle. In this embodiment, in order to cope with the impact of input parameter fluctuations on carbon emission calculation results, the Monte Carlo simulation method is used; First, we define the input parameters and distributions so that the vehicle flow V(t) follows a Poisson distribution, the visibility C(t) follows a normal distribution, and the grid carbon emission factor λ(t) follows a uniform distribution. Next, the number of iterations of the Monte Carlo model is set, for example, to 1000. In each iteration, simulated traffic flow values, simulated visibility values, and simulated grid carbon emission factor values ​​are obtained through sampling. These sampled simulated values ​​are substituted into the required illumination calculation formula and the tunnel carbon emission model to obtain simulated carbon emission values, which are then stored in an array to form a carbon emission simulation data set. Finally, the simulated mean AE and simulated standard deviation σ are calculated based on the carbon emission simulation data set. E , using formula E CO2 ∈[AE-1.96σE ,AE+1.96σ E ] Calculate the confidence interval E of carbon emissions CO2 , where AE+1.96σ E Place an upper limit on the confidence interval for carbon emissions; The confidence intervals for carbon emissions generated through Monte Carlo simulations effectively quantify the impact of input parameter fluctuations on carbon emissions and provide tunnel operators with a clear risk assessment range. When the upper limit of the confidence interval approaches or exceeds a preset threshold, operators can promptly identify potential carbon emissions exceeding the permitted limit and take appropriate measures, such as adjusting lighting control parameters and optimizing ventilation equipment operating modes, to avoid exceeding the permitted limit and ensure the environmental friendliness of tunnel operations.

[0037] Therefore, when the upper limit of the carbon emission confidence interval is greater than the preset threshold, the environmental optimization module will be activated, and the equipment control parameters will be optimized using the reinforcement learning framework to obtain the optimal equipment control parameters, thereby achieving effective control of the lighting and ventilation equipment in the tunnel, reducing carbon emissions and ensuring driving comfort and safety.

[0038] In one embodiment, the workflow of the environment optimization module may include the following steps: Set the state space: Set the traffic flow gradient , predicted visibility , grid carbon emission factor λ(t) and remaining carbon budget E re (t) is the state space of the reinforcement learning framework; where the remaining carbon budget E re (t) is calculated by formula E re (t) = E quota -E total (t) is calculated, E quota indicating a preset carbon emission limit; Setting the action space: dimming level of lighting equipment i , fan speed f of ventilation equipment j And the equipment maintenance instruction m, the action vector a(t)=[l i ∈{l1,l2,…,l p},f j ∈{f1,f2,…,f q},m∈{0,1}]; where p represents the number of dimming level parameters, q represents the number of fan speed parameters, and l i It can be the illuminance value or the percentage of the lighting brightness value output by the demand analysis module. jIt can be the fan speed value or the percentage of the initial fan speed value. The equipment maintenance instruction is used to send a signal whether the lighting equipment should be replaced. When m=1, it indicates sending a lighting equipment replacement signal, and when m=0, it indicates maintaining the current state of the lighting equipment. Setting up the reward function: Building the reward function , where R j Represents each reward function, w j Represents the weight coefficient determined based on historical experience; Specifically, the carbon emission reward function , used to encourage the reduction of carbon emissions; comfort reward function , aimed at ensuring driving comfort; safety penalty function ,in It is an indicative function that imposes a penalty when visibility C(t)<50 to ensure driving safety; Environment interaction and experience storage learning: Initialize the reinforcement learning framework: Use the multi-objective reinforcement learning algorithm to initialize the parameters of the reinforcement learning framework's policy network, value network, and target network, and set hyperparameters such as the learning rate, discount factor, and soft update coefficient; Create an experience replay buffer: This is used to store the experience of the agent (the system that controls the lighting and ventilation equipment in the tunnel) interacting with the environment. The agent generates action vectors based on the state vector through the decision network and observes the feedback from the environment. Generate action vector: Input the state vector s(t) into the reinforcement learning framework and generate the action vector a(t) according to the policy network.

[0039] Execute actions and get feedback: Execute action a(t) to get the next state vector s(t+1), the calculation result of the reward function r(t), and the mark d(t+1) indicating whether to terminate learning.

[0040] Store experience tuple: store the experience tuple (s(t), a(t), r(t), s(t+1), d(t+1)) into the experience replay buffer; Experience sampling and framework update: Repeat the above steps until the number of experience tuples in the experience replay buffer reaches the batch size threshold, randomly sample a batch of experience tuples from the experience replay buffer to obtain experience sampling. Use experience sampling to update the reinforcement learning framework and calculate the average reward of the last n rounds. and the average reward value of the first n rounds ; And determine whether the difference between the two is less than the preset change threshold. If so, terminate the learning and obtain the output device optimization parameters; otherwise, repeat the above process; It should be noted that the equipment optimization parameters include the dimming level of the lighting equipment, the fan speed of the ventilation equipment, and the equipment maintenance instructions.

[0041] In one embodiment, the visibility is predicted The visibility prediction model is constructed based on a machine learning algorithm, and the construction method of the visibility prediction model may include the following steps: D1, collect historical meteorological data, traffic flow and visibility, and arrange them in chronological order to obtain time series data; meteorological data includes air humidity, temperature, wind speed and air pressure, visibility C(t) is calculated by PM in the tunnel 2.5 The value PM(t) is obtained, and the calculation formula is: ; D2, build a time series model based on machine learning algorithms and define the model's loss function and training parameters; D3, preprocess the time series data and divide the preprocessed data into training sets, validation sets, and test sets according to the preset ratio; D4: Input the training set and validation set into the model for iterative training and parameter update. Use the test set to evaluate whether the accuracy of the updated model is greater than the preset accuracy threshold. If yes, save the updated model and output it as a visibility prediction model for visibility prediction. If no, adjust the model's loss function and training parameters, and repeat D4. The input of the visibility prediction model includes meteorological data and traffic flow, and the output is predicted visibility.

[0042] Within a reinforcement learning framework, the system dynamically responds to changes in the tunnel's internal and external environments by collecting real-time data such as traffic flow gradients, predicted visibility, grid carbon emission factors, and remaining carbon budgets. Furthermore, through continuous interaction with the environment and stored experience, the system autonomously learns, accumulating experience from actual operations and gradually optimizing control strategies to accurately determine the optimal timing for lighting equipment replacement, balancing equipment lifespan with energy costs. Furthermore, by leveraging a multi-objective reward function encompassing carbon emissions, comfort, and safety, the system can effectively balance multiple objectives while reducing energy consumption and carbon emissions while ensuring driving comfort and safety, thereby enhancing the overall benefits of tunnel operations.

[0043] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0044] Working principle of the present invention: The present invention collects traffic flow, illumination and PM in real time through the demand analysis module2.5 The required illumination of each segment is obtained through calculation and adjustment based on data such as values, thus realizing adaptive lighting control. The carbon efficiency analysis module is used to collect lighting equipment and power grid data, build a carbon emission model, and use Monte Carlo simulation to obtain the carbon emission placement confidence interval to quantify the uncertainty of carbon emissions. Finally, the environmental optimization module is activated when the upper limit of the carbon emission placement confidence interval exceeds the threshold. The state, action space and reward function are set through the reinforcement learning framework. The equipment control parameters are optimized through data collection and interactive learning to realize precise equipment control and maintenance decision optimization, thereby improving the green and intelligent level of tunnel operation.

[0045] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent carbon emission analysis system for highway tunnel lighting, characterized in that: include: Demand analysis module: used to collect real-time traffic flow, natural light illumination outside the tunnel, PM inside the tunnel, etc. 2.5 Value, analyze and control the lighting brightness of several lighting equipment in the tunnel; Carbon efficiency analysis module: used to build a tunnel carbon emission model, quantify the uncertainty of the tunnel carbon emission model using Monte Carlo simulation, and obtain the carbon emission confidence interval; Environmental optimization module: when the upper limit of the carbon emission placement confidence interval is greater than a preset threshold, the device control parameters are optimized using a reinforcement learning framework to obtain device optimization parameters, and the device is controlled using the device optimization parameters; wherein the device includes lighting equipment and ventilation equipment.

2. According to claim 1, the carbon emission intelligent analysis system for highway tunnel lighting is characterized in that: The analysis and control of the lighting brightness of several lighting devices in the tunnel include: A1: Divide the tunnel into several sections according to the preset distance, and collect the traffic flow V(t) and PM in the tunnel in real time. 2.5 Value PM(t), natural light illumination outside the tunnel L ext (t); where t represents the current time; A2, based on PM in the tunnel 2.5 The value PM(t) is used to calculate the visibility C(t), and the calculation formula is: ; A3, based on the traffic volume V(t), visibility C(t), and natural light intensity L outside the tunnel ext (t), calculate the basic lighting brightness I of several segments base (t), the calculation formula is: ; Among them, α, β, γ represent dynamic learning parameters, which are obtained by online updating of the neural network model, V max represents the maximum traffic capacity of the tunnel, L base (x) represents the segment reference illumination, and x represents the segment index; A4, set segment correction coefficients S(x) for several segments, calculate segment lighting brightness according to the segment correction coefficients, and control several lighting devices according to the segment lighting brightness.

3. According to claim 2, the carbon emission intelligent analysis system for highway tunnel lighting is characterized in that: The step of calculating the segmented lighting brightness according to the segmented correction coefficient comprises: A41, according to the formula I'(t)=I base (t)×S(x) is used to calculate the required illumination I'(t) of several segments; A42: Determine whether the required illumination I'(t) of several segments is greater than or equal to the segment reference illumination L base (x); if yes, the required illuminance I'(t) is marked as the segmented lighting brightness I(t); if no, the segmented reference illuminance L base (x) is marked as the segment illumination intensity I(t).

4. According to claim 2, the carbon emission intelligent analysis system for highway tunnel lighting is characterized in that: The dynamic learning parameters are obtained by online updating of the neural network model, including: A31, according to the data collection cycle, uses the sliding window method to obtain the time series data within the most recent preset time period to obtain the input features; The time series data include traffic volume V, visibility C, natural light illumination outside the tunnel L ext , Traffic flow gradient , and the traffic flow gradient is expressed by the formula get, Indicates a preset time interval; A32, performing standardization preprocessing on the input features to obtain preprocessed data; A33, constructing a recurrent neural network based on a deep learning algorithm, and setting a dynamic loss function, online incremental training parameters, and initial dynamic parameters α0, β0, and γ0 of the recurrent neural network to obtain an online update model; wherein the online incremental training parameters include an optimizer, a number of iterations, and a gradient norm threshold; A34, according to the update cycle, inputs the preprocessed data into the online update model for parameter update to obtain dynamic parameter increments , , ; A35, update formula according to parameters , and obtain dynamic learning parameters; among them, Indicates the learning rate for parameter update.

5. According to claim 3, the carbon emission intelligent analysis system for highway tunnel lighting is characterized in that: The construction process of the tunnel carbon emission model includes: B1-1, collect the rated power P of lighting equipment rated , Rated illumination I rated , unit production carbon emission M and design life L, and obtain the grid carbon emission factor λ(t) according to the State Grid real-time data interface; B1-2, calculate the real-time power P(t) of the lighting equipment based on the rated power. The calculation formula is: ; B1-3, calculate the equivalent operating time T(t) of the lighting equipment based on the rated illumination, the calculation formula is: ;in, represents the integral variable, t0 represents the starting time of carbon emission monitoring; B1-4, construct the tunnel carbon emission model as follows: ; where i represents the lighting device index, n represents the number of lighting devices, and t represents the time. The implicit carbon emission allocation coefficient of the i-th lighting device is expressed by the formula Calculated, Indicates the real-time carbon emissions in the tunnel. Indicates installation time.

6. The carbon emission intelligent analysis system for highway tunnel lighting according to claim 1 is characterized in that: The use of Monte Carlo simulation to quantify uncertainty in the tunnel carbon emission model includes: B2-1, define the input parameters and parameter distribution as follows: vehicle flow V(t) follows Poisson distribution, visibility C(t) follows normal distribution, and grid carbon emission factor λ(t) follows uniform distribution; B2-2, setting the number of iterations of the Monte Carlo model, and in each iteration, obtaining the simulated traffic flow value, the simulated visibility value, and the simulated grid carbon emission factor value by sampling to obtain the sampled simulation value; B2-3, substituting the sampled simulation value of each iteration into the tunnel carbon emission model to obtain the simulated carbon emission value, and storing it in an array to obtain a carbon emission simulation data group; B2-4, obtain the simulated mean AE and simulated standard deviation σ based on the carbon emission simulation data set E , calculate E using the formula CO2 ∈[AE-1.96σ E ,AE+1.96σ E ]Get the carbon emission placement confidence interval E CO2 ; Among them, AE+1.96σ E Place an upper bound on the confidence interval for carbon emissions.

7. The carbon emission intelligent analysis system for highway tunnel lighting according to claim 3 is characterized in that: The optimization of the equipment control parameters by using the reinforcement learning framework includes: C1, set the state space of the reinforcement learning framework to: traffic flow gradient , Forecast visibility , grid carbon emission factor λ(t) and remaining carbon budget E re (t); where the remaining carbon budget E re The calculation formula of (t) is: re (t) = E quota -E total (t), E quota represents the preset carbon emission limit, t represents the time, Indicates a preset time interval; C2, set the action space of the reinforcement learning framework to: the dimming level l of the lighting equipment i , fan speed f of ventilation equipment j And the equipment maintenance instruction m, we get the action vector a(t)=[l i ∈{l1,l2,…,l p },f j ∈{f1,f2,…,f q },m∈{0,1}]; where p represents the number of dimming level parameters, q represents the number of fan speed parameters, and the equipment maintenance instruction is used to send a signal whether the lighting equipment is replaced, and when m=1, it means sending a lighting equipment replacement signal, and when m=0, it means keeping the current state of the lighting equipment; C3, set the reward function of the reinforcement learning framework to: ; Among them, R1 represents the carbon emission reward function, R2 represents the comfort reward function, R3 represents the safety penalty function, and w j Represents the weight coefficient, which is determined based on historical experience; C4, collects real-time state space data, obtains the state vector, inputs the state vector into the reinforcement learning framework for environmental interaction and experience storage learning, and obtains the equipment optimization parameters; The equipment optimization parameters include dimming levels of lighting equipment, fan speeds of ventilation equipment, and equipment maintenance instructions.

8. The carbon emission intelligent analysis system for highway tunnel lighting according to claim 7 is characterized in that: The step of inputting the state vector into the reinforcement learning framework for environment interaction and experience storage learning includes: C41, using the multi-objective reinforcement learning algorithm to initialize the reinforcement learning framework, including initializing the parameters of the policy network, value network, and target network of the reinforcement learning framework, and setting the learning rate, discount factor, and soft update coefficient; C42, create experience replay buffer; C43, inputs the state vector s(t) into the reinforcement learning framework and generates the action vector a(t) according to the policy network; C44, executes action a(t), obtains the next state vector s(t+1), the calculation result of the reward function r(t), and the mark d(t+1) indicating whether to terminate learning; C45, store the experience tuple (s(t), a(t), r(t), s(t+1), d(t+1)) into the experience replay buffer; C46, repeat C43 to C45 until the number of experience tuples in the experience replay buffer reaches the batch size threshold, randomly sample a batch of experience tuples from the experience replay buffer to obtain experience sampling; C47, using experience sampling to update the reinforcement learning framework and calculate the average reward value of the last n rounds and the average reward value of the first n rounds ,judge and Is the difference less than the preset change threshold? If yes, the learning is terminated and the output result is obtained; if no, C43 to C46 are repeated.

9. The carbon emission intelligent analysis system for highway tunnel lighting according to claim 7 is characterized in that: The predicted visibility is obtained by predicting a visibility prediction model constructed based on a machine learning algorithm; The visibility prediction model is constructed as follows: D1, collect historical meteorological data, traffic flow and visibility, and arrange them in chronological order to obtain time series data; wherein the meteorological data includes air humidity, temperature, wind speed and air pressure, and the visibility C(t) is measured by PM in the tunnel. 2.5 The value PM(t) is obtained by the following calculation formula: ; D2, build a time series model based on machine learning algorithm and define the model's loss function and training parameters; D3, preprocessing the time series data, and dividing the preprocessed data into data sets according to a preset ratio to obtain a training set, a validation set, and a test set; D4, input the training set and the validation set into the model for iterative training and parameter update, and use the test set to evaluate whether the accuracy of the model after parameter update is greater than the preset accuracy threshold; if yes, save the model after parameter update and output it as a visibility prediction model for predicting visibility; if no, adjust the loss function and training parameters of the model and repeat D4; The input of the visibility prediction model includes meteorological data and vehicle flow, and the output is predicted visibility.

10. The carbon emission intelligent analysis system for highway tunnel lighting according to claim 7 is characterized in that: The reward function of the reinforcement learning framework includes: The carbon emission reward function R1 is: ; The comfort reward function R2 is: ; The security penalty function is R3: ;in, Represents an indicative function.

Images