A computing method for realizing fast updating of a large-scale dynamic noise map system

Abstract

The present application relates to a kind of computing method for realizing wide range dynamic noise map system fast update, comprising the following steps: using the data information of system database, based on existing day and night period and city road grade, simplify the structure of traditional road sound level mechanism model;Using big data technology, using hybrid modeling method, regression each grade road sound source (day and night) sound level contribution amount prediction model;Road noise fast calculation method based on sound level contribution amount prediction model is constructed, the fast update of dynamic noise map is realized.Compared with prior art, while ensuring the calculation precision, the operation and update efficiency of noise map system is greatly improved, and the technical constraints of foreign core computing software are broken, the localization of noise map core computing technology is realized, which provides technical support for wide range deep development of noise map.

Description

Technical Field

[0001] This invention relates to the field of urban environmental noise control and management technology, and in particular to a calculation method for rapidly updating a large-scale dynamic noise map system. Background Technology

[0002] With rapid urbanization, environmental noise pollution has become an increasingly prominent public concern. Noise maps display the distribution of noise pollution within urban areas using both numerical and graphical methods. In recent years, with the continuous expansion of urban areas and rising population density in my country, traditional noise maps have become insufficient for real-time monitoring of urban environmental noise. Therefore, the Shanghai Academy of Environmental Sciences pioneered a dynamic noise map management system. This system dynamically updates the noise map by calculating noise hours at grid points, supplemented by on-site monitoring and a road noise correction system to correct sound levels. In noise maps covering small to medium areas (below 20 square kilometers), it has effectively achieved hourly noise level updates and corresponding management functions.

[0003] However, at present, the core calculations of dynamic noise maps largely rely on mature foreign acoustic software. Domestic research on core calculation technologies for noise maps is still in its early stages, with no independent intellectual property rights yet. The closed nature of foreign core software also restricts the in-depth development and large-scale application of noise map systems. In fact, mature acoustic calculation software cannot currently achieve large-scale, rapid updates to dynamic noise map management systems. This is mainly because noise maps have certain accuracy requirements for calculation grid points (generally not less than 10m × 10m). A 20 square kilometer calculation grid contains 200,000 grid points. Currently, using a high-efficiency server, the calculation time for one operation is approximately 45 minutes. Adding the approximately 20 minutes for data and image processing, this exceeds the minimum requirement for hourly map updates. Therefore, the massive computational load of grid points restricts the dynamic updates of large-scale noise maps, necessitating a computational method that enables rapid updates of large-scale noise maps. Summary of the Invention

[0004] The purpose of this invention is to optimize the update efficiency of existing dynamic noise map systems and to achieve the localization of core noise map technologies. Based on the existing database of noise map systems, this invention provides a calculation method for rapidly updating large-scale dynamic noise map systems using a big data-driven approach.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A computational method for rapidly updating a large-scale dynamic noise map system includes the following steps:

[0007] S1: Collect and analyze traffic data and noise contribution data of road noise sources in the system database, and construct a noise map road noise source sound level contribution prediction model according to day and night time and different urban road grades.

[0008] S2: The parameters of the prediction model for the contribution of road noise sources (day and night) at each level are regressed using a multivariate nonlinear regression modeling method.

[0009] S3: Based on a data-driven approach, compare the calculated values ​​of the prediction models for the sound level contribution of road noise sources of different levels with the calculated values ​​of the original core software, and calculate the sound level contribution correction of all road noise sources in the noise map.

[0010] S4: Introduce the road noise source sound level contribution prediction model and the sound level contribution correction amount into the noise map system, and calculate the sound level contribution of each road noise source and the road noise at the prediction point in the noise map system in real time.

[0011] Furthermore, the system database is a database of a noise map system, including a road traffic flow and vehicle speed database, a core software calculation value database of road noise source contribution, and a prediction point latitude and longitude database.

[0012] Further, step S1 includes the following sub-steps:

[0013] S11: Collect traffic data, noise contribution data, and prediction point latitude and longitude data of all road noise sources in the system database to construct a traffic sample set and a noise contribution sample set for road noise sources.

[0014] S12: Analyze the distribution characteristics of the traffic data and noise contribution data, and classify road noise source samples by combining the day and night time periods of the noise map system and the urban road classification.

[0015] S13: Combining the emission principle of road traffic noise sources and the distribution characteristics of traffic data from the noise map system, construct prediction models for the sound power levels of road noise sources (day and night) at different urban road levels and during different time periods;

[0016] S14: Based on the constructed road noise source power level prediction model, combined with my country's road traffic noise assessment methods and field experience, a noise map road noise source power level prediction model is constructed.

[0017] Furthermore, in step S11, the road sound sources of this local map system include R1,…,R k (A total of k road sound sources), prediction points include P1,…,P l (A total of 1 point), the collected traffic data includes the hourly average flow of large vehicles in group M. Average hourly traffic flow of cars Average speed of large vehicles per hour and the average speed of cars per hour The road noise source contribution data represents the core software calculation value of the road's hourly noise contribution to the M groups of prediction points.

[0018] Furthermore, the different urban road levels in step S12 include q urban road levels, and the day and night time periods include daytime 06:00-22:00 and nighttime 22:00-06:00.

[0019] Furthermore, in step S13, the noise sources R at each level of the road i The sound power level prediction model is obtained by superimposing the line source sound power levels of the large traffic flow and the small traffic flow, and its expression is:

[0020]

[0021]

[0022] Where i (=1,…,k) represents the road noise source index, m (=1,…,q) represents the urban road grade, d represents the daytime period 06:00-22:00, and n represents the nighttime period 22:00-06:00. R represents the m-th level road noise source. i Daytime line source sound power level, This represents the line source sound power level during periods of heavy daytime traffic. The line source sound power level represents the daytime traffic flow. R represents the m-th level road noise source. i Nighttime sound power level, This indicates the sound power level of heavy traffic at night. This indicates the sound power level of light traffic flow at night; These represent the sound power levels of a single large vehicle and a single small vehicle at the m-th level road noise source during the daytime. Let represent the sound power level of a single large vehicle and a single small vehicle at night, respectively, representing the sound power level of a single road noise source of level m at night. Here, q represents the number of road grades in the city. This represents the average hourly flow rate of large vehicles. This represents the average hourly traffic flow for vehicles. The average hourly speed of large vehicles, and This represents the car's average speed per hour.

[0023] Further, in step S14, the prediction model for the sound level contribution of the road noise source is the sound source R of the i-th road. i For the j-th prediction point P j The prediction model for the hourly equivalent A-weighted sound level is expressed as follows:

[0024]

[0025] Where j (=1,…,l) represents the prediction point index, R represents the m-th level road noise source. i For prediction point P j The contribution of daytime sound level R represents the m-th level road noise source. i For prediction point P j The contribution of nighttime sound level;

[0026] Sound level contribution correction ΔL ij R represents the sound source of the i-th road. i To the j-th prediction point P j The sound level attenuation (i.e., the attenuation during sound wave propagation), the sound level contribution correction includes distance attenuation. Building shading attenuation and the amount of attenuation caused by other factors Its expression is:

[0027]

[0028] Further, step S2 includes the following sub-steps:

[0029] S21: Analyze the distribution characteristics of traffic data for road noise sources at each level, and select representative road noise sources R from q levels of roads. 1 ,…,R q ;

[0030] S22: Search for each sound source road R 1 ,…,R q Corresponding prediction point P 1 ,…,P q Collect the corresponding traffic sample set and noise contribution sample set;

[0031] S23: Using a multivariate nonlinear fitting method based on the MATLAB platform, the parameters of the prediction model for the contribution of q levels of road noise sources (day and night) constructed in step S1 are regressed.

[0032] Furthermore, in step S23, the parameters of the road noise source sound level contribution prediction model include: the daytime sound power level of a single large vehicle on the m-th level road. Daytime sound power level of a single car on a Class m road Nighttime sound power level of a single large vehicle on a Class m road The sound power level of a single car on a Class m road at night. and the prediction point P of the q-class road sound source. 1 ,…,P q The sound level contribution correction ΔL1 ,…,ΔL q .

[0033] Furthermore, step S3 includes the following sub-steps:

[0034] S31: Randomly select a subset M′ from the M sample set, and using the collected latitude and longitude data, statistically analyze the correspondence between l prediction points and k road noise sources, and establish a one-to-one identification "i-j" between road noise sources and prediction points in the noise map system. i ”;

[0035] S32: Based on the one-to-one identifier "i-j i “The sound level contribution of the q-level road noise source (day and night) obtained by regression is used to calculate the sound level contribution of all road noise sources in the M′ group of traffic samples.

[0036] S33: Compare the model-calculated values ​​of the sound level contribution of the M′ group of samples with the core software-calculated values, and evaluate the difference between the model-calculated values ​​and the software-calculated values ​​of the sound level contribution of the k road sound sources. That is, the road sound source R in the map. i To its jth i Prediction points The sound level contribution correction amount.

[0037] Further, step S4 includes the following sub-steps:

[0038] S41: Integrate the q-level road noise source (day and night) sound level contribution prediction model into the noise map system backend, and assign a one-to-one identifier i-j between the road noise source and the prediction point. i And sound level contribution correction Import system database;

[0039] S42: Automatically run the road noise source sound level contribution prediction model and sound level contribution correction calculation every hour to obtain the hourly calculated value of the sound level contribution of each road noise source in the noise map.

[0040] S43: Based on the real-time calculated value of the sound level contribution of each road sound source, the road noise of all predicted points in the noise map system is calculated and updated in real time using the road traffic sound level superposition calculation formula.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1) Optimizing the core calculation method for noise maps and achieving domestic production of core noise map technology: This invention is the first to systematically study the core calculation method in urban noise map management systems. It breaks away from the traditional approach of relying on foreign acoustic software for grid point calculation in map updates, and adopts a big data-driven method to construct a core calculation method for rapid updates of large-scale dynamic noise maps. While maintaining prediction accuracy, this invention significantly improves the computational speed of the map system (for an area of ​​approximately 20 square kilometers, a single calculation takes less than 1 minute, and with pre- and post-processing of data and images, the computation time is significantly reduced). Figure 1 This invention (with updates occurring within 3 minutes) overcomes the limitation of previous noise map updates, which were hampered by excessively long core software computation times, thus preventing large-scale hourly updates. This provides strong technical support for future expansion of the noise map's geographical scope. Furthermore, this invention achieves the localization of core computational technology for the noise map system, avoiding reliance on foreign black-box core computational software and breaking through the limitations of noise map system research and application. This invention further optimizes existing noise map management systems, enabling more timely and accurate reflection of urban noise information, scientific evaluation of the current state of noise pollution and its environmental impact, and providing precise data for planning decisions, noise source management, and noise pollution prevention and control.

[0043] 2) A Real-Time Fast Calculation Method for Road Noise Based on Big Data: This invention fully utilizes the massive database of a noise map system and employs a hybrid modeling method combining big data-driven approaches and mechanistic modeling. It constructs prediction models for the sound level contribution of road noise sources (day and night) at different time periods and for different urban road grades. This replaces the identification of sound power mechanism models for individual vehicles with complex mechanistic structures. The single-vehicle sound power mechanism model is parameterized, avoiding the large number of parameters and empirical coefficients required to evaluate traditional (complex and nonlinear) road sound level mechanism models, while significantly reducing the computational complexity of the model. The constructed model is applied to the real-time calculation of road noise and its source contribution in the noise map system, ultimately achieving hourly rapid updates of the noise map.

[0044] 3) Road Noise Source Sound Level Contribution Prediction Modeling Based on Multivariate Nonlinear Regression: This invention utilizes a multivariate nonlinear regression modeling method to regress the road noise source sound level contribution prediction model, greatly simplifying the model parameters and avoiding the selection of empirical coefficients. By keeping the core software mechanism model a black box, the invention ensures the consistency of the constructed sound level contribution prediction model with the parameters of the original core calculation software, avoiding errors caused by the selection of empirical coefficients in traditional mechanism-based sound level prediction models, and ensuring the computational accuracy of the proposed road noise source sound level contribution prediction model. Attached Figure Description

[0045] Figure 1 This is a flowchart of the present invention;

[0046] Figure 2 A flowchart for predicting the contribution of road sound sources to sound levels in the noise map system of this invention;

[0047] Figure 3 This is a schematic diagram illustrating the hourly rapid update operation principle of the dynamic noise map system of the present invention.

[0048] Figure 4 This is a diagram illustrating the effect of road noise calculation and updating in the noise map system of this invention. Detailed Implementation

[0049] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0050] like Figure 1 and Figure 2 As shown, this invention provides a computational method for rapidly updating a large-scale dynamic noise map system, aiming to optimize the computational efficiency of existing dynamic noise map systems and achieve the localization of core noise map technologies. Based on big data-driven modeling, this method combines traditional road sound level prediction mechanism models with multivariate nonlinear regression modeling to construct a real-time road noise calculation method based on a road sound source sound level contribution prediction model. This significantly simplifies the complex model structure and parameters of traditional mechanism models. The road sound source sound level contribution prediction model is integrated into the system backend, replacing the black-box foreign core computational software, to calculate the road sound source sound level contribution of the noise map system in real time. Furthermore, it calculates and updates the road noise at all prediction points.

[0051] like Figure 1 As shown, the real-time fast calculation method for the dynamic noise map system proposed in this invention includes the following steps:

[0052] Step 1: Collect and analyze a large amount of historical traffic data and noise contribution data in the system database to construct a prediction model for the sound level contribution of road noise sources;

[0053] Specifically, it includes:

[0054] Step 11: Collect historical traffic data (including high traffic volume) from group M in the noise map system database. Traffic flow Large vehicle speed and the speed of the car ), road noise source contribution data (core software calculated value) and prediction point latitude and longitude data;

[0055] Step 12: Analyze the distribution trend of historical traffic data and noise contribution of group M, and combine day and night time periods and the classification of urban road level q to construct a road source sound power level prediction model for noise map;

[0056] Step 13: Based on the road noise source power level prediction model and combined with my country's road traffic noise assessment method, construct a noise map road noise source power level prediction model.

[0057] Step 2: Based on big data-driven technology, the parameters of the prediction model for the sound level contribution of road noise sources (day and night) at each level are regressed using a multivariate nonlinear regression modeling method.

[0058] Specifically, it includes:

[0059] Step 21: Analyze the distribution characteristics of road traffic data at each level, and select representative road noise sources R for q levels of road types. 1 ,…,R q ;

[0060] Step 22: Using traffic data and noise contribution data of these q representative road noise sources, adopt multivariate nonlinear regression modeling based on the MATLAB platform to regress the parameters of the prediction model for the sound level contribution of the q-level road noise sources.

[0061] Step 23: Collect the latitude and longitude of the prediction points and the contributing road noise source information, and establish a one-to-one identification system for "road noise source - prediction point" (i-j). i ”;

[0062] Step 24: Based on the one-to-one identifier "i-j i Based on the q-level road noise source (day and night) noise level contribution prediction model, a big data-driven approach is used to evaluate the noise level contribution correction of all "road noise source-prediction points" in the system.

[0063] Step 3: Introduce the q-level road noise source (day and night) sound level contribution prediction model, the one-to-one identification of "road noise source - prediction point" and its corresponding sound level contribution correction into the background of the existing noise map system to realize the real-time and rapid calculation of the sound level contribution of road noise sources and the road noise at the prediction point in the dynamic noise map system.

[0064] Specifically, it includes:

[0065] Step 31: Connect the q-level road noise source (day and night) sound level contribution prediction model to the existing noise map system backend, and assign a one-to-one identifier "i-j" to each "road noise source - prediction point". i Import the corresponding sound level contribution correction value into the system database;

[0066] Step 32: Automatically trigger the prediction model and correction calculation of road noise source sound level contribution every hour, and calculate the sound level contribution of all road noise sources in real time.

[0067] Step 33: Based on the calculated sound level contribution of road sound sources, run the road traffic sound level superposition calculation to calculate the road noise at all predicted points of the system in real time.

[0068] The specific implementation steps are as follows:

[0069] First, a large amount of historical data, such as traffic flow, vehicle speed, sound level contribution, and latitude and longitude of prediction points, are collected from the noise map system database.

[0070] Then, the data distribution characteristics were analyzed, and the principle of traffic noise emission was combined with the simplification of the road noise source power prediction model structure according to day and night time periods and urban road levels. Furthermore, the q-level road noise source (day and night) sound level contribution prediction model was constructed based on the principle of road traffic noise assessment.

[0071] Subsequently, based on the database technology and big data-driven methods of the noise map system, a hybrid modeling method combining multiple nonlinear regression and traditional mechanistic modeling was employed to regress the parameters of the proposed road noise source sound level contribution prediction model. Specifically, a multivariate nonlinear regression modeling method based on the MATLAB platform was used to regress the sound level contribution prediction model for representative road noise sources at each level; then, a big data-driven calculation method was used to evaluate the sound level contribution correction of all "road noise source-prediction point" points in the system; finally, a q-level road noise source (day and night) sound level contribution prediction model and a database of arbitrary road noise source sound level contribution corrections were obtained.

[0072] Finally, the constructed q-level road noise source (day and night) sound level contribution prediction model is integrated into the existing noise map system backend, and the "road noise source - prediction point" is identified one-to-one as "i-j". i The sound level contribution correction is imported into the system database and triggered every hour to calculate the sound level contribution of each road sound source in real time. Then, the calculated sound level contribution values ​​of each road sound source are superimposed to calculate the final real-time and efficient calculation and update of road traffic noise in the dynamic noise map system.

[0073] like Figure 2 The diagram shows the construction flowchart of the road noise source sound level contribution prediction model of the noise map system of the present invention. The specific modeling steps are as follows:

[0074] 1. Collect road noise source traffic data, noise contribution data, and prediction point latitude and longitude data from the database to construct road noise source traffic sample sets and noise contribution sample sets:

[0075] 1) Collect all prediction points P1,…,P in the database. l (A total of l points) Latitude and longitude data, and M noise contribution prediction data (core calculation software), to evaluate all road noise sources R1,…,R in the system. k ;

[0076] 2) Collect road noise sources R1,…,R from the database. k From the historical traffic data of group M, we can obtain the large traffic volume of group M. Traffic flow Large vehicle speed and the speed of the car Constructing a traffic sample set in Indicates the road noise source R i The t-th set of historical traffic data;

[0077] 3) Statistical analysis of road noise sources R1,…,R k and their corresponding prediction points P1,…,P l Based on the noise contribution data, construct a sample set of road noise source contribution data.

[0078] 2. Analyze the distribution characteristics of M sets of traffic data for road noise sources. Combining the emission principles of road traffic noise sources and the Chinese Guidelines for Road Traffic Noise Assessment (HJ2.4-2021), construct a prediction model for the sound level contribution of road noise sources in the noise map. The specific modeling steps are as follows:

[0079] 1) Evaluate the traffic sample set of group M Based on the data distribution trend and the distribution characteristics of traffic flow and speed of each road noise source in group M, and according to the urban road level and day / night time, a prediction model for the sound power level of a q-level road noise source (day and night) is constructed as follows, parameterizing the single-vehicle sound power mechanism model with a complex mechanism structure:

[0080]

[0081]

[0082] Among them, R i Let m represent the sound source of the i-th road, m (=1,…,q) represent the urban road class, d represent the daytime period 06:00-22:00, and n represent the nighttime period 22:00-06:00. R represents the m-th level road noise source. i Daytime line source sound power level, which consists of the line source sound power levels of large and small traffic flows during the daytime. Similarly, the road noise source R is obtained by superposition. i Nighttime line source sound power level The sound power levels of heavy traffic and light traffic at night Obtained by superposition. and The average hourly flow rate for both large and small vehicles. and The average hourly speeds of large vehicles and small vehicles are given. Let be the parameters to be estimated, where These represent the sound power levels of a single large vehicle and a single small vehicle at the m-th level road noise source during the daytime. This represents the sound power level of a single large vehicle and a single small vehicle on a road of class m at night.

[0083] 2) Based on the constructed q-level road noise source (day and night) sound power level prediction model (Equation (1)-Equation (2)), combined with the road traffic noise evaluation principle (HJ2.4-2021), the noise map road noise source sound level contribution prediction model is constructed as follows:

[0084]

[0085] Among them, P j This represents the j-th prediction point. R represents the m-th level road noise source. i For prediction point P j The contribution of daytime sound level R represents the m-th level road noise source. i For prediction point P j Nighttime sound level contribution, sound level contribution correction ΔL ij Let R be the parameter to be estimated, representing the road noise source R. i To the predicted point P j The sound level attenuation (i.e., the attenuation during sound wave propagation), including distance attenuation. and building shading attenuation and the amount of attenuation caused by other factors Its expression is as follows:

[0086]

[0087] This map system contains l prediction points, corresponding to k road sound sources, i.e., i = 1, ..., k, j = 1, ..., l.

[0088] 3. Based on the collected traffic sample set Using a big data-driven approach, the regression model for predicting the sound level contribution of q-level road noise sources (day and night) constructed in this invention is as follows:

[0089] 1) Analyze the distribution characteristics of traffic flow and speed of M groups of noise sources at each level of road, and select representative noise source roads R at each level. 1 ,…,R q and prediction point P 1 ,…,P q Collect corresponding traffic samples Samples of noise contribution

[0090] 2) Using a multivariate nonlinear model fitting method based on the MATLAB platform, the parameters of the prediction model for the sound level contribution of q road noise sources were regressed, including the sound power levels of single large and small vehicles at daytime and nighttime for q types of road noise sources. and the sound source road R 1 ,…,R q The corresponding sound level contribution correction ΔL 1 ,…,ΔL q ;

[0091] 3) Using the collected latitude and longitude information of the prediction points and road noise source information, statistically analyze the correspondence between l prediction points and k noise source roads, and establish a one-to-one identification "i-j" between the road noise source and the prediction point in the noise map system. i ”;

[0092] 4) Randomly collect a subset of group M′ from the traffic sample set of group M. Using the regression-derived q-level road noise source (day and night) sound level contribution prediction model, calculate the sound level contribution of group M′.

[0093]

[0094] 5) Evaluate the error between the model calculation value and the core software calculation value of the sound level contribution of group M′;

[0095] The calculated values ​​from the model are:

[0096]

[0097] The core software calculation value is:

[0098]

[0099] We obtain k road sound sources R1,…,R k To l prediction points P1,…,P l sound level contribution correction amount in R represents the sound source of the i-th road. i up to its corresponding j-th i Prediction points The sound level contribution correction is calculated using the following formula:

[0100]

[0101] like Figure 3 As shown, the hourly rapid update process of the dynamic noise map system is as follows:

[0102] The system updates road traffic flow and vehicle speed hourly, automatically triggering the background road noise source contribution prediction model and calling the one-to-one identifier "i-j" of the "road noise source - prediction point" in the database. i "and the corresponding sound level contribution correction amount" The system calculates and stores the sound level contribution of all road noise sources in real time for the current hour, triggers the superposition calculation of the sound level contribution of road noise sources, calculates and stores the road noise of all predicted points for the current hour, and then updates the traffic noise level of grid points and the contribution of each road noise source in real time on the front-end noise map webpage. The system then ends its operation for the current hour and waits for the trigger of the next hour.

[0103] Example 1:

[0104] The q-level road noise source (day and night) sound level prediction model constructed in this invention is integrated into the backend of the noise map system, and the "road noise source - prediction point" is identified one-to-one as "i-j". i "and sound level contribution correction amount ΔL" iji The system database was imported, and traffic data from hour T1 was randomly accessed for testing. This triggered the execution of a background sound contribution prediction model and real-time correction calculations for sound level contributions. The road sound source contribution for that hour was calculated in real time, and the contribution data was stored in the database. Then, the road sound source contribution superposition calculation was performed to obtain the road noise at all predicted points in the map system for hour T1. Further, following the noise prediction method for hour T1, traffic data from times T2, ..., T10 were randomly selected and accessed into the system to complete real-time road noise calculations, resulting in a total of 10 road noise calculations. According to the Technical Guidelines for Environmental Impact Assessment in my country (HJ2.4-2021), the model calculation error must be within 3 dB(A). The sound level calculation results of the model and the original core software at times T1, ..., T10 were compared. Figure 4 As shown, for the existing noisy map range, the proportion of predicted points with a difference of less than 3dB(A) reached over 97.9%, and the average effective point ratio of the 10 model calculations was 97.93%, with a variance of 2.84*10. -8 As can be seen, the real-time fast road noise calculation method of the dynamic noise map system constructed in this invention has superior calculation accuracy and robustness, and the calculation time can be reduced to about 2% of that of grid point calculation.

[0105] Example 2:

[0106] Select any prediction point P on the noise map, at any time T0, collect the calculated noise value of point P from the database and the sound level contribution of all road noise sources R1, R2, ..., R9, as shown in Table 1. The model prediction error of the contribution of all road noise sources at this prediction point is within 1 dB. Among them, the errors between the model calculation value and the software calculation value of the contribution of the main road noise sources R1, R5, R8, and R9 and the sound level of this prediction point are all within 0.5 dB(A) (the difference in sound level contribution between other road noise sources and the largest road noise source R5 is greater than 10 dB, and its impact on the sound level of this prediction point can be ignored). It can be seen that the real-time fast calculation method of noise map based on the road noise source sound level contribution prediction model constructed in this invention has superior calculation accuracy.

[0107] Table 1. Calculated noise values ​​at prediction point P and road noise contribution at time T0.

[0108]

[0109] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A computational method for rapidly updating a large-scale dynamic noise map system, characterized in that, Includes the following steps: S1: Collect and analyze traffic data and noise contribution data of road noise sources in the system database, and construct a noise map road noise source sound level contribution prediction model according to day and night time and different urban road grades. S2: The parameters of the prediction model for the sound level contribution of road noise sources at each level are regressed using a multivariate nonlinear regression modeling method. S3: Based on a data-driven approach, compare the calculated values ​​of the prediction models for the sound level contribution of road noise sources of different levels with the calculated values ​​of the original core software, and calculate the sound level contribution correction of all road noise sources in the noise map. S4: Introduce the road noise source sound level contribution prediction model and the sound level contribution correction amount into the noise map system, and calculate the sound level contribution of each road noise source and the road noise at the prediction point in the noise map system in real time. Step S1 includes the following sub-steps: S11: Collect traffic data, noise contribution data, and prediction point latitude and longitude data of all road noise sources in the system database to construct a traffic sample set and a noise contribution sample set for road noise sources. S12: Analyze the distribution characteristics of the traffic data and noise contribution data, and classify road noise source samples by combining the day and night time periods of the noise map system and the urban road classification. S13: Combining the emission principles of road traffic noise sources and the distribution characteristics of traffic data from noise map systems, construct prediction models for the sound power levels of road noise sources at different levels and during different times of day and night in different cities. S14: Based on the constructed road noise source power level prediction model, construct a noise map road noise source power level prediction model; In step S13, the noise sources at all levels of the road R i The sound power level prediction model is obtained by superimposing the line source sound power levels of the large traffic flow and the small traffic flow, and its expression is: in, Indicates road noise source indicators. Indicates the city road classification. Indicates the daytime period. Indicates nighttime period, Indicates the first m Class I road noise source Daytime line source sound power level, This represents the line source sound power level during periods of heavy daytime traffic. The line source sound power level represents the daytime traffic flow. Indicates the m-th level road noise source Nighttime line source sound power level, This indicates the line source sound power level for heavy traffic at night. This represents the line source sound power level of light traffic flow at night; , They represent the daytime number m The sound power level of a single large vehicle and a single small vehicle in a Class II road noise source. , They represent the nighttime number of... m Line source sound power level of a single large vehicle and a single small vehicle in a road noise source class. This represents the average hourly flow rate of large vehicles. This represents the average hourly traffic flow for vehicles. The average hourly speed of large vehicles, and The average hourly speed of the car; In step S14, the prediction model for the sound level contribution of road noise sources is the first... i Road noise source R i For the j Prediction points P j The prediction model for the hourly equivalent A-weighted sound level is expressed as follows: in, Indicates the predictive point index. Indicates the first m Class I road noise source For prediction points The contribution of daytime sound level Indicates the first m Class I road noise source For prediction points The contribution of nighttime sound level; Sound level contribution correction Indicates the first i Road noise source To the j Prediction points The sound level attenuation, the sound level contribution correction includes the distance attenuation. Building shading attenuation and the amount of attenuation caused by other factors Its expression is: 。 2. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 1, characterized in that, The system database is a database for a noise map system, including a database of road traffic flow and vehicle speed, a database of core software calculation values ​​of road noise source contribution, and a database of latitude and longitude of prediction points.

3. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 1, characterized in that, In step S11, the road noise source contribution data represents the contribution of the current road to the predicted point. M The core software calculation value of the group's hourly noise contribution.

4. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 1, characterized in that, Step S2 includes the following sub-steps: S21: Analyze the distribution characteristics of traffic data for road noise sources at various levels, and select... q Road noise sources in each level of road R 1 ,…, R q ; S22: Search for roads containing various sound sources R 1 , …, R q Corresponding prediction point P 1 ,…,P q Collect the corresponding traffic sample set and noise contribution sample set; S23: Using a multivariate nonlinear fitting method based on the MATLAB platform, the regression model constructed in step S1 is regressed. q Parameters of the prediction model for the sound level contribution of road noise sources at each level.

5. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 4, characterized in that, In step S23, the parameters of the road noise source sound level contribution prediction model include: the first... m Level 1 Road Daytime Single Vehicle Large Vehicle Line Sound Source Power , No. m Level 1 Road Daytime Single Vehicle Line Sound Source Power , No. m Level 1 Road Nighttime Single Vehicle Large Vehicle Line Sound Source Power Level , No. m Level 1 Road Nighttime Single Vehicle Line Sound Source Power Level and q Road-like sound sources and their prediction points P 1 ,…,P q sound level contribution correction amount .

6. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 1, characterized in that, Step S3 includes the following sub-steps: S31: In Randomly selected from the group sample set Group subsets, using collected latitude and longitude data, to statistically analyze l Each prediction point and k The correspondence between each road noise source is established, and a one-to-one identification between road noise sources and prediction points is created in the noise mapping system. i — j i ; S32: Based on one-to-one identification i — j i The results obtained by applying regression q A model for predicting the sound level contribution of all road noise sources is used to calculate the sound level contribution of all road noise sources. The sound level contribution of the traffic sample group; S33: Comparison The model-calculated values ​​of the sound level contribution of the sample group are compared with the calculated values ​​of the original core software. k The difference between the model-calculated value and the software-calculated value of the sound level contribution of each road noise source. That is, the road sound source in the map. R i To its first j i Prediction points The sound level contribution correction amount.

7. The calculation method for rapidly updating a large-scale dynamic noise map system according to claim 1, characterized in that, Step S4 includes the following sub-steps: S41: Will q The noise level contribution prediction model for road noise sources is integrated into the noise map system backend, establishing a one-to-one correspondence between the road noise sources and the prediction points. i — j i And sound level contribution correction Import system database; S42: Automatically run the road noise source sound level contribution prediction model and sound level contribution correction calculation every hour to obtain the hourly calculated value of the sound level contribution of each road noise source in the noise map. S43: Based on the real-time calculated value of the sound level contribution of each road sound source, the road noise of all predicted points in the noise map system is calculated and updated in real time using the road traffic sound level superposition calculation formula.

Images