阅读说明：本技术 基于微生物的列车车厢空气调控方法、系统及存储介质 (Method and system for regulating and controlling air in train compartment based on microorganisms and storage medium ) 是由 刘辉 李燕飞 夏雨 陈超 于 2020-12-30 设计创作，主要内容包括：本发明公开了一种基于微生物的列车车厢空气通风调节方法、系统及存储介质,根据各测点间微生物扩散情况调节通风系统,进而降低乘客所在区域的微生物污染指标。该方法对铁路列车空气质量调控有着指导作用。本发明研究了微生物污染与大气污染物浓度的映射关系,能有效解决微生物检测的实时性问题,确保对列车车厢内微生物污染的实时调控。(The invention discloses a method and a system for regulating air ventilation of a train compartment based on microorganisms and a storage medium, wherein a ventilation system is regulated according to the microorganism diffusion condition among measuring points, so that the microorganism pollution index of the area where passengers are located is reduced. The method has a guiding function on the air quality control of the railway train. The invention researches the mapping relation between the microbial pollution and the concentration of the atmospheric pollutants, can effectively solve the problem of real-time property of microbial detection, and ensures real-time regulation and control of the microbial pollution in the train carriage.)

1. A method for regulating and controlling air in a train compartment based on microorganisms is characterized by comprising the following steps:

1) detecting PM at train supply air outlet, exhaust outlet and seat_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration and total number of bacterial colonies;

2) according to PM of each measuring point in the carriage_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration and total number of bacterial colonies, and establishing a mapping relation between the total number D of the bacterial colonies in each micro-environment unit and the concentration D of the atmospheric pollutants; wherein the micro environment unit is a measuring point;

3) selecting an actually-measured atmospheric pollutant concentration data set with the time length of N minutes, calculating the total number of bacterial colonies according to the mapping relation, and recording the time sequence of the total number of the bacterial colonies at the ith seat asThe time series of the total bacterial colonies at the j-th air supply outlet or air exhaust outlet is recorded asPerforming hypothesis test and judgment by Glangel causal relationship testAndwhether a causal relationship exists or not is judged, and then a test result set of each seat test point and m air supply outlets and n air exhaust outlets is obtained;

4) acquiring a nonlinear description model library of all seat detection points according to the mapping relation and the inspection result set;

5) and taking the ventilation rates of all air supply outlets and air outlets of the train as the input of a wolf optimization algorithm, calculating the total bacterial colony number fitting results of the air supply outlets/the air outlets at different ventilation rates, taking the fitting results as the input of the nonlinear description model library to obtain the total bacterial colony number fitting results of all seats, and determining the ventilation rates of all the air supply outlets and the air outlets by using the total bacterial colony number fitting results of all the seats.

2. The method for regulating and controlling air in train carriages based on microorganisms as claimed in claim 1, wherein the concrete implementation process of establishing the mapping relation between the total number D of bacterial colonies and the concentration D of atmospheric pollutants in each micro-environmental unit in the step 2) comprises the following steps:

A. reading air pollutant concentration and bacterial colony total number index data sets of current micro-environment units in M continuous historical moments, and dividing the data sets into a training set and a testing set;

B. constructing a microorganism-air pollutant model by adopting a deep confidence network, taking the concentration of the air pollutants as the input of the deep confidence network, taking the total number of bacterial colonies at the same moment as the output of the deep confidence network, and training the deep confidence network;

C. taking the test set as the input of the trained deep confidence network, and selecting a group of parameters with highest description precision on the test set as a microorganism-air pollutant mapping model of the micro environmental unit;

D. repeating the steps A to C until all the tiny environment units are traversed, and obtaining the mapping relation between the total number of the bacterial colonies in the m + n + p detection points and the air pollutants; m, n and p are respectively the detection points of the air supply outlet, the air exhaust outlet and the seat.

3. The microorganism-based air conditioning method for train carriages as claimed in claim 1, wherein in step 3), the test result setWherein the content of the first and second substances,test resultsA value of 0 or 1; GCT () stands for glandor causal relationship test.

4. The method for regulating and controlling air in train carriages based on microorganisms as claimed in claim 1, wherein the concrete implementation process of the step 4) comprises the following steps:

I) reading the concentrations of a seat, an air supply outlet, an air exhaust outlet PM2.5, PM10, CO, NO2, SO2 and O3 at P continuous historical moments, and calculating the total number of bacterial colonies at each detection point in the P continuous historical moments according to the mapping relation;

II) reading the total number O of bacterial colonies at the detection point of the ith seat_iAnd the total bacterial colony number I of the air supply outlet/air exhaust outlet in causal relation with the ith seat detection point_i；

III) mixing I_iAs input to a deep echo state network, with O_iFor the output of the deep echo state network, learning the corresponding relation between the seats and the total bacterial colonies of the air supply outlet/the air exhaust outlet at different historical moments;

IV) repeating the steps I) to III) until all the seat detection points are traversed, and obtaining a nonlinear description model library of all the seat detection points.

5. The method for regulating and controlling the air in the train compartment based on the microorganisms as claimed in any one of claims 1 to 4, wherein the concrete implementation process for calculating the result of the bacterial colony total fitting of the air supply opening/air exhaust opening at different ventilation rates in the step 5) comprises the following steps:

i) increasing the ventilation rate at equal intervals, and determining the total number of bacterial colonies at the corresponding ventilation rate;

ii) performing least square fitting on the total number of bacterial colonies of the k air supply outlet/exhaust outlet to obtain the total number of bacterial coloniesWith respect to the rate v of ventilation_kThe polynomial expression method of (1);

iii) repeating the steps i) and ii) and traversing to all the air supply openings and the air exhaust openings to obtain a polynomial fitting result of the total bacterial colony number of all the air supply openings and the air exhaust openings along with the change of the ventilation ratem and n are respectively the detection points of the air supply outlet and the air exhaust outlet.

6. The microorganism-based air conditioning method for train carriages as claimed in claim 5, wherein in step 5), the optimization objective is set to simultaneously minimize the total number of bacterial colonies at each seat as a fitting result, and the optimization function is:u_kand l_kThe upper limit and the lower limit of the ventilation rate of the k-th air supply outlet/exhaust outlet are respectively.

7. The microbe-based air conditioning method for train carriages as claimed in claim 2, wherein in step 5), the evaluation index is selectedAchieving a minimum non-dominated solution NS^*Argmin E for determining the ventilation rate NS of all supply and exhaust openings^*(ii) a Wherein the content of the first and second substances,l_k≤v_k≤u_k；variance of total number of bacterial colonies at all seats in the test set; u. of_kAnd l_kThe ventilation rate v of the k-th air supply outlet/exhaust outlet respectively_kUpper and lower limits of (1).

8. A train compartment air conditioning system based on microorganisms, characterized by comprising computer equipment; the computer device is configured or programmed for carrying out the steps of the method according to one of claims 1 to 7.

9. A computer storage medium characterized by storing a program; the program is configured for carrying out the steps of the method according to one of claims 1 to 7.

Technical Field

The invention relates to the field of train environment monitoring, in particular to a method and a system for regulating and controlling air in a train compartment based on microorganisms and a storage medium.

Background

With the continuous development of the rail transit industry in China, the requirement of passenger train comfort is gradually concerned by the public. Under the influence of air pressure waves, proper internal and external pressure difference is required to be ensured when a train carriage runs at high speed, so that a high-speed train usually adopts a sealed train body structure, and all windows cannot be opened. In this case, the air pollutant treatment inside the cabin is entirely dependent on the ventilation system. Thus, the quality of the ventilation system and the regulation strategy will directly affect passenger comfort. How to monitor the train environment and adjust the ventilation system accordingly becomes an urgent problem to be solved.

The existing patents on the environment of train carriages mainly relate to the following two aspects:

1. novel air quality detection device and purifier's installation. For example, patent application with publication number CN101885338A proposes an intelligent sampling detection and air purification device for a train air conditioning ventilation system, which comprises a variable-frequency cyclone dust catcher, a high-efficiency filter and the like. Patent publication No. CN105172818A proposes a special air purifier for trains, which comprises an upper box body and a lower box body matched with the upper box body.

2. A carriage environment adjusting method based on atmospheric pollutant detection. The patent application with the publication number of CN110239577A provides a train passenger health protection system and a method thereof in an in-vehicle polluted environment, and the system comprises a basic data acquisition module, a train external air quality prediction module, a train internal air quality prediction module and a ventilation strategy formulation module. The patent application with the publication number of CN104608785A provides an intelligent management and control method for an air conditioning system of a high-speed train.

The method mainly uses the air pollutants such as PM2.5 and the like in the train as the basis for evaluating the air quality, however, the harm of biological pollution in the air environment to human health is not concerned, and no method pays attention to the biological pollutants in the closed environment of the train carriage at the present stage. In addition, because the microorganism measurement mechanism is different from pollutants such as PM2.5, long-time colony culture is required, and direct detection and real-time regulation and control measures are difficult to take.

Disclosure of Invention

The invention aims to solve the technical problem that the prior art is insufficient, and provides a method, a system and a storage medium for regulating the air ventilation of a train carriage based on microbial diffusion, so that the mapping relation between microbial pollution and atmospheric pollutant concentration in the carriage is learned, and optimal-level protection measures are taken for the health of passengers according to the microbial distribution condition in the carriage.

In order to solve the technical problems, the technical scheme adopted by the invention is as follows: a method for regulating and controlling air in a train compartment based on microorganisms comprises the following steps:

1) detecting PM at train supply air outlet, exhaust outlet and seat_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration and total number of bacterial colonies;

2) according to PM of each measuring point in the carriage_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration and total number of bacterial colonies, and establishing a mapping relation between the total number D of the bacterial colonies in each micro-environment unit and the concentration D of the atmospheric pollutants; wherein the micro environment unit is a measuring point;

3) selecting an actually-measured atmospheric pollutant concentration data set with the time length of N minutes, calculating the total number of bacterial colonies according to the mapping relation, and recording the time sequence of the total number of the bacterial colonies at the ith seat asThe time series of the total bacterial colonies at the j-th air supply outlet or air exhaust outlet is recorded asPerforming hypothesis test and judgment by Glangel causal relationship testAndwhether a causal relationship exists or not is judged, and then a test result set of each seat test point and m air supply outlets and n air exhaust outlets is obtained;

4) acquiring a nonlinear description model library of all seat detection points according to the mapping relation and the inspection result set;

5) and taking the ventilation rates of all air supply outlets and air outlets of the train as the input of a wolf optimization algorithm, calculating the total bacterial colony number fitting results of the air supply outlets/the air outlets at different ventilation rates, taking the fitting results as the input of the nonlinear description model library to obtain the total bacterial colony number fitting results of all seats, and determining the ventilation rates of all the air supply outlets and the air outlets by using the total bacterial colony number fitting results of all the seats.

The invention learns the mapping relation between the microbial pollution and the air pollutant concentration in the carriage and makes the optimal protection measure for the health of passengers in real time according to the microbial distribution condition in the carriage. The method innovatively detects and analyzes microorganisms in the train compartment. And adjusting the ventilation system according to the distribution condition of microorganisms among the measuring points, so as to reduce the microbial pollution of the area where the passengers are located. The method has a guiding function on the air quality control of the railway train.

In the step 2), the specific implementation process of establishing the mapping relation between the total number D of the bacterial colonies in each micro-environment unit and the concentration D of the atmospheric pollutants comprises the following steps:

A. reading air pollutant concentration and bacterial colony total number index data sets of current micro-environment units in M continuous historical moments, and dividing the data sets into a training set and a testing set;

B. constructing a microorganism-air pollutant model by adopting a deep confidence network, taking the concentration of the air pollutants as the input of the deep confidence network, taking the total number of bacterial colonies at the same moment as the output of the deep confidence network, and training the deep confidence network;

C. taking the test set as the input of the trained deep confidence network, and selecting a group of parameters with highest description precision on the test set as a microorganism-air pollutant mapping model of the micro environmental unit;

D. repeating the steps A to C until all the tiny environment units are traversed, and obtaining the mapping relation between the total number of the bacterial colonies in the m + n + p detection points and the air pollutants; m, n and p are respectively the detection points of the air supply outlet, the air exhaust outlet and the seat.

The invention researches the mapping relation between the microbial pollution and the air pollutant concentration, can effectively solve the problem of real-time property of microbial detection, and ensures real-time regulation and control of the microbial pollution in the train carriage.

In step 3), test result setWherein the content of the first and second substances,test resultsThe value is 0 or 1.

According to the method, the causal relationship inspection is carried out on the microorganism time sequence data among different measuring points, the measuring points closely related to the seats where passengers are located are further screened for subsequent modeling, the compression of the space dimensions of the measuring points is realized, and the provided data characteristics have strong characterization capability.

The specific implementation process of the step 4) comprises the following steps:

I) reading the concentrations of a seat, an air supply outlet, an air exhaust outlet PM2.5, PM10, CO, NO2, SO2 and O3 at P continuous historical moments, and calculating the total number of bacterial colonies at each detection point in the P continuous historical moments according to the mapping relation;

II) reading the total number of bacterial colonies at the ith seat detection pointAnd the total bacterial colony number of the air supply outlet/air exhaust outlet in causal relation with the ith seat detection point

III) mixing I_iAs input to a deep echo state network, with O_iFor the output of the deep echo state network, learning the corresponding relation between the seats and the total bacterial colonies of the air supply outlet/the air exhaust outlet at different historical moments;

IV) repeating the steps I) to III) until all the seat detection points are traversed, and obtaining a nonlinear description model library of all the seat detection points.

The invention adopts the deep neural network to describe the microorganism-air pollutant concentration and the seat-air supply outlet/air outlet microorganism nonlinear mapping relation, thereby ensuring the description precision.

In the step 5), the concrete implementation process for calculating the total bacterial colony fitting result of the air supply outlet/air outlet at different ventilation rates comprises the following steps:

i) increasing the ventilation rate at equal intervals, and determining the total number of bacterial colonies at the corresponding ventilation rate;

ii) performing least square fitting on the total number of bacterial colonies of the k air supply outlet/exhaust outlet to obtain the total number of bacterial coloniesWith respect to the rate v of ventilation_kThe polynomial expression method of (1);

iii) repeating the steps i) and ii) and traversing to all the air supply openings and the air exhaust openings to obtain a polynomial fitting result of the total bacterial colony number of all the air supply openings and the air exhaust openings along with the change of the ventilation ratem and n are respectively the detection points of the air supply outlet and the air exhaust outlet.

In step 5), setting an optimization target to simultaneously minimize the total number of bacterial colonies at each seat, wherein the optimization function is as follows:

the invention adopts a multi-objective optimization method to minimize the total number of bacterial colonies at each seat, so that the microbial pollution of the area where passengers are positioned is totally optimal, and the secondary pollution phenomenon of partial areas caused in the adjusting process of a ventilation system is avoided.

In step 5), selecting the evaluation indexAchieving a minimum non-dominated solution NS^*Argmin E for determining the ventilation rate NS of all supply and exhaust openings^*(ii) a Wherein the content of the first and second substances,variance of total number of bacterial colonies at all seats in the test set; u. of_kAnd l_kThe ventilation rate v of the k-th air supply outlet/exhaust outlet respectively_kUpper and lower limits of (d).

The evaluation index is a combination of a cumulative fit of the total number of bacterial colonies at all seats representing the degree of microbial contamination after ventilation control and a variance representing the degree of dispersion of microbial contamination between the seats. Choosing the non-dominant solution that minimizes the evaluation index ensures that: (1) the microbial pollution degree in the carriage is minimum overall; (2) the difference of microbial contamination between the seats is minimized, avoiding extreme contamination in individual areas.

The invention also provides a train carriage air ventilation and conditioning system based on microbial diffusion, which comprises computer equipment; the computer device is configured or programmed for performing the steps of the method of the invention.

As an inventive concept, the present invention also provides a computer storage medium storing a program; the program is configured for performing the steps of the method of the invention.

Compared with the prior art, the invention has the beneficial effects that:

1) the invention detects and analyzes the microorganisms in the train compartment. And adjusting the ventilation system according to the microbial diffusion condition among the measuring points, so as to reduce the microbial pollution index of the area where the passenger is located. The method has a guiding function on the air quality control of the railway train.

2) The invention researches the mapping relation between the microbial pollution and the concentration of the atmospheric pollutants, can effectively solve the problem of real-time property of microbial detection, and ensures real-time regulation and control of the microbial pollution in the train carriage.

3) The invention adopts a mode of comprehensively detecting the air supply outlet, the air exhaust outlet and the seat multiple measuring points, can effectively describe the distribution condition of atmospheric pollutants and microorganisms in the internal environment of the carriage, and ensures the depicting authenticity of the detection result on the actual spatial distribution condition.

4) According to the method, the causal relationship inspection is carried out on the microorganism time sequence data among different measuring points, the measuring points closely related to the seats where passengers are located are further screened for subsequent modeling, the compression of the space dimensions of the measuring points is realized, and the provided data characteristics have strong characterization capability.

5) The invention adopts the deep neural network to describe the microorganism-atmospheric pollutant concentration and the seat-air supply outlet/air outlet microorganism nonlinear mapping relation, thereby ensuring the description precision.

6) The invention adopts a multi-objective optimization method to minimize the total number of bacterial colonies at each seat, so that the microbial pollution of the area where passengers are positioned is totally optimal, and the secondary pollution phenomenon of partial areas caused in the adjusting process of a ventilation system is avoided.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

Detailed Description

As shown in fig. 1, the specific implementation process of the embodiment of the present invention is as follows:

step 1: multi-point acquisition of pollution data

The interior of the train compartment includes PM_2.5、PM₁₀、CO、NO₂、SO₂And O₃Six kinds of air pollutants and microbial pollution such as bacteria, fungi, viruses and the like. The microorganisms are closely related to the air quality condition, and generally, the bacterial colonies in the air are alwaysThe number is positively correlated with the existence probability of pathogenic microorganisms (bacteria, fungi and viruses), so the patent measures the pathogenicity of the microorganisms by the total number of bacterial colonies. Arranging TS WES-C air pollutant detector (for measuring PM) at multiple air supply outlet, air outlet and seats of train carriage_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration, real-time detection) and anderson impact air microorganism sampler (measuring total bacterial colony count, 48h microbial culture).

The obtained data includes air supply outlet, air outlet and PM at the seat_2.5Concentration, PM₁₀Concentration, CO concentration, NO₂Concentration, SO₂Concentration, O₃Concentration and total number of bacterial colonies, which can be expressed asAndin the formulaRepresenting the concentration of atmospheric pollutants at the mth blower outlet,representing the concentration of atmospheric pollutants at the nth exhaust outlet,represents the concentration of atmospheric pollutants at the p-th seat, whereinRepresenting the total number of bacterial colonies at the mth supply air outlet,represents the total number of bacterial colonies at the n-th exhaust port,represents the total number of bacterial colonies at the p-th seat, i represents six atmospheric pollutants of PM2.5, PM10, CO, NO2, SO2 and O3, and m, n and O are the number of measuring points of an air supply opening, an air exhaust opening and a seat respectively. And regarding each detection point as a micro environment unit, corresponding detection data to a carriage number, and recording a time stamp of the detection data, wherein the interval between adjacent data is 5 minutes. And the collected data is transmitted to the data storage platform in a 4G mode.

Step 2: microorganism-atmospheric pollutant mapping learning

And establishing a model to learn the mapping relation between the total number D of the bacterial colonies in each micro-environment unit and the concentration D of the atmospheric pollutants according to the historical pollution data of the measuring points of the carriage. The specific modeling process is as follows:

a1: selecting a micro-environment unit, and reading an index data set of the concentration of the atmospheric pollutants and the total number of bacterial colonies of the micro-environment unit in 200 continuous historical moments.

A2: and (4) dividing the data set. The data set comprises 200 continuous historical moments, data at 1-160 moments are used as a training set, and data at 161-200 moments are used as a testing set.

A3: and (3) constructing a microorganism-atmospheric pollutant model by adopting a deep confidence network, taking the concentration of the atmospheric pollutants as the input of the deep confidence network, and taking the total number of bacterial colonies at the same moment as the output of the deep confidence network. The number of layers of the deep confidence network is determined by 5-fold cross validation, and the range is selected to be [1,2,3,4,5 ].

A4: and performing model inspection on the test set by using the well-trained deep confidence network, and selecting a group of parameters with highest description precision on the test set as a microorganism-atmospheric pollutant mapping model of the micro environmental unit.

A5: traversing a 1-a 4 to all tiny environment units (i.e., detection points), and obtaining a mapping relation between the total number of bacterial colonies and the atmospheric pollutants in all m + n + p detection points { D (D) | i ═ 1,2, 3.

And step 3: point-of-effect testing based on microbial diffusion mechanism

The space distribution and the diffusion of microorganisms in the carriage are influenced by air movement, and the total number of bacterial colonies among all measuring points has a causal relationship of a data layer. And analyzing the causal relationship between the seats and the time series of the total bacterial colonies of the air supply outlets and the air exhaust outlets aiming at each carriage.

And (3) selecting an actually measured atmospheric pollutant concentration data set with the time length of N minutes, and calculating the total number of bacterial colonies according to the mapping relation obtained in the step (2). The time series of the total number of bacterial colonies at the ith seat was recordedThe time series of the total bacterial colonies at the j-th air supply outlet or air exhaust outlet is recorded asThe judgment was made by performing hypothesis testing using the Granger Causality Test (GCT)Andwhether a causal relationship exists. Test resultsThe output is 0 or 1, where 0 represents the time series of the total number of bacterial colonies at the seatTime sequence of air supply outlet/air exhaust outlet bacterial colony countNo causal relationship exists, whereas 1 represents the existence of a causal relationship:

GCT () stands for glandor causal relationship test. Obtaining a test result set of each seat detection point and m air supply outlets and n air exhaust outlets:

and 4, step 4: causal association measurement point nonlinear description modeling

Establishing a non-linear description model of a related air supply outlet/air outlet aiming at each seat detection point, wherein the specific modeling process is as follows:

b1: and (3) reading the concentrations of the seat, the air supply outlet, the air exhaust outlet PM2.5, the PM10, the CO, the NO2, the SO2 and the O3 at 100 continuous historical moments, and calculating the total bacterial colony number of each measuring point in the 100 continuous historical moments according to the mapping relation obtained in the step (2).

B2: and (4) dividing the data set. The data set comprises 100 continuous historical moments, data at 1-60 moments are used as a training set, data at 61-80 moments are used as a verification set, and data at 81-100 moments are used as a test set.

B3: reading the total number of bacterial colonies at the ith seat detection pointAnd the total bacterial colony number of the air supply outlet/air exhaust outlet in causal relation with the ith seat detection point

B4: constructing a nonlinear description model by adopting a deep echo state network, wherein the input of the model is I_iThe model output is O_iSo as to learn the corresponding relation between the seats and the total bacterial colonies of the air supply outlet/the air exhaust outlet at different historical moments. Setting the number of the storage pool nodes of the deep echo state network to be 10, determining the number of the storage pool layers and the spectrum radius of the storage pool matrix in each layer by adopting 5-fold cross validation, wherein the selection ranges of the two parameters are [1,2,3,.. multidot.. 10 ] respectively]And [0.1,0.3,0.5,0.7,0.9 ]]Selecting a group of parameters with highest description precision on the verification set to obtain a completely trained nonlinear description model h (I)_i)。

B5: traversing A1-A4 to all the seat detection points to obtain a nonlinear description model library { h (I) of all the seat detection points_i)|i＝1,2,3,...,p}。

And 5: compartment ventilation adjusting strategy based on multi-objective optimization

C1: and measuring the change relation of the total bacterial colony number at all the air supply openings/air exhaust openings along with the ventilation rate. The method comprises the following steps:

1) the ventilation rate is increased at equal intervals, and the total number of bacterial colonies at the corresponding ventilation rate is determined and recorded in the data storage platform in a data format of time stamp-ventilation rate-total number of bacterial colonies.

2) Aiming at the kth air supply outlet/air outlet, performing least square fitting to obtain the total number of bacterial coloniesWith respect to the rate v of ventilation_kThe polynomial expression method of (1):

3) repeating the steps, traversing to all the air supply outlets and the air outlets, and obtaining a polynomial fitting result of the total bacterial colony number of all the air supply outlets and the air outlets changing along with the ventilation rate

C2: establishing a multi-objective optimization model, wherein the specific implementation details are as follows:

1) selecting an optimization algorithm and setting initial hyper-parameters: multi-objective Grey wolf optimization is employed, With a leader selection mechanism and archive storage mechanism embedded to improve convergence (MIRJALILI S, SAREMI S, MIRJALILI S M, et al. Multi-objective greenwolf optimizer [ J. Expert Systems With Applications,2016,47: 106-19.). The number of search populations, maximum number of iterations, and archive size for the multi-objective grayish wolf optimization are set to 200, 100, and 50, respectively.

2) Optimizing variables to be the ventilation rate v of all the air supply openings and the air exhaust openings, wherein the search range of the variables meets the following formula:

l_k≤v_k≤u_k

wherein u is_kAnd l_kThe upper and lower limits of the ventilation rate of the k-th air supply outlet/exhaust outlet are respectively.

3) And calculating the total bacterial colony fitting results of the air supply openings/the air exhaust openings at different ventilation rates according to a polynomial fitting method of the total bacterial colony numbers of the air supply openings and the air exhaust openings, which is obtained by the step C1, along with the change of the ventilation rates. And inputting the total number of the obtained bacterial colonies into a non-linear description model library of the detection points of the seats obtained by B5, and outputting a fitting result of the total number of the bacterial colonies at each seat. Setting an optimization target to simultaneously minimize the total number of bacterial colonies fitting results at each seat, wherein the optimization function is as follows:

4) multi-objective optimization (MIRJALILI S, SAREMI S, MIRJALILI S M, et al, multi-objective greeny wolf optimizer [ J ]. Expert Systems With Applications,2016,47:106-19.) was performed, and the number of iterations, Itr, was recorded as 1. And calculating the optimization function values of all the search results, and selecting a non-dominated solution to store in the file.

5) And updating the search path and generating a new ventilation rate search result.

6) If It is less than the maximum iteration number, returning to step 4); otherwise, the multi-objective optimization algorithm is ended, and the non-dominated solution set NS in the final archive is output.

7) Performance of the non-dominated solution on the test set was evaluated as a combination of the total number of bacterial colonies at the seat and the cumulative fit and variance (Var) of the total number of bacterial colonies for all seats:

selecting a non-dominated solution NS which minimizes the evaluation index^*Argmin E, for determining the ventilation rate of all supply and exhaust ports.

Step 6: and after the ventilation adjustment of the train carriage is completed according to the obtained ventilation rate, continuously detecting the total number of the bacterial colonies by each detection point and transmitting the data to the data storage platform.

And 7: model training is not required to be carried out again within a period of time after the first ventilation adjustment is finished, and only calculation is carried out according to subsequent detection data and an optimal ventilation adjustment strategy is output. The distribution conditions of the air microorganisms can change due to different crowd behaviors, the causal relationship test, the nonlinear description and the multi-objective optimization model all need to be retrained periodically, parameters are updated to ensure the effectiveness of the model, and the retraining time interval can be set to be 3 hours.