A digital twin system for analyzing air pollution distribution within transportation tracks

Abstract

The invention discloses a digital twin system for analyzing air pollution distribution in a traffic track, relates to the technical field of experimental analysis of traffic tracks, and discloses a collection end, a simulation end, and a pollution distribution data prediction model; the collection end includes: a collection wind speed sensor, a collection wind direction sensor, a collection formaldehyde detector, a collection haze gas dust sensor, a collection CO2 sensor, a collection temperature and humidity sensor, a collection infrared sensor, a contact sensor, an infrared camera, an ordinary camera, a collection central control, and a GPS; through the setting of the pollution distribution data prediction model, it is calculated where the dust is concentrated and where the density is the highest under different dust inputs, different wind speeds and directions, and different door opening and closing times, so as to know the further control scheme of using a low-temperature plasma purifier and an air conditioner.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic track experimental analysis, and more specifically, to a digital twin system for analyzing air pollution distribution in a traffic track. Background Art

[0002] Low-temperature plasma purifiers are widely used in waste gas treatment projects. They are equipped with a unique collision and suction unit to intercept and remove particulate matter in the waste gas. The multi-element gas collected by the waste gas collection system passes through the plasma active oxygen purification device. Under the action of the high-voltage plasma electric field, the initial oxygen is ionized to ionize and charge the waste gas ions therein, and the charged tiny ions (dust particles) are collected by the adsorption unit and flow into and deposited in the dust storage box of the gas treatment device. The harmful gas in the gas is sterilized by the ozone generated in the electric field, and the odor is removed. The harmful gas is removed to achieve the purpose of waste gas treatment.

[0003] At present, low-temperature plasma purifiers are widely used in transportation rail carriages to remove volatile organic compounds (VOCs). These VOCs mainly include building materials, interior decoration materials, and daily and office supplies. For example: organic solvents, paints and water-containing coatings; second, human excrement; third, outdoor industrial waste gas, automobile exhaust, photochemical smog, etc. Among them, the organic solvents, paints and water-containing coatings in the carriages will be put into use after the carriages are produced and left to stand for a period of time, and then meet the standards. Therefore, when the carriages are used, the main pollutants they face come from human excrement, various exhaust gases that enter when the doors are opened, and various odors when people are using them.

[0004] At present, people install low-temperature plasma purifiers and air conditioners inside the carriages to adjust the air quality inside the carriages, but there is no further research on how to set the parameters and location of the low-temperature plasma purifiers, as well as the resulting lifespan costs. If this type of research is to be carried out, people need an experimental system, one part of which is installed on a real train to collect and analyze data from the actual carriages, and the other part is installed on the simulation end, which can generate the corresponding air pollution inside the actual carriages, and then use low-temperature plasma purifiers and air conditioners to eliminate this pollution, analyze the elimination process, study its parameters, and realize data twins. Summary of the invention

[0005] In view of the shortcomings of the prior art, the object of the present invention is to provide a digital twin system for analyzing the distribution of air pollution in transportation tracks.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A digital twin system for analyzing the distribution of air pollution in a transportation track, including a collection end, a simulation end, and a pollution distribution data prediction model;

[0008] The collection end includes: collection wind speed sensor, collection wind direction sensor, collection formaldehyde detector, collection haze gas dust sensor, collection CO2 sensor, collection temperature and humidity sensor, collection infrared sensor, contact sensor, infrared camera, ordinary camera, collection central control, GPS;

[0009] The analog end includes: analog wind speed sensor, analog wind direction sensor, analog formaldehyde detector, analog haze gas dust sensor, analog CO2 sensor, analog temperature and humidity sensor, analog infrared sensor, analog central control, haze gas dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier, air conditioner;

[0010] The collection end is arranged in the collection carriage, and the simulation end is arranged in the experimental carriage;

[0011] The wind speed sensor, wind direction sensor, formaldehyde detector, haze gas and dust sensor, CO2 sensor, temperature and humidity sensor, infrared sensor, contact sensor, infrared camera, ordinary camera and GPS are marked as collection sensors. The collection sensors are used to collect the internal data of the collection compartment and send the internal data to the collection central control. The collection central control sends the internal data to the simulation central control through the network. The simulation central control controls the operation of the haze gas and dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier and air conditioner.

[0012] The pollution distribution data prediction model includes LRCN network module, hole sampling data preprocessing module, smoothing prediction module, and fusion output module;

[0013] The LRCN network module has a 5-layer structure, each layer includes 1024 standard LSTMs, which are used for basic learning data sets and predicting output data based on input data;

[0014] The hole sampling data preprocessing module is used to cover the required working time with 1024 data. The working time is the time when the haze gas and dust emitted by the haze gas and dust generator is blown into the experimental compartment by the wind, and then the exit door and entrance door inside the experimental compartment are closed. The haze gas and dust float stably inside the experimental compartment. The working time is marked as T1. Using the formula The hole sampling rate HS is obtained, and the simulated wind speed sensor, simulated wind direction sensor, simulated formaldehyde detector, simulated haze gas dust sensor, simulated CO2 sensor, simulated temperature and humidity sensor, and simulated infrared sensor are marked as simulated sensors; where Sam is the unified sampling time of the simulated sensors;

[0015] The smoothing prediction module uses the data of the dint(ih, Hs+ih, …, nh*Hs+i)th of the total data set input part Dint for prediction each time, and obtains Hs prediction results; among them, ih is the smoothing number, from 1 to Hs-1; nh is dow(Dt / Hs)-1, dow is rounded down; the first group of data is dint(1, Hs+1, …, nh*Hs+1), the second group of data is dint(2, Hs+2, …, nh*Hs+2), and so on to obtain Hs groups of input data, and obtain the output Dfoutt(hsout) of the Hs group prediction, hsout from 1 to HS;

[0016] The fusion output module is used to add all Dfoutt(hsout) and take the average value.

[0017] Furthermore, wind speed sensors and wind direction sensors are installed at the two entrances of the collection compartment and at the upper middle position, with a total of three groups for obtaining wind speed and wind direction data. The temperature and humidity sensors are installed together with the wind speed sensors and wind direction sensors.

[0018] Furthermore, the formaldehyde detector includes two groups, one of which is placed at the upper middle position of the collection carriage to obtain formaldehyde data, and the other is installed in the toilet of the collection carriage.

[0019] Furthermore, there are thirteen groups of haze gas and dust sensors, two groups are installed at the two entrances of the collection carriage, ten groups are installed at equal distances on both sides of the collection carriage, and one group is installed in the toilet of the collection carriage. The CO2 collection sensor is installed together with the haze gas and dust collection sensor.

[0020] Furthermore, a collection infrared sensor is installed above each seat in the collection carriage to analyze whether there is someone in the seat. A contact sensor is installed on the two entrance doors of the collection carriage to analyze whether the entrance doors are closed. A group of infrared cameras and ordinary cameras are used to analyze the actual number of people in the collection carriage. GPS is used to obtain the location of the collection end.

[0021] Furthermore, the acquisition central control is used to obtain the data of the acquisition sensor, analyze whether the data is normal, save the data after adding a time tag, and then send the data to the simulation central control through the network according to the instruction of the simulation central control. The specific steps of analyzing whether the data is normal are:

[0022] Determine whether the acquisition sensor obtains data according to the normal time interval FT(i). If it exceeds 3 times FT(i), it is judged to be disconnected, where i is the number of the acquisition sensor;

[0023] Determine whether the data exceeds the set threshold range.

[0024] Furthermore, haze gas dust generators were installed at two entrances and exits of the experimental carriage, and under the control of the simulated central control, air pollutants of PM2.5 and PM10 were spread into the air.

[0025] Furthermore, a nitrogen generator and a CO2 generator are used to simulate the exhaust of abdominal gas by passengers.

[0026] Furthermore, simulated wind speed sensors and simulated wind direction sensors are installed at the two entrances of the experimental carriage and at the upper middle position, with a total of 3 groups. Simulated temperature and humidity sensors are installed together with the simulated wind speed sensor and the simulated wind direction sensor. The simulated formaldehyde detector includes two groups, one of which is placed at the upper middle position of the experimental carriage to obtain formaldehyde data, and the other group is installed in the bathroom. There are thirteen groups of simulated haze gas and dust sensors, two of which are installed at the two entrances of the experimental carriage, ten groups are installed at equal distances on both sides of the carriage, and one group is installed in the bathroom. The simulated CO2 sensor is installed together with the simulated haze gas and dust sensor, and the simulated infrared sensor is installed above each seat.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. Through the collection end, simulation end and settings, the air condition of the actual running rail carriage is collected, and the data will be collected and provided to the simulation end. After the simulation end obtains the data, its control system generates the corresponding air condition, and also controls the low-temperature plasma and air conditioning. The algorithm module continues to monitor the air condition inside the simulation end to avoid the impact on passengers caused by improper handling in the actual carriage;

[0029] 2. By setting up the pollution distribution data prediction model, we calculated where the dust is concentrated and where the density is highest under different dust inputs, different wind speeds and directions, and different door opening and closing times, so as to know the further control plan of using low-temperature plasma purifiers and air conditioners. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a principle block diagram of the present invention;

[0031] Figure 2 It is a processing flow chart of the pollution distribution data prediction model of the present invention;

[0032] Figure 3 A top view of the module distribution of the acquisition end of the present invention;

[0033] Figure 4 A top view of the module distribution of the analog end of the present invention;

[0034] Figure 5This is an internal structure diagram of the haze gas dust generator of the present invention;

[0035] Figure 6 For the present invention Figure 5 Left side view;

[0036] Figure 7 For the present invention Figure 5 Right side view;

[0037] Figure 8 It is a schematic diagram of the structure of the nitrogen generator and CO2 generator of the present invention. DETAILED DESCRIPTION

[0038] Reference Figure 1

[0039] A digital twin system for analyzing the distribution of air pollution in a transportation track includes a collection end, a simulation end, and a pollution distribution data prediction model.

[0040] Reference Figure 3 , the collection end includes: collection wind speed sensor ( Figure 3 1A, 2, 3 positions), wind direction sensor ( Figure 3 1A, 2, 3 positions), formaldehyde detector ( Figure 3 Middle 2 position), collecting haze gas dust sensor ( Figure 3 1A, 3, 4 positions), CO2 collection sensor ( Figure 3 1A, 3, 4 positions) position, collect temperature and humidity sensors ( Figure 3 1A, 2, 3 positions), infrared sensor, contact sensor ( Figure 3 1A, 3 positions), infrared camera ( Figure 3 1A position), ordinary camera ( Figure 3 1A position), acquisition and central control ( Figure 3 2 positions), GPS( Figure 3 2 position), can be marked as a collection sensor.

[0041] Reference Figure 4 , the analog end includes: analog wind speed sensor ( Figure 4 1B, 2, 3 positions), analog wind direction sensor ( Figure 4 1B, 2, 3 positions), simulated formaldehyde detector ( Figure 4 Middle 2 position), simulated haze gas dust sensor ( Figure 4 1B, 3, 4 positions), analog CO2 sensor ( Figure 4 1B, 3, 4 positions), analog temperature and humidity sensor ( Figure 4 1B, 2, 3 positions), simulated infrared sensor, simulated central control ( Figure 3Middle 2 position), haze gas dust generator ( Figure 4 Middle 5 position), nitrogen generator ( Figure 4 6A, 6B position), CO2 generator ( Figure 4 6A, 6B position), low temperature plasma purifier ( Figure 4 Middle 21 position), air conditioning ( Figure 4 21 position), the simulated wind speed sensor, simulated wind direction sensor, simulated formaldehyde detector, simulated haze gas dust sensor, simulated CO2 sensor, simulated temperature and humidity sensor, and simulated infrared sensor are marked as simulated sensors.

[0042] The collection end is set in the collection carriage, and the simulation end is set in the experimental carriage.

[0043] The wind speed sensor, wind direction sensor, formaldehyde detector, haze gas and dust sensor, CO2 sensor, temperature and humidity sensor, infrared sensor, contact sensor, infrared camera, ordinary camera and GPS are marked as collection sensors. The collection sensors are used to collect the internal data of the collection car and send the internal data to the collection central control. The collection central control sends the internal data to the simulation central control through the network. The simulation central control controls the operation of the haze gas and dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier and air conditioner.

[0044] The collection end is placed on the collection carriage to collect the air conditions of the actual running rail carriage. The data will be collected and provided to the simulation end. The train carriage used for actual testing is called the collection carriage. The simulation end is placed on the experimental carriage. After the simulation end obtains the data, the simulation central control controls the low-temperature plasma purifier, air conditioner, CO2 generator, nitrogen generator, and haze gas dust generator, and the simulation sensor performs continuous monitoring. The reason for not directly experimenting on the collection end is to prevent improper handling during the experiment from affecting passengers.

[0045] The acquisition sensor provides the data to the acquisition central control at the acquisition end, which summarizes the data, submits it to the network, and transmits it to the simulation central control at the simulation end.

[0046] The wind speed sensor and wind direction sensor are installed at the two entrances of the collection compartment and the upper middle position, with a total of 3 sets, for obtaining wind speed and wind direction data. In this embodiment, the sensor uses the HBOBOP brand micro 4-element weather station, which includes 4 elements such as wind speed, wind direction, temperature, and humidity, and sends the data to the collection control through the serial port.

[0047] Only one set of formaldehyde detector is needed to be placed in the middle of the carriage to obtain formaldehyde data. Because the carriage will be put into use after being produced and left to stand for a period of time and the formaldehyde reaches the standard, only one set is needed to deal with unexpected situations. In this embodiment, the sensor uses the MEF500 formaldehyde detector of the Slezhi brand, and sends the data to the collection control through the serial port.

[0048] The haze gas dust sensor is for PM2.5 and PM10 pollutants, which mainly come from the opening and closing of doors, as well as the dust brought by the people coming in. 12 groups are needed, 2 groups are installed at the two entrances of the carriage, and 10 groups are divided into 5m*2m and installed on the two sides of the carriage at equal distances. In this embodiment, the sensor uses the VMS-3002-PM haze gas dust sensor of the VMS brand, and sends the data to the collection control through the serial port.

[0049] The CO2 sensor is aimed at the source of CO2, which comes from the outside when the door is opened and closed, and is more oriented to activities such as people farting, because the main component of farts is nitrogen, followed by CO2. If the change of CO2 can be analyzed, the emission of farts can basically be analyzed. It has a total of 12 groups, which are installed together with the haze gas dust sensor. In this embodiment, the sensor uses the DK model CO2 sensor of the Dekong Senshe brand, and sends the data to the collection control through the serial port.

[0050] There are 3 sets of temperature and humidity sensors for collecting the temperature and humidity information of the carriage, which are installed together with the wind speed sensor and wind direction sensor. In this embodiment, the sensor uses the HBOBOP brand micro 4-element weather station, which includes 4 elements such as wind speed, wind direction, temperature and humidity.

[0051] The infrared sensor is installed above each seat to analyze how many passengers are actually sitting in the carriage. Taking hard seats as an example, 118 sensors are installed. In this embodiment, the sensor uses the E3F-DS30C4 model, three-wire sensor, NPN, normally open, 24V infrared sensor.

[0052] The contact sensor is installed on the doors of the two entrances of the train carriage to analyze whether they are closed. In this embodiment, the sensor uses the PM-L45P type, and the limit is close to the 24 volt infrared emitting tube.

[0053] One set of infrared camera and one set of ordinary camera are placed in the middle of one side of the collection carriage, with the axis pointing to the longest side of the carriage, to analyze the activities of people, how many people are not seated, but just standing in the aisle, so as to analyze the actual number of people in the collection carriage. In this embodiment, an Intel RealSense D455 stereo depth camera is used, which has binoculars, built-in infrared camera and ordinary camera, and comes with a human body analysis algorithm module, and obtains data through USB.

[0054] GPS is used to obtain the location of the acquisition end, because different locations are also affected by the external air quality. In this embodiment, the BT-F9PK3 board of BeiDou BeiTian brand is used to obtain data through the serial port.

[0055] The acquisition control is used to obtain the data of the acquisition sensor, analyze whether the data is normal, save the data after adding a time tag, and then send the data to the simulation control through the network according to the instruction of the simulation control. In this embodiment, the acquisition control uses an industrial computer of Advantech brand, with 2 sets of 100-channel switch acquisition cards (Shandong Piyi brand PY-810 model) and 1 set of 48-channel digital acquisition cards (Taiwan ICP DAS brand PIO-D48U model).

[0056] The specific analysis of whether the data is normal is as follows:

[0057] Determine whether the acquisition sensor obtains data according to the normal time interval FT(i). If it exceeds 3 times FT(i), it is judged to be disconnected, where i is the number of the acquisition sensor. These values ​​are set by the staff. In this embodiment, the acquisition cycle of each sensor is 50 milliseconds and 1 minute, so the corresponding FT(i) is 50 milliseconds and 1 minute.

[0058] To determine whether the data exceeds the set threshold range, the maximum value DTMax(i) and the minimum value DTMin(i) of each sensor are set by the staff. In this embodiment, these values ​​are set by the staff.

[0059] The network is a channel for receiving instructions and sending data. The normal wifi network is used here.

[0060] The interior environment of the experimental carriage is exactly the same as that of the collection carriage. It is used to simulate the environment of the collection carriage. It can be a room or a real carriage. It is used to install all the equipment on the simulation end.

[0061] The equipment model, quantity, installation location, and purpose of use of the simulated wind speed sensor, simulated wind direction sensor, simulated formaldehyde detector, simulated haze gas and dust sensor, simulated CO2 sensor, simulated temperature and humidity sensor, and simulated infrared sensor are exactly the same as those in the collection carriage.

[0062] The simulation central control is used to issue instructions for data transmission to the acquisition central control under the control of the operator, receive data from the acquisition central control, and display this data in three dimensions; under the operation of the operator, the actions and parameters of the haze gas dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier, and air conditioner are set, and then the data from the simulated sensors is used to observe the process and analyze the control parameters.

[0063] Reference Figure 5-Figure 7 The haze gas dust generator is used at the two entrances and exits of the experimental carriage. Under the control of the simulated central control, common PM2.5 and PM10 air pollutants are spread into the air as required. Its structure includes a hopper ( Figure 5 Middle 7 position), hopper electric control switch ( Figure 5 Middle 8 position), hair dryer blowing from bottom to top 1 ( Figure 5 9 position), hair dryer blowing from the side 2 ( Figure 5 10 position), single chip microcomputer 1 ( Figure 6 Middle 14 position), controllable switch door 1 ( Figure 6 Center 12 position), exhaust fan 1 ( Figure 5 Middle 11 position), controllable switch door 2 ( Figure 7 13 position) etc.

[0064] The hopper is a container for placing haze gas dust. It is Y-shaped with a partition on top. Personnel can open the partition to add dust. The partition is usually closed. There is an electric control switch for the hopper under the Y-shaped, and the switch is opened and closed by the single-chip computer 1. When it is opened, the dust will fall into the inner bin of the generator due to its own gravity. When it is closed, the dust cannot fall into the inner bin and is isolated in the hopper. In this embodiment, the hopper is customized.

[0065] Among them, the hopper electric control switch is controlled by the single-chip computer 1. According to the area and opening time of the switch, the mass of the dust falling into the inner bin can be calculated, so as to achieve the quantity required by the simulated central control. In this embodiment, there is a metal tube under the hopper. An electric control switch is installed on one side of the tube, which is a 20kg high-torque servo. A metal sheet is installed on the servo rotor. By controlling the rotation of the servo, the metal sheet can block the metal tube. There is a groove on the other side of the metal tube servo. When the metal tube is blocked, it stays on the groove.

[0066] The blower 1 blowing from bottom to top provides buoyancy for the dust falling into the inner bin, so that the dust is evenly distributed in the inner bin and diffused in the air. In this embodiment, an industrial electric fan is used.

[0067] The blower 2 blowing from the side is a device that blows from the side to the inside of the compartment. When the wind speed reaches a stable level, the single chip computer 1 opens the controllable switch door 1 of the inner compartment, so that the dust blows a certain amount of pollutants into the compartment at a specified speed. In this embodiment, an industrial electric fan is used.

[0068] The controllable switch door 1 is controlled by the single chip computer 1, and when it is opened, the dust from the haze gas dust generator will be blown into the interior of the carriage. In this embodiment, an electric sliding door with a size half that of the carriage door is used, which moves to the side.

[0069] Among them, the exhaust fan 1 and the controllable switch door 2 cooperate with another group of haze gas dust generators in the carriage to simulate the state of adding dust to the carriage at the same time, or adding and closing at the same time, or adding and discharging at the same time. If one side is added and closed at the same time, then the exhaust fans 1 and controllable switch doors 2 of the two groups of haze gas dust generators are all closed. If one side is added and discharged at the same time, then the exhaust fan 1 and controllable switch door 2 of the working haze gas dust generator are closed, and the exhaust fan 1 and controllable switch door 2 of the other non-working haze gas dust generator are opened to remove polluted air. Finally, if the experiment is over and the format is cleaned, the exhaust fans 1 and controllable switch doors 2 of the two groups of haze gas dust generators are all opened to remove polluted air. In this embodiment, the exhaust fan 1 adopts an industrial exhaust fan, and the controllable switch door 2 adopts an electric sliding door that is half the size of the carriage door and moves to the side.

[0070] Among them, the single-chip microcomputer 1 controls the operation of the haze gas dust generator and is controlled by the analog central control. In this embodiment, the STM32 single-chip microcomputer development board is used.

[0071] Reference Figure 8 The nitrogen generator and CO2 generator are used to simulate the exhaust of abdominal gas by passengers and are placed on the seats in the experimental compartment. Its structure includes a nitrogen and CO2 mixed gas cylinder ( Figure 8 15 position), electric switch 1 ( Figure 8 16 position), warehouse body ( Figure 8 17 position), CO2 sensor ( Figure 8 Middle 18 position), air pump ( Figure 8 19 position), electric switch 2 ( Figure 8 Middle 20 position), hose ( Figure 8 20 position), single chip microcomputer 2 ( Figure 8 18 position) etc.

[0072] The nitrogen and CO2 mixed gas cylinder is a compressed gas cylinder, which contains a mixture of nitrogen and CO2 at a mass ratio of 4 to 1, and tries to simulate the gas excretion of the human body. The cylinder mouth is equipped with an electric switch 1 to control its switch. In the open state, the compressed gas can enter the warehouse through the electric switch; in the closed state, the compressed gas cannot enter the warehouse through the electric switch.

[0073] The electric switch 1 is used to be controlled by the single chip microcomputer 2 to control the compressed gas to enter the warehouse through the electric switch, thereby ensuring the concentration of the compressed gas in the warehouse.

[0074] Among them, the warehouse body is a container for storing compressed gas, CO2 sensor, and air pump.

[0075] The CO2 sensor is placed inside the warehouse to detect the concentration of the compressed gas inside and send the concentration data to the single chip microcomputer 2. If the concentration is not enough, the single chip microcomputer 2 will turn on the electric switch 1 to put the compressed gas into the warehouse; if the concentration meets the standard, the electric switch 1 will be turned off. The sensor uses the DK model CO2 sensor of the Dekong Senshe brand, and sends the data to the single chip microcomputer 2 through the serial port.

[0076] Among them, the air pump is placed inside the warehouse body, one is sound insulation, and the other is to directly obtain the gas input source from the inside of the warehouse body. Under the control of the single-chip computer 2, pressure is generated and connected to the seats in the carriage and the inside of the toilet through multiple hoses. The electric switch 2 is installed at the end of each hose and is also controlled by the single-chip computer 2. It is used to release and close the seats in the carriage and the gas in the hose inside the toilet to simulate the human body to discharge abdominal gas. In this embodiment, the air pump uses a Feibao silent oil-free air compressor small industrial-grade air pump, which has 188 sub-heads, and the switch uses a 10mm hose electromagnetic control switch valve from Chongqing Yisai, which is controlled by the single-chip computer 2.

[0077] Among them, the single chip microcomputer 2 ensures the operation of the nitrogen generator and the CO2 generator, which is controlled by the analog central control. In this embodiment, the single chip microcomputer 2 adopts the STM32 single chip microcomputer development board, plus 2 groups of 100 IO port data acquisition cards.

[0078] The low-temperature plasma purifier and air conditioner are air purification devices, and their parameters are controlled by the simulation central control, and the operator of the simulation central control sets their parameters according to needs. In this embodiment, the low-temperature plasma purifier and air conditioner customized by Jiangsu Wellsaf Electrical Technology Co., Ltd. are used.

[0079] The pollution distribution data prediction model is used to analyze and predict the pollution inside the experimental compartment by combining the pollution data inside the experimental compartment with the wind direction and speed data. Its main contents include the parameters of the retention and diffusion of different air pollutants in the air at different wind speeds.

[0080] For the external dust problem, the main problem is the proportion of PM2.5 and PM10 that fall on seats, clothes and other attachments after entering the test compartment, as well as the problem of diffusion in the air. For the internal nitrogen and CO2 problem, the method of random emission within the test time is adopted according to the average emission value of normal people. This problem can be solved through experiments and algorithms. The main experimental process is:

[0081] Step 1. The experiment is designed according to the requirements, with pollutants input on one side of the experimental carriage and discharged on the other side, and pollutants input on both sides. According to the average value of the above data collected by the collection end of the system, the simulation central control issues data instructions to the two haze gas dust generators, which include the mass of PM2.5 and PM10, and the wind speed at the inlet. Then the two haze gas dust generators execute the corresponding input PM2.5 and PM10 mass, as well as the wind speed at the inlet according to the instruction requirements.

[0082] Step 2. The simulation central control continuously collects data from 12 simulated haze gas dust sensors and simulated CO2 sensors inside the experimental compartment, and establishes a data diffusion and distribution model with time. The data set for model learning includes:

[0083] The input data are: all temperatures, all humidity, all wind speeds, all wind directions, time, GPS, 2 simulated haze gas dust sensors at the entrance and exit, 2 simulated CO2 sensors at the entrance and exit, 1 simulated formaldehyde detector, number of people (simulated infrared sensor plus infrared camera and ordinary camera to analyze the number of people in the aisle). The emissions of nitrogen generator and CO2 generator are randomly generated, and the emission of a single one is 500 ml / day, which is set according to the average value of normal people.

[0084] The output data are: the values ​​of 10 simulated haze gas dust sensors and 10 simulated CO2 sensors in the car when no low-temperature plasma purifier, air conditioning or other air purification devices are turned on.

[0085] The structure of the pollution distribution data prediction model is LRCN network module, hole sampling data preprocessing module, smoothing prediction module and fusion output module.

[0086] The LRCN network module is a standard LRCN network with a structure of 5 layers, each layer has 1024 standard LSTMs, which are used for basic learning data sets and predicting output data based on input data.

[0087] The role of the void sampling data preprocessing module is that there is the following mathematical principle: "If the structure of the LRCN network can approximately cover the required working time, then the accuracy is the highest", that is, 1024 data just cover the required working time. Here, the working time is the time when the haze gas dust emitted by the haze gas dust generator at the entrance and exit is blown into the interior of the carriage by the wind, and then the entrance and exit doors are closed. The haze gas dust is stably floating inside the experimental carriage. This time is assumed to be T1. The present invention obtains the void sampling rate according to the following formula:

[0088] Where Sam is the uniform sampling time of the analog sensor.

[0089] In this embodiment, Sam is 50 ms and T1 is 10 minutes, so Hs is approximately equal to 10.

[0090] The smooth prediction module is because the preprocessing of the hole sampling data only uses 1 / Hs data. The present invention smoothes the data, that is, using the same LRCN network and a hole sampling data preprocessing structure, each prediction uses the data of the dint(ih, Hs+ih, ..., nh*Hs+ih)th of the total data set input part Dint to perform prediction, and obtain Hs prediction results. Among them, ih is the smoothing number, from 1 to Hs-1. nh is dow(Dt / Hs)-1, dow is rounded down. In this way, the first group of data is dt(1, Hs+1, ..., nh*Hs+1), and the second group of data is dt(2, Hs+2, ..., nh*Hs+2), and so on to obtain Hs groups of input data, and then obtain the output Dfoutt(hsout) of the Hs group prediction, hsout from 1 to HS.

[0091] Finally, the fusion output module adds all Dfoutt(hsout) and takes the average value. Figure 2 As shown, hsout ranges from 1 to 10 at this moment.

[0092] In this way, through this algorithm, we can know where the dust is concentrated and where the density is highest under different dust inputs, different wind speeds and directions, and different door opening and closing times, so as to know the further control plan of using low-temperature plasma purifiers and air conditioners, providing an ideal environment for the research of its air quality measurement and control equipment.

[0093] The above is only a preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of this template.

Claims

1. A digital twin system for analyzing the distribution of air pollution in transportation tracks, characterized in that: Including collection end, simulation end, and pollution distribution data prediction model; The collection end includes: collection wind speed sensor, collection wind direction sensor, collection formaldehyde detector, collection haze gas dust sensor, collection CO2 sensor, collection temperature and humidity sensor, collection infrared sensor, contact sensor, infrared camera, ordinary camera, collection central control, GPS; The analog end includes: analog wind speed sensor, analog wind direction sensor, analog formaldehyde detector, analog haze gas dust sensor, analog CO2 sensor, analog temperature and humidity sensor, analog infrared sensor, analog central control, haze gas dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier, air conditioner; The collection end is arranged in the collection carriage, and the simulation end is arranged in the experimental carriage; The wind speed sensor, wind direction sensor, formaldehyde detector, haze gas and dust sensor, CO2 sensor, temperature and humidity sensor, infrared sensor, contact sensor, infrared camera, ordinary camera and GPS are marked as collection sensors. The collection sensors are used to collect the internal data of the collection compartment and send the internal data to the collection central control. The collection central control sends the internal data to the simulation central control through the network. The simulation central control controls the operation of the haze gas and dust generator, nitrogen generator, CO2 generator, low-temperature plasma purifier and air conditioner. The pollution distribution data prediction model includes LRCN network module, hole sampling data preprocessing module, smoothing prediction module, and fusion output module; The LRCN network module has a 5-layer structure, each layer includes 1024 standard LSTMs, which are used for basic learning data sets and predicting output data based on input data; The hole sampling data preprocessing module is used to cover the required working time with 1024 data. The working time is the time when the haze gas and dust emitted by the haze gas and dust generator is blown into the experimental compartment by the wind, and then the exit door and entrance door inside the experimental compartment are closed. The haze gas and dust float stably inside the experimental compartment. The working time is marked as T1. Using the formula The hole sampling rate HS is obtained, and the simulated wind speed sensor, simulated wind direction sensor, simulated formaldehyde detector, simulated haze gas dust sensor, simulated CO2 sensor, simulated temperature and humidity sensor, and simulated infrared sensor are marked as simulated sensors; where Sam is the unified sampling time of the simulated sensors; The smoothing prediction module uses the data of the dint(ih, Hs+ih, ..., nh*Hs+ih)th of the total data set input part Dint for prediction each time, and obtains Hs prediction results; among them, ih is the smoothing number, from 1 to Hs-1; nh is dow(Dt / Hs)-1, dow is rounded down; the first group of data is dint(1, Hs+1, ..., nh*Hs+1), the second group of data is dint(2, Hs+2, ..., nh*Hs+2), and so on to obtain Hs groups of input data, and obtain the output Dfoutt(hsout) of the Hs group prediction, hsout from 1 to HS; The fusion output module is used to add all Dfoutt(hsout) and take the average value.

2. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: Wind speed sensors and wind direction sensors are installed at the two entrances of the collection compartment and at the upper middle position. There are three groups in total, which are used to obtain wind speed and wind direction data. The temperature and humidity sensors are installed together with the wind speed sensors and wind direction sensors.

3. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: The formaldehyde detectors include two groups, one of which is placed above the middle of the collection carriage to obtain formaldehyde data, and the other is installed in the toilet of the collection carriage.

4. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: There are thirteen groups of sensors for collecting haze, gas and dust. Two groups are installed at the two entrances of the collection carriage, ten groups are installed at equal distances on both sides of the collection carriage, and one group is installed in the toilet of the collection carriage. The CO2 sensor and the haze, gas and dust sensor are installed together.

5. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: The collection infrared sensor is installed above each seat in the collection carriage to analyze whether there is someone in the seat. The contact sensor is installed on the two entrance doors of the collection carriage to analyze whether the entrance doors are closed. A group of infrared cameras and ordinary cameras are used to analyze the actual number of people in the collection carriage. GPS is used to obtain the location of the collection end.

6. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: The acquisition central control is used to obtain the data of the acquisition sensor, analyze whether the data is normal, save the data after adding a time tag, and then send the data to the simulation central control through the network according to the instructions of the simulation central control. The specific steps of analyzing whether the data is normal are as follows: Determine whether the acquisition sensor obtains data according to the normal time interval FT(i). If it exceeds 3 times FT(i), it is judged to be disconnected, where i is the number of the acquisition sensor; Determine whether the data exceeds the set threshold range.

7. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: Haze gas and dust generators are installed at the two entrances and exits of the experimental carriage. Under the control of the simulated central control, they spread PM2.5 and PM10 air pollutants into the air.

8. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: Nitrogen generator and CO2 generator are used to simulate passengers' discharge of abdominal gas.

9. A digital twin system for analyzing air pollution distribution in a transportation track according to claim 1, characterized in that: Simulated wind speed sensors and simulated wind direction sensors are installed at the two entrances of the experimental carriage and at the upper middle position, with a total of 3 groups. Simulated temperature and humidity sensors are installed together with simulated wind speed sensors and simulated wind direction sensors. The simulated formaldehyde detector includes two groups, one of which is placed at the upper middle position of the experimental carriage to obtain formaldehyde data, and the other group is installed in the bathroom. There are thirteen groups of simulated haze gas and dust sensors, two groups are installed at the two entrances of the experimental carriage, ten groups are installed at equal distances on both sides of the carriage, and one group is installed in the bathroom. Simulated CO2 sensors are installed together with simulated haze gas and dust sensors, and simulated infrared sensors are installed above each seat.

Images