Comprehensive air sterilization and purification method in dynamic air duct

Abstract

The invention discloses a comprehensive air sterilization and purification method in a dynamic air duct, and the method comprises the steps: arranging a sensor array in a first flow channel, sampling and monitoring air flowing through the first flow channel, collecting the concentration value of pollutants in the air, and building an air quality prediction model through a machine learning method. An air quality index prediction value is output through pollutant concentration detection, and whether the air quality reaches the standard or not is judged. Activated carbon purification and ultraviolet sterilization methods are comprehensively applied, and the air purification rate is increased; an air quality prediction model is established by using a decision tree algorithm, air quality data in an air duct can be collected in real time and processed and predicted, dynamic sterilization and purification are realized, and the method has the characteristics of short time consumption, high efficiency, high accuracy and the like.

Description

Technical Field

[0001] The invention relates to an air sterilization and purification method, and in particular to an air dynamic sterilization and purification method which is applicable to an air duct and integrates multiple sterilization methods. Background Art

[0002] Air sterilization and purification devices mainly refer to the use of a series of technical means to remove various pollutants in the air. At present, the air purification methods used by factories mostly use fresh air systems to filter and ventilate the air at the primary level, and use activated carbon plates for adsorption. When the air quality requirements are high, additional separate UV sterilization, plasma sterilization and other methods are introduced. The overall sterilization and purification capacity is limited. During the purification process, the carrying capacity of adsorption and filtration components such as activated carbon is limited. If a single purification method is used, the purification may not be thorough due to the limited purification capacity at one time, and the monitoring requirements are high and the fault tolerance rate is low.

[0003] Fresh air system is an air treatment system, the main function of which is to realize the circulation exchange of indoor and outdoor air to improve indoor air quality. Fresh air system can be divided into ducted fresh air system (central fresh air system) and ductless fresh air system (wall-mounted fresh air system). Ducted fresh air system mechanically forces ventilation through equipment such as air ducts and blowers, which is suitable for industrial and large-area installation and has a wider application. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an air sterilization and purification method suitable for a ducted fresh air system with better sterilization and purification effect.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A comprehensive air sterilization and purification method in a dynamic air duct of the present invention comprises the following steps:

[0007] Step 1: Install a control system in a closed space at the intersection of the cross air ducts, the control system comprising:

[0008] A first valve and a second valve are respectively installed between the inner wall of the longitudinal air duct near the air inlet of the longitudinal air duct and the inner wall of the transverse air duct, and between the inner wall of the longitudinal air duct near the air outlet of the longitudinal air duct and the inner wall of the transverse air duct to divide the longitudinal air duct into a first flow channel and a second flow channel, and a one-way valve is installed in the second flow channel so that air can only flow in the direction from the second valve to the first valve; a baffle is installed in the longitudinal air duct near the air outlet of the longitudinal air duct to separate the first flow channel near the air outlet of the longitudinal air duct into two independent areas, namely the third flow channel and the fourth flow channel, and along the air flow direction, a third valve, a silver ion photocatalyst filter, and a fourth valve are installed in sequence on the third flow channel, and a fifth valve is installed at the inlet of the fourth flow channel, and the direction from the inlet of the first flow channel to the first flow outlet is , a centrifugal blower, an activated carbon filter, and an ultraviolet LED are sequentially installed in the first flow channel, an ultrasonic wind speed sensor is installed in the first flow channel just in front of the air outlet of the centrifugal blower, a wind speed control module is installed in the centrifugal blower, and the wind speed control module is composed of a controller and a frequency converter, the signal input end of the controller is connected to the ultrasonic wind speed sensor, the signal output end of the controller is connected to the signal input end of the frequency converter, the signal output end of the frequency converter is connected to the power supply at the stator end of the motor of the centrifugal blower, the frequency of the input power supply at the stator end of the motor of the centrifugal blower is changed by the controller to adjust the frequency, thereby adjusting the speed of the motor rotor, and then adjusting the outlet wind speed of the centrifugal blower, and the wind speed control module is connected to a computer;

[0009] Step 2: Arrange an array of SO2, NO2, O3, CO, PM2.5, and PM10 sensors in the first flow channel, sample and monitor the air flowing through the first flow channel, collect the concentration values ​​of pollutants in the air, and form a training data set sample;

[0010] Step 3: Use machine learning methods to establish an air quality prediction model, and then use the same method as step 2 to collect pollutant concentration values ​​in the air in the first flow channel in real time, and input them into the air quality prediction model. The output value obtained is the predicted value of the air quality index AQI;

[0011] Step 4: Start the centrifugal blower, turn on the computer, and input the given centrifugal blower outlet wind speed v n , ultrasonic wind speed sensor measures the outlet wind speed v of centrifugal blower in real time w , the computer performs the judgment: If v w >v n , the computer outputs a control signal to the controller of the wind speed control module, and the controller reduces the speed of the motor rotor by controlling the inverter until If v w <v n, the computer outputs a control signal to the controller of the wind speed control module. The controller controls the frequency converter to increase the rotational speed of the motor rotor until If v w = v n , no adjustment is required;

[0012] Step Five: Input the concentrations of pollutants such as SO2, NO2, O3, CO, PM2.5, and PM10 detected by the SO2, NO2, O3, CO, PM2.5, PM10 sensor array into the trained air quality prediction model, and output the predicted value PDT of the air quality index;

[0013] Step Six: Set the qualified standard value of the air quality as a, and compare the predicted value PDT of the air quality index with the qualified standard value a. If PDT > a, the air quality does not meet the standard, the third valve opens, the air is further purified through the silver ion photocatalyst filter, and then Step Seven is executed; if PDT < a, the air quality meets the standard, the fifth valve opens, and then Step Eight is executed;

[0014] Step Seven: Detect again whether the air quality after being purified by the silver ion photocatalyst filter meets the standard in the manner of Step Five and Step Six. If the air quality does not meet the standard, open the second valve and the first valve, and the air returns to the entrance of the first flow channel along the second flow channel, and then Step Eight is executed; if the air quality meets the standard, Step Nine is executed;

[0015] Step Eight: Repeatedly execute the computer judgment in Step Four, and then execute Step Five, Step Six. During the execution of Step Seven, if it is judged that the air quality meets the standard, close the second valve and the first valve, and then execute Step Nine;

[0016] Step Nine: The air is discharged from the longitudinal air duct outlet; repeatedly execute the computer judgment in Step Four, and then execute Steps Five to Seven.

[0017] The beneficial effects of the present invention are as follows:

[0018] The present invention comprehensively applies the methods of activated carbon purification and ultraviolet sterilization, improving the air purification rate; uses the decision tree algorithm to establish an air quality prediction model, which can collect and process the air quality data in the air duct in real time for prediction, realizing dynamic sterilization and purification, and having the characteristics of short time, high efficiency, high accuracy, etc. Description of the Drawings

[0019] Figure 1 is a flowchart of a method for comprehensive air sterilization and purification in a dynamic air duct according to an embodiment of the present invention;

[0020] Figure 2 is a flowchart of establishing an air quality prediction model according to an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of the layout of the air sterilization and purification device in the air duct. DETAILED DESCRIPTION

[0022] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] As shown in the accompanying drawings, a comprehensive air sterilization and purification method in a dynamic air duct of the present invention comprises the following steps:

[0024] Step 1: Install a control system in the closed space at the intersection of the cross-crossed air ducts, the control system comprising: installing a first valve 7-1 and a second valve 7-2 respectively between the inner wall of the longitudinal air duct and the inner wall of the transverse air duct near the air inlet of the longitudinal air duct, and between the inner wall of the longitudinal air duct and the inner wall of the transverse air duct near the air outlet of the longitudinal air duct to divide the longitudinal air duct into a first flow channel and a second flow channel, a one-way valve 9 is installed in the second flow channel to ensure that the air can only flow in the direction from the second valve 7-2 to the first valve 7-1; a baffle is installed in the longitudinal air duct near the air outlet of the longitudinal air duct to separate the first flow channel near the air outlet of the longitudinal air duct into two independent areas, namely the third flow channel and the fourth flow channel. Along the air flow direction, the third valve 4, the silver ion photocatalyst filter 6, and the fourth valve 8 are sequentially installed on the third flow channel, and the fifth valve 5 is installed at the inlet of the fourth flow channel. From the inlet of the first flow channel to the first flow outlet (i.e., the inlet of the third valve and the fifth valve), the centrifugal blower 1, the activated carbon filter 2, and the ultraviolet LED 3 are sequentially installed in the first flow channel, and an ultrasonic wind speed sensor is installed in the first flow channel just in front of the air outlet of the centrifugal blower 1. A wind speed control module is installed in the centrifugal blower, and the wind speed control module is composed of a controller and a frequency converter. The signal input end of the controller is connected to the ultrasonic wind speed sensor, the signal output end of the controller is connected to the signal input end of the frequency converter, and the signal output end of the frequency converter is connected to the power supply of the motor stator end of the centrifugal blower. The frequency converter is controlled by the controller to adjust the frequency, and the frequency of the input power supply of the motor stator end of the centrifugal blower is changed, thereby adjusting the speed of the motor rotor, and then adjusting the outlet wind speed of the centrifugal blower. The wind speed control module is connected to a computer.

[0025] Step 2: Arrange an array of SO2, NO2, O3, CO, PM2.5, and PM10 sensors in the first flow channel, sample and monitor the air flowing through the first flow channel, collect the concentration values ​​of pollutants in the air, and form a training data set sample;

[0026] Step 3: Use machine learning methods (reference: Zhao Yiwei. Research on air quality prediction model based on machine learning [D]. Anhui University of Science and Technology, 2024.) to establish an air quality prediction model. Then, use the same method as step 2 to collect the pollutant concentration value in the air in the first flow channel in real time, and input it into the air quality prediction model. The output value obtained is the predicted value of the air quality index AQI, as follows:

[0027] Step 301, reference the network open source training data set, the number of samples is N, and the network open source data set expression is TDS = {(x1, y1), (x2, y 2) …(x i ,y i )...(x N ,y N )}, where array x i ={ρ iSO2 ,ρ iNO2 ,ρ iO3 ,ρ iCO ,ρ iPM2.5 ,ρ iPM10}, ρ iSO2 ,ρ iNO2 ,ρ iO3 ,ρ iCO ,ρ iPM2.5 ,ρ iPM10 Respectively represent the concentrations of SO2, NO2, O3, CO, PM2.5, and PM10 pollutants, y1, y2, …y N is the air quality index AQI corresponding to the array. The network open source dataset is divided into a training set and a test set in a ratio of 7:3.

[0028] Step 302: Use the training set to train the decision tree model in the machine learning method to obtain a trained decision tree model.

[0029] The principle of the decision tree model is to improve the stability and accuracy of the model by constructing multiple tree structures, each internal node represents a feature or attribute, each branch represents a value of this feature or attribute, and each leaf node represents a classification or regression result. Among them, the construction of each decision tree is independent, and the data set is divided into multiple subsets through the decision tree. Each subset contains data with the same features or attributes to increase diversity. Finally, each subset is analyzed to obtain the average value of all decision tree predictions.

[0030] Step 303: x in the test set i Input the trained decision tree model for testing to obtain the model prediction value of the air quality index; use the formula Evaluate the model predictions of the air quality index, where y iis the value of the air quality index of the i-th sample data in the test set, is the model prediction value of the air quality index in the i-th sample data in the test set, and n is the number of samples in the test set. If the MSE is not greater than 5%, the trained decision tree model is determined to be qualified and used as the final air quality prediction model; otherwise, it is unqualified, and step 302 is repeated.

[0031] Step Four: Start the centrifugal blower, turn on the computer, and input the given outlet wind speed v of the centrifugal blower n , and the ultrasonic wind speed sensor measures the outlet wind speed v of the centrifugal blower in real time w , and the computer makes a judgment: If v w >v n , the computer outputs a control signal to the controller of the wind speed control module, and the controller reduces the rotational speed of the motor rotor by controlling the frequency converter until If v w <v n , the computer outputs a control signal to the controller of the wind speed control module, and the controller increases the rotational speed of the motor rotor by controlling the frequency converter until If v w =v n , no adjustment is required.

[0032] Step Five: Input the concentrations of SO2, NO2, O3, CO, PM2.5, and PM10 pollutants detected by the SO2, NO2, O3, CO, PM2.5, and PM10 sensor arrays into the trained air quality prediction model, and output the air quality index prediction value PDT;

[0033] Step Six: Refer to GB / T 18883 "Indoor Air Quality Standard" to set the air quality pass standard value as a, and compare the air quality index prediction value PDT with the pass standard value a. If PDT > a, the air quality fails to meet the standard, the third valve 4 opens, the air is further purified through the silver ion photocatalyst filter screen 6, and then step seven is executed; if PDT < a, the air quality meets the standard, the fifth valve 5 opens, and then step eight is executed;

[0034] Step Seven: Detect again whether the air quality after purification by the silver ion photocatalyst filter screen meets the standard in the manner of step five and step six. If the air quality fails to meet the standard, open the second valve and the first valve, and the air returns to the inlet of the first flow channel along the second flow channel, and then step eight is executed; if the air quality meets the standard, step nine is executed;

[0035] Step Eight: Repeatedly execute the computer judgment in step four, and then execute step five and step six. During the execution of step seven, if it is determined that the air quality meets the standard, close the second valve and the first valve, and then execute step nine.

[0036] Step 9. The air is discharged from the longitudinal air duct outlet; repeat the computer judgment of step 4, and then execute steps 5 to 7.

Claims

1. A comprehensive air sterilization and purification method in a dynamic air duct, characterized in that: The following steps are involved: Step 1: Install a control system in a closed space at the intersection of the cross air ducts, the control system comprising: A first valve (7-1) and a second valve (7-2) are respectively installed between the inner wall of the longitudinal air duct near the air inlet of the longitudinal air duct and the inner wall of the transverse air duct, and between the inner wall of the longitudinal air duct near the air outlet of the longitudinal air duct and the inner wall of the transverse air duct to divide the longitudinal air duct into a first flow channel and a second flow channel, and a one-way valve (9) is installed in the second flow channel so that air can only flow in the direction from the second valve (7-2) to the first valve (7-1); a baffle is installed in the longitudinal air duct near the air outlet of the longitudinal air duct to divide the first flow channel near the air outlet of the longitudinal air duct into two independent areas, namely a third flow channel and a fourth flow channel, and a third valve (4), a silver ion photocatalyst filter (6), and a fourth valve (8) are installed in sequence on the third flow channel along the air flow direction, and a fifth valve (5) is installed at the inlet of the fourth flow channel to separate air from the first flow channel. In the direction from the inlet to the first outlet, a centrifugal blower (1), an activated carbon filter (2), and an ultraviolet LED 3 are installed in sequence in the first outlet; an ultrasonic wind speed sensor is installed in the first outlet of the centrifugal blower (1); a wind speed control module is installed in the centrifugal blower; the wind speed control module consists of a controller and a frequency converter; the signal input end of the controller is connected to the ultrasonic wind speed sensor; the signal output end of the controller is connected to the signal input end of the frequency converter; the signal output end of the frequency converter is connected to the power supply at the stator end of the motor of the centrifugal blower; the frequency converter is controlled by the controller to adjust the frequency, the frequency of the input power supply at the stator end of the motor of the centrifugal blower is changed, thereby adjusting the speed of the motor rotor, and then adjusting the outlet wind speed of the centrifugal blower; the wind speed control module is connected to a computer; Step 2: Arrange an array of SO2, NO2, O3, CO, PM2.5, and PM10 sensors in the first flow channel, sample and monitor the air flowing through the first flow channel, collect the concentration values ​​of pollutants in the air, and form a training data set sample; Step 3: Use machine learning methods to establish an air quality prediction model, and then use the same method as step 2 to collect pollutant concentration values ​​in the air in the first flow channel in real time, and input them into the air quality prediction model. The output value obtained is the predicted value of the air quality index AQI; Step 4: Start the centrifugal blower, turn on the computer, and input the given centrifugal blower outlet wind speed v n , ultrasonic wind speed sensor measures the outlet wind speed v of centrifugal blower in real time w , the computer performs the judgment: If v w >v n , the computer outputs a control signal to the controller of the wind speed control module, and the controller reduces the speed of the motor rotor by controlling the inverter until If v w <v n , the computer outputs a control signal to the controller of the wind speed control module, and the controller increases the speed of the motor rotor by controlling the inverter until If v w =v n , no adjustment is required; Step 5: Input the concentrations of SO2, NO2, O3, CO, PM2.5, and PM10 pollutants detected by the SO2, NO2, O3, CO, PM2.5, and PM10 sensor arrays into the trained air quality prediction model, and output the predicted value of the air quality index PDT; Step 6: Set the qualified standard value of air quality as a, and compare the predicted value PDT of the air quality index with the qualified standard value a. If PDT > a, the air quality fails to meet the standard, and the third valve (4) is opened. The air is further purified by the silver ion photocatalyst filter (6), and then Step 7 is executed. If PDT < a, the air quality meets the standard, and the fifth valve (5) is opened, and then Step 8 is executed; Step 7: Detect again whether the air quality after being purified by the silver ion photocatalyst filter meets the standard in the manner of Step 5 and Step 6. If the air quality fails to meet the standard, open the second valve and the first valve, and the air returns to the inlet of the first flow channel along the second flow channel, and then Step 8 is executed. If the air quality meets the standard, Step 9 is executed; Step 8: Repeatedly execute the computer judgment in Step 4, and then execute Step 5 and Step 6. If it is judged that the air quality meets the standard during the execution of Step 7, close the second valve and the first valve, and then execute Step 9; Step 9: The air is discharged from the longitudinal air duct outlet; Repeatedly execute the computer judgment in Step 4, and then execute Steps 5 to 7.

2. The method for comprehensive air sterilization and purification in a dynamic air duct according to claim 1, characterized in that: The said Step 3 includes the following steps: Step 301, reference an open source training data set on the Internet, the number of samples is N, and the expression of the open source data set on the Internet is TDS = {(x1, y1), (x2, y2) ... (x i ,y i )...(x N ,y N )}, where array x i ={ρ iSO2 ,ρ iNO2 ,ρ iO3 ,ρ iCO ,ρ iPM2.5 ,ρ iPM10 }, ρ iSO2 ,ρ iNO2 ,ρ iO3 ,ρ iCO ,ρ iPM2.5 ,ρ iPM10 Respectively represent the concentrations of SO2, NO2, O3, CO, PM2.5, and PM10 pollutants, y1, y2, …y N is the air quality index AQI corresponding to the array; the network open source data set is divided into a training set and a test set in a ratio of 7:3; Step 302: Use the training set to train the decision tree model in the machine learning method to obtain the trained decision tree model; Step 303: x in the test set i Input the trained decision tree model for testing to obtain the model prediction value of the air quality index; use the formula Evaluate the model predictions of the air quality index, where y i is the value of the air quality index of the i-th sample data in the test set, is the model prediction value of the air quality index in the i-th sample data in the test set, n is the number of samples in the test set, if the MSE is not greater than 5%, the trained decision tree model is determined to be qualified and is used as the final air quality prediction model, otherwise it is unqualified and step 302 is repeated.

Images