A method and system for indoor radon detection

Abstract

The application provides an indoor radon detection method and system, and relates to the technical field of indoor radon detection. In view of the problems that a CZT detector is sensitive to temperature and humidity changes, continuous measurement has a residual memory effect, and the detection result is only a single concentration value and cannot guide the user to effectively reduce radon, the application adopts electrostatic enrichment and alpha energy spectrum analysis to obtain original measurement data, synchronously collects temperature, humidity and air pressure parameters, carries out one-stage temperature and humidity compensation on the original data through an artificial neural network model, solves the problem of count efficiency drift caused by temperature and humidity, carries out two-stage correction on intermediate concentration data through an iterative correction factor algorithm, eliminates the interference of the residual memory effect in the detection cavity on the current measurement, inversely evaluates the ventilation condition in combination with air pressure data and radon concentration change rate, and generates a comprehensive detection report. The application significantly improves the accuracy and environmental adaptability of radon concentration measurement, and reduces the lag error of continuous measurement.

Description

Technical Field

[0001] This invention relates to the field of indoor radon detection technology, and in particular to an indoor radon detection method and system. Background Technology

[0002] With the promotion of modern building energy-saving technologies and the continuous improvement of building airtightness, indoor radon gas is prone to accumulation and difficult to dilute, making indoor radon gas detection a crucial aspect of safeguarding public health. In recent years, with the rapid development of semiconductor detection technology, radon gas detection equipment based on cadmium zinc telluride (CZT) detectors has gradually become a research hotspot and development direction in the field of continuous indoor radon gas monitoring due to its advantages such as high energy resolution, room temperature operation, and compact size.

[0003] Currently, existing indoor radon detection methods have the following shortcomings: First, while traditional detection methods such as the activated carbon box method are low-cost, they cannot achieve real-time continuous monitoring, and the detection cycle is long and the operation is cumbersome. Second, although existing detection equipment based on CZT detectors can achieve alpha spectrum analysis, the CZT detectors themselves are extremely sensitive to temperature changes. Fluctuations in ambient temperature can lead to measurement errors of up to 40%, and humidity changes also affect the detector's counting efficiency. Existing technologies lack effective temperature and humidity compensation mechanisms. Third, during continuous measurement, radon progeny from the previous cycle can easily remain in the detection chamber, producing a memory effect that leads to an overestimation of the current measurement result, and existing methods do not effectively correct for this. Fourth, indoor radon concentration is affected by multiple environmental factors such as ventilation conditions and air pressure changes. Most existing detection methods only provide a single concentration value and fail to comprehensively consider the coupled effects of these factors, making it difficult to provide targeted ventilation recommendations.

[0004] Therefore, there is an urgent need for a more accurate indoor radon detection method. Summary of the Invention

[0005] To address these issues, this invention provides an indoor radon detection method and system to overcome the limitations of existing technologies, such as the inability to achieve real-time continuous monitoring, long detection cycles, cumbersome operation, lack of effective temperature and humidity compensation mechanisms, the tendency for residual radon progeny from the previous cycle to remain in the detection chamber, resulting in a memory effect and leading to higher current measurement results. Furthermore, indoor radon concentration is affected by multiple environmental factors, including ventilation conditions and air pressure changes, and most existing detection methods only provide a single concentration value, failing to comprehensively consider the coupling effect of these factors.

[0006] To achieve the above objectives, the present invention provides an indoor radon gas detection method, comprising: S1, radon progeny in indoor air is collected onto the surface of a CZT detector by electrostatic enrichment, and α-energy spectrum analysis is performed using the CZT detector to continuously acquire raw measurement data of radon concentration for a first predetermined period. S2, through the built-in environmental sensor, synchronously collects environmental parameter data including temperature, humidity and air pressure at the first predetermined cycle; S3, clean the original measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbance, and align the cleaned multi-source data on the time axis to generate a synchronous data frame sequence. S4, the synchronous data frame sequence is input into the pre-built and trained artificial neural network temperature and humidity correction model to perform a first-level correction on the original measurement data of the current period, so as to eliminate the influence of temperature and humidity changes on the counting efficiency of the CZT detector and generate intermediate concentration data compensated for temperature and humidity. S5. Based on the final concentration data of the previous measurement cycle and the intermediate concentration data, a second-level correction is performed using a preset iterative correction factor algorithm to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. S6. Based on the final concentration data from multiple consecutive periods, calculate the average radon concentration and confidence interval during the entire measurement period. Combine the synchronously collected air pressure data and the rate of change of radon concentration over time, couple the ventilation conditions and radon concentration during the measurement period, and finally generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations.

[0007] Furthermore, the process of collecting radon progeny from indoor air onto the surface of a CZT detector via electrostatic enrichment, and then performing alpha spectrum analysis using the CZT detector, includes: S11 introduces indoor air into the measurement chamber at a constant flow rate through a built-in air pump. Existing radon progeny and aerosol particles in the air are filtered out by a filter membrane, allowing only newly generated radon progeny to enter the detection area. S12, a high-voltage electrostatic field is applied in the measurement chamber, causing polonium produced by radon decay to migrate directionally under the action of the electric field force and be enriched and deposited on the sensitive surface of the CZT detector; S13 uses a CZT detector under high pressure to continuously detect alpha particles released during the decay of deposited polonium. The detected signal is converted into a voltage pulse by a charge-sensitive preamplifier, and after being shaped by the main amplifier, the alpha energy spectrum is collected by a multichannel analyzer. S14 identifies and extracts the characteristic peak count of polonium near the 6.00 MeV energy. Based on the net count rate of the characteristic peaks and the preset efficiency calibration coefficient, it calculates and outputs the raw measurement data of radon concentration in the current cycle.

[0008] Furthermore, the step of cleaning the original measurement data and the environmental parameter data, removing abnormal data points caused by instrument disturbances, and aligning the cleaned multi-source data on the time axis to generate a synchronized data frame sequence includes: S31, perform outlier detection on the collected raw measurement data and environmental parameter data respectively, identify and remove abrupt data points; S32 uses linear interpolation to repair and fill in the data gaps caused by the removal of mutation data points, ensuring the continuity and integrity of the data. S33 uses the sampling time point of the original measurement data as a reference, and aligns the repaired temperature, humidity and air pressure data to the same time coordinate system through a time matching algorithm to generate a synchronous data frame sequence containing the original radon concentration measurement value and its corresponding temperature, humidity and air pressure values ​​in each measurement cycle.

[0009] Furthermore, the first-level correction of the original measurement data for the current period includes: S41, extract the original measurement data, temperature data and humidity data of the current measurement cycle from the synchronous data frame sequence, and use them as the input feature vector of the artificial neural network temperature and humidity correction model; S42, the input feature vector is input into a pre-constructed and trained multi-layer feedforward neural network, the neural network comprising an input layer, a hidden layer and an output layer, wherein the hidden layer uses a non-linear activation function to map and transform the input features layer by layer; S43, through forward propagation calculation of neural network, obtains intermediate concentration data after temperature and humidity compensation in the output layer, so as to characterize the estimated value of radon concentration in the current period after eliminating the influence of temperature and humidity. S44, output and store the intermediate concentration data as input data for the next level of correction.

[0010] Furthermore, the secondary correction based on the final concentration data from the previous measurement period and the intermediate concentration data, using a preset iterative correction factor algorithm, includes: S51, read the final concentration data of the previous measurement cycle and obtain the intermediate concentration data after temperature and humidity compensation; S52, calculate the difference between the intermediate concentration data of the current period and the final concentration data of the previous measurement period, and determine the residual memory effect influence coefficient of the current period based on the difference. S53, Substitute the residual memory effect influence coefficient into the preset iterative correction factor algorithm model to perform secondary correction on the intermediate concentration data of the current period, so as to eliminate the measurement deviation caused by the accumulation of residual radon progeny in the detection chamber. S54 outputs the corrected result as the final concentration data for the current cycle and stores it as a historical basis for secondary correction in the next measurement cycle.

[0011] Furthermore, based on the final concentration data from multiple consecutive periods, the average radon concentration and confidence interval are calculated throughout the measurement period. Combined with synchronously acquired air pressure data and the rate of change of radon concentration over time, the ventilation conditions during the measurement period are coupled with the radon concentration to ultimately generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations, including: S61. Extract the radon concentration change curve over time from the final concentration data sequence of multiple consecutive measurement periods, calculate the radon concentration change rate between adjacent periods, and generate a radon concentration change rate time series. S62, synchronously acquire the air pressure data sequence within the corresponding time period, and determine the potential impact of air pressure on radon concentration based on the air pressure data sequence; S63, based on the time series of radon concentration change rate, identifies the period of continuous concentration decrease, and determines the ventilation period during the measurement by analyzing the rate of concentration decrease; S64, Couple the potential impact of the air pressure on radon concentration with the ventilation period to generate a ventilation recommendation that includes the number of ventilations and the ventilation period; S65, integrate the ventilation recommendations, radon concentration values, and environmental parameter change curves into a comprehensive test report for output.

[0012] Furthermore, the potential impact of the air pressure on radon concentration is coupled with the ventilation period to generate ventilation recommendations that include the number of ventilations and the ventilation period, including: S641, based on the ventilation periods during the determined measurement period, extract the start time, duration and corresponding radon concentration decrease rate of each ventilation period, and statistically obtain the user's ventilation habit characteristic parameters. The ventilation habit characteristic parameters include at least the distribution of typical ventilation periods, average ventilation duration and average ventilation air exchange rate. S642, Obtain meteorological forecast data for the next preset number of days, wherein the meteorological forecast data includes at least an hourly sequence of predicted air pressure values; S643, Based on the ventilation habit characteristic parameters, the air pressure prediction value sequence, and the potential impact of air pressure on radon concentration, generate an hourly variation curve of indoor radon concentration under the current ventilation conditions; S644, compare the hourly change curve of the indoor radon concentration with the preset safe concentration threshold to determine whether there is a risk of radon concentration exceeding the standard; S644a: If there is a risk of radon concentration exceeding the standard, then based on the air pressure prediction value sequence and user habits, the recommended ventilation period and recommended ventilation frequency that can effectively reduce radon concentration are determined through an optimization algorithm; return to S643, replace the ventilation habit feature parameters with the recommended ventilation period and recommended ventilation frequency, and re-execute S643 and S644 until it is determined that there is no risk of radon concentration exceeding the standard, and output the current recommended ventilation period and recommended ventilation frequency; S644b If there is no risk of radon concentration exceeding the standard, the current ventilation habit characteristic parameters will be output as ventilation recommendations.

[0013] Furthermore, the iterative correction factor algorithm model specifically includes: ; in, This is the final concentration data for the current period. This is the intermediate concentration data for the current period. The final concentration data for the previous measurement period is given, and k is the residual memory effect coefficient, ranging from 0.1 to 0.3.

[0014] Furthermore, the mutation data points are abnormal data points that exceed a preset threshold range due to instrument vibration, electromagnetic interference, or momentary power outages. The preset threshold range is the distribution interval of normal values ​​determined based on the statistical distribution of historical data.

[0015] The present invention also provides an indoor radon gas detection system, the system being used to implement any of the indoor radon gas detection methods described above, the system comprising: The raw data acquisition module is used to collect radon progeny in indoor air onto the surface of the CZT detector by electrostatic enrichment, and to perform alpha spectrum analysis using the CZT detector to continuously acquire raw measurement data of radon concentration for a first predetermined period. An environmental data acquisition module, connected to the original data acquisition module, is used to synchronously acquire environmental parameter data, including temperature, humidity and air pressure, through a built-in environmental sensor at the first predetermined cycle. The data preprocessing module, connected to the environmental data acquisition module, is used to clean the raw measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbances, and align the cleaned multi-source data on the time axis to generate a synchronous data frame sequence. The temperature and humidity compensation module, connected to the data preprocessing module, is used to input the synchronous data frame sequence into the pre-built and trained artificial neural network temperature and humidity correction model, and to perform a first-level correction on the original measurement data of the current period to eliminate the influence of temperature and humidity changes on the counting efficiency of the CZT detector, and generate intermediate concentration data after temperature and humidity compensation. The iterative correction module, connected to the temperature and humidity compensation module, is used to perform secondary correction based on the final concentration data of the previous measurement cycle and the intermediate concentration data, using a preset iterative correction factor algorithm, in order to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. The comprehensive report output module, connected to the iterative correction module, is used to calculate the average concentration and confidence interval of radon during the entire measurement period based on the final concentration data of multiple consecutive periods. It also combines the synchronously collected air pressure data and the rate of change of radon concentration over time to couple the ventilation status and radon concentration during the measurement period, and finally generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: Firstly, this invention inputs the synchronous data frame sequence into a pre-built and trained artificial neural network temperature and humidity correction model to perform a first-level correction on the original measurement data of the current cycle. It uses a multi-layer feedforward neural network to model and compensate for the nonlinear relationship between temperature, humidity and CZT detector counting efficiency. This solves the technical problems that CZT detectors are extremely sensitive to changes in temperature and humidity, that fluctuations in ambient temperature can lead to measurement errors of up to 40% or more, and that existing technologies lack effective temperature and humidity compensation mechanisms. This significantly improves the adaptability and accuracy of radon concentration measurement under different environmental conditions, enabling the measurement results to truly reflect the indoor radon concentration level. Secondly, this invention uses a preset iterative correction factor algorithm to perform secondary correction based on the final concentration data of the previous measurement cycle and the intermediate concentration data of the current cycle. It calculates the difference between the current cycle and the previous cycle and dynamically determines the residual memory effect influence coefficient. The residual interference from the previous cycle is subtracted from the intermediate concentration data. This solves the technical problem that radon progeny from the previous cycle is easily left in the detection chamber during continuous measurement, and the memory effect causes the current measurement result to be systematically high. It significantly reduces the hysteresis error and memory interference of the continuous measurement sequence, and improves the self-consistency of the measurement sequence and the tracking response capability to dynamic changes in radon concentration. Thirdly, this invention extracts the radon concentration variation curve over time from a series of final concentration data over multiple consecutive periods, calculates the radon concentration change rate and generates a change rate time series, simultaneously acquires air pressure data sequences and determines the potential impact of air pressure on radon concentration, identifies periods of continuous concentration decline and analyzes the rate of decline to determine ventilation periods, couples the influence of air pressure with ventilation periods to generate ventilation recommendations including ventilation frequency and ventilation periods, and finally generates a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations. This solves the technical problems of existing detection methods that mostly only provide a single concentration value, fail to comprehensively consider the coupling effect of multiple environmental factors such as ventilation conditions and air pressure changes, and are difficult to guide users to effectively reduce radon. It significantly improves the practical application value and user guidance of the test results, and provides a scientific basis for users to formulate precise radon reduction plans. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart of an indoor radon gas detection method provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of an indoor radon gas detection system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an indoor radon gas detection system provided in an embodiment of the present invention; The attached figures are labeled as follows: 1. Temperature sensor; 2. CZT detector; 3. Humidity sensor. Detailed Implementation

[0019] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0020] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0021] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0022] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0023] Example 1 like Figure 1 As shown, the present invention provides an indoor radon gas detection method, comprising: S1, radon progeny in indoor air is collected onto the surface of CZT detector 2 via electrostatic enrichment, and α-energy spectroscopy analysis is performed using CZT detector 2 to continuously acquire raw measurement data of radon concentration for a first predetermined period, including: S11 introduces indoor air into the measurement chamber at a constant flow rate through a built-in air pump. Existing radon progeny and aerosol particles in the air are filtered out by a filter membrane, allowing only newly generated radon progeny to enter the detection area. S12, a high-voltage electrostatic field is applied in the measurement chamber, causing polonium produced by radon decay to migrate directionally under the action of the electric field force and be enriched and deposited on the sensitive surface of CZT detector 2; S13, using the CZT detector 2 under high pressure to continuously detect alpha particles released during the decay of deposited polonium, the detected signal is converted into voltage pulses by a charge-sensitive preamplifier, and after being shaped by the main amplifier, the alpha energy spectrum is collected by a multichannel analyzer. S14 identifies and extracts the characteristic peak count of polonium near the 6.00 MeV energy. Based on the net count rate of the characteristic peaks and the preset efficiency calibration coefficient, it calculates and outputs the raw measurement data of radon concentration in the current cycle.

[0024] In one possible implementation, a built-in miniature diaphragm air pump with a rated voltage of 12V and a stable flow rate of 5L / min is activated. The air pump inlet is connected to a PTFE (polytetrafluoroethylene) pipe, introducing indoor air into the measurement chamber at a constant flow rate. A glass fiber filter membrane with a pore size of 0.45μm is installed between the air inlet and the measurement chamber to filter out existing radon progeny and aerosol particles in the air, ensuring that only newly generated radon progeny enters the detection area, thus avoiding interference from environmental aerosols on the measurement results. The filter membrane is replaceable.

[0025] The measurement chamber is made of stainless steel with smooth inner walls to reduce radon progeny adsorption. A ring-shaped high-voltage electrode is arranged inside the chamber, and a +1.5kV DC high voltage is applied to create a uniform electrostatic field. The sensitive surface of the CZT detector 2 is grounded and serves as the collection electrode. The first-generation progeny produced by radon decay carries a positive charge in the early stages of decay and migrates directionally under the influence of the electric field, accumulating and depositing efficiently on the sensitive surface of the CZT detector 2. The high-voltage power supply employs a modular design to ensure a constant electric field strength.

[0026] The CZT detector 2 has a crystal size of 5mm × 5mm × 5mm and operates at room temperature with a bias voltage of -800V. Polonium deposited on the detector surface undergoes alpha decay, releasing alpha particles with an energy of 6.00MeV. When these alpha particles are incident on the CZT crystal, they generate electron-hole pairs, which migrate under the influence of an electric field to form a charge signal. This signal is first fed into a charge-sensitive preamplifier, which converts the weak charge signal into a voltage pulse with a gain set to 100mV / MeV. The preamplifier output signal is then fed into the main amplifier, with a shaping time constant set to 0.5μs, for filtering and pulse shaping, outputting a quasi-Gaussian waveform. The shaped signal is then subjected to analog-to-digital conversion and energy spectrum acquisition by a multichannel analyzer with 1024 channels covering an energy range of 0-8MeV. The energy spectrum acquisition time for each measurement cycle is set to 1 hour, i.e., the first predetermined cycle is 1 hour. After acquisition, the alpha energy spectrum data file for the current cycle is generated.

[0027] The acquired alpha spectrum was smoothed using a three-point smoothing algorithm to reduce statistical fluctuations. Peak values ​​were searched within the 5.8-6.2 MeV energy range, identifying the 6.00 MeV characteristic peak of polonium. The net count of the characteristic peak was calculated using a linear background subtraction method: 10 channels were selected on each side of the characteristic peak as background intervals, a background straight line was fitted, and the background was subtracted from the total peak area to obtain the net peak area count. Based on a pre-calibrated efficiency scale coefficient, the raw radon concentration measurement data for the current period was calculated. In this embodiment, the effective volume of the measurement chamber was 0.5 L, the measurement period was t = 3600 s, and the net count of the characteristic peak measured in a certain period was 576. Therefore, the raw radon concentration measurement data was calculated to be 1000 Bq / m³.

[0028] This invention introduces indoor air into the measurement chamber at a constant flow rate using a built-in air pump. Existing radon progeny and aerosol particles are filtered out by a filter membrane, solving the background interference problem caused by existing radon progeny in the ambient air. Only newly generated radon progeny are allowed into the detection area, significantly improving the signal-to-noise ratio and the purity of the raw data. By continuously detecting alpha particles released during the decay of deposited polonium using a high-pressure CZT detector 2, and performing signal conversion, shaping, and energy spectrum acquisition via a charge-sensitive preamplifier, main amplifier, and multichannel analyzer, this invention solves the problem that traditional detectors cannot achieve high-energy-resolution alpha spectrum analysis at room temperature. This significantly improves the energy resolution capability for characteristic alpha particles, laying the foundation for accurate identification of radon progeny characteristic peaks. Furthermore, by identifying and extracting polonium at 6.00... The counting of characteristic peaks near MeV energy, based on the net count rate of the characteristic peaks and the preset efficiency calibration coefficient, calculates the radon concentration. This solves the problems of conventional measurement methods, such as the inability to distinguish different radionuclides and susceptibility to interference from other rays. It significantly reduces the background interference and nuclide crosstalk error of the measurement results, and realizes the specific quantitative detection of radon concentration.

[0029] S2, through the built-in environmental sensor, synchronously collects environmental parameter data including temperature, humidity and air pressure at the first predetermined cycle; S3, clean the original measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbances, and align the cleaned multi-source data on the time axis to generate a synchronized data frame sequence, including: S31, perform outlier detection on the collected raw measurement data and environmental parameter data respectively, identify and remove abrupt data points; S32 uses linear interpolation to repair and fill in the data gaps caused by the removal of mutation data points, ensuring the continuity and integrity of the data. S33 uses the sampling time point of the original measurement data as a reference, and aligns the repaired temperature, humidity and air pressure data to the same time coordinate system through a time matching algorithm to generate a synchronous data frame sequence containing the original radon concentration measurement value and its corresponding temperature, humidity and air pressure values ​​in each measurement cycle.

[0030] The abrupt change data point is an abnormal data point that exceeds the preset threshold range due to instrument vibration, electromagnetic interference or momentary power failure. The preset threshold range is the distribution range of normal values ​​determined based on the statistical distribution of historical data.

[0031] In one possible implementation, temperature sensor 1 uses an SHT35 digital temperature and humidity sensor with a temperature measurement range of -40℃ to 125℃, an accuracy of ±0.3℃, and a resolution of 0.01℃. Humidity sensor 3 also uses an SHT35 digital temperature and humidity sensor with a measurement range of 0% to 100%, a resolution of 0.01%, and a data update frequency of 1Hz. The barometric pressure sensor uses a BMP388 digital barometric pressure sensor with a measurement range of 300hPa to 1250hPa, an accuracy of ±0.1hPa, a resolution of 0.01hPa, and a data update frequency of 1Hz. The environmental sensors and the CZT detector 2 acquisition system share the same system clock to ensure consistent time references. Using a first predetermined period, i.e., 1 hour, as the time window, the raw data collected by the sensors every second is processed. At the end of each 1-hour period, the arithmetic mean of the temperature, humidity, and barometric pressure for that hour is calculated as the environmental parameter data for that period. For example, if 3600 temperature data points are collected in a certain period, their average value is taken as the temperature value for that period. The calculated periodic average temperature, average humidity, and average air pressure, along with the original radon concentration measurement data obtained during the same period, are stored together to form the original dataset.

[0032] Outlier detection is performed on both the raw measurement data and environmental parameter data. This embodiment employs a strategy combining the sliding window method and the threshold method. Specifically, for the raw radon concentration measurement data sequence, the sliding window width is set to 5 periods. For the current period i, the mean and standard deviation of the 5 data points within the window are calculated. If the difference between the current data point and the mean is greater than three times the standard deviation, the current data point is determined to be an anomalous data point and is removed. Simultaneously, combined with physical threshold judgment, if the current data point exceeds the normal physical range, such as showing a negative value or exceeding the upper limit of the instrument's measurement range, it is also determined to be an outlier and removed.

[0033] For temperature, humidity, and air pressure data sequences, the principle of three standard deviations is also used for anomaly detection. In addition, specific judgment rules are set for anomaly types specific to the sensors: if temperature data shows a sudden change rate exceeding 5°C / minute, exceeding the normal environmental change rate, it is judged as an anomaly caused by poor sensor contact or electromagnetic interference; if humidity data shows a sudden change exceeding 20% / minute, or remains unchanged at 100% for a long time, it is judged as an anomaly; if air pressure data shows a sudden change exceeding 5 hPa / minute, it is judged as an anomaly. All abnormal data points are marked and removed from the original sequence. For example, in a continuous 24-hour measurement that obtained 24 cycles of radon concentration data, a data point with a sudden change to 3500 Bq / m³ appeared in the 8th cycle, while the average value of the preceding and following cycles was approximately 800 Bq / m³. This point exceeded three standard deviations and was judged as an anomaly and removed.

[0034] For data gaps caused by the removal of abnormal data points, linear interpolation is used for repair. If the gap appears at the beginning of the sequence, it is filled with the next nearest neighbor value, i.e., the first valid data point is used to fill the preceding gap; if the gap appears at the end of the sequence, it is filled with the next nearest neighbor value, i.e., the last valid data point is used to fill the following gap. In this embodiment, after the radon concentration data of the aforementioned 8th period was removed, linear interpolation was performed using the 7th period (780 Bq / m³) and the 9th period (820 Bq / m³) to calculate the repaired value of the 8th period as 800 Bq / m³. At the same time, the temperature data of the 5th period was removed due to instantaneous sensor interference, and was repaired to 23.8℃ using interpolation using the 4th period (23.5℃) and the 6th period (24.1℃).

[0035] Using the original radon concentration measurement data of the current measurement cycle as a reference, the temperature, humidity, and air pressure data, after being repaired, correspond to a cycle average value for each environmental parameter within the same time interval. Since the environmental sensor and CZT detector 2 use the same system clock and both calculate the average value with a 1-hour cycle, the three naturally have a basis for time alignment. To ensure strict synchronization, this embodiment uses a timestamp matching algorithm to generate a unified time identifier for each measurement cycle. The end time of the cycle is taken as the timestamp, and the temperature, humidity, and air pressure data within the current cycle are combined with the timestamp to form a synchronization data frame for the current cycle.

[0036] Multiple consecutive periods of synchronized data frames constitute a synchronized data frame sequence, which serves as the input for S4. In this embodiment, 168 synchronized data frames are generated over 7 consecutive days of operation. Each frame contains a timestamp, period number, original radon concentration measurement, temperature, humidity, and air pressure values, ensuring strict alignment of multi-source data on the time axis and providing high-quality training and inference data for subsequent temperature and humidity compensation based on neural networks.

[0037] This invention solves the technical problem in traditional radon detection methods that fail to adequately consider the impact of environmental factors on measurement results, leading to inaccurate data and a lack of environmental correlation. This is achieved by using a built-in environmental sensor to synchronously collect environmental parameter data, including temperature, humidity, and air pressure, at a predetermined first cycle. This improves the accuracy and reliability of radon concentration measurement data under different environmental conditions, enabling the detection results to more realistically reflect indoor radon concentration levels and enhancing the accuracy and practicality of the results.

[0038] This invention addresses the technical problem of abnormal data mutations affecting the accuracy of subsequent data analysis and processing caused by unexpected events such as instrument vibration, electromagnetic interference, or momentary power outages. It involves detecting outliers in both raw measurement data and environmental parameter data, identifying and removing abrupt data points. This reduces the interference of abnormal data on overall measurement results and evaluation, improving the overall data quality. Furthermore, by employing linear interpolation to repair and fill in data gaps caused by the removal of abrupt data points, it resolves the issue of gaps in the data sequence caused by the removal of abnormal data, which disrupts data continuity and integrity, and consequently affects the assessment of radon levels. The technical problem of accurately analyzing radon concentration trends has been solved, improving the accuracy and reliability of data analysis. By aligning the repaired temperature, humidity, and air pressure data to the same time coordinate system using a time matching algorithm, synchronous data frames containing the original radon concentration measurement value and its corresponding temperature, humidity, and air pressure values ​​within each measurement cycle are generated. This solves the technical problem that the original measurement data and environmental parameter data cannot be accurately correlated and analyzed due to asynchronous acquisition time, making it difficult to effectively assess the impact of environmental factors on radon concentration measurement. This improves the scientificity and accuracy of the entire detection method and enhances the ability of the detection results to reflect the actual indoor environmental conditions.

[0039] S4, the synchronous data frame sequence is input into the pre-built and trained artificial neural network temperature and humidity correction model to perform a first-level correction on the original measurement data of the current period, so as to eliminate the influence of temperature and humidity changes on the counting efficiency of CZT detector 2 and generate intermediate concentration data compensated for temperature and humidity. The original measurement data for the current period undergoes a first-level correction, including: S41, extract the original measurement data, temperature data and humidity data of the current measurement cycle from the synchronous data frame sequence, and use them as the input feature vector of the artificial neural network temperature and humidity correction model; S42, the input feature vector is input into a pre-constructed and trained multi-layer feedforward neural network, the neural network comprising an input layer, a hidden layer and an output layer, wherein the hidden layer uses a non-linear activation function to map and transform the input features layer by layer; S43, through forward propagation calculation of neural network, obtains intermediate concentration data after temperature and humidity compensation in the output layer, so as to characterize the estimated value of radon concentration in the current period after eliminating the influence of temperature and humidity. S44, output and store the intermediate concentration data as input data for the next level of correction.

[0040] In one possible implementation, a pre-built and trained artificial neural network temperature and humidity correction model is used to perform a first-level correction on the raw measurement data of the current cycle, thereby eliminating the impact of temperature and humidity changes on the counting efficiency of the CZT detector 2. Specifically, after the detection device completes data acquisition for a measurement cycle, the data preprocessing module has cleaned and aligned the raw radon concentration measurement data, temperature data, humidity data, and air pressure data for that cycle, generating a synchronous data frame sequence. For example, in a synchronous data frame sequence, the raw radon concentration measurement value for the current measurement cycle is recorded as 50 Bq / m³, the temperature as 25℃, the humidity as 50%RH, and the air pressure as 1013 hPa. From the above synchronous data frame sequence, the raw measurement data, temperature data, and humidity data for the current measurement cycle are specifically extracted, and these three data are combined into a vector in a certain order as the input feature vector of the artificial neural network temperature and humidity correction model. For example, the input feature vector can be represented as [50, 25, 50].

[0041] The pre-built and trained artificial neural network temperature and humidity correction model is a multilayer feedforward neural network, specifically a fully connected multilayer perceptron (MLP), comprising an input layer, hidden layers, and an output layer. The input layer has three neurons, corresponding to the three dimensions of the input feature vector: the original radon concentration measurement, temperature, and humidity. The hidden layer has two layers, each with ten neurons. The hidden layers employ a non-linear activation function, such as ReLU, to map and transform the input features layer by layer, capturing the complex non-linear relationship between temperature / humidity and radon concentration measurements. The output layer has one neuron, used to output the intermediate concentration data after temperature and humidity compensation.

[0042] The extracted input feature vector [50, 25, 50] is input into the input layer of the artificial neural network temperature and humidity correction model. The neurons in the input layer pass the received data to the first hidden layer. Each neuron in the first hidden layer performs calculations and transformations using the ReLU function based on the input data and its own weights and biases, and then passes the results to the second hidden layer. The second hidden layer performs similar calculations and transformations, and finally passes the processed data to the output layer.

[0043] The neurons in the output layer calculate the intermediate radon concentration data after temperature and humidity compensation, based on the data transmitted from the second hidden layer and their own weights and biases. This data represents the estimated radon concentration for the current period after eliminating the influence of temperature and humidity. For example, after forward propagation calculation by the neural network, the output layer obtains a value of 48 Bq / m³, which is the intermediate concentration data after temperature and humidity compensation. This means that under the temperature (25℃) and humidity (50%RH) conditions of the current measurement period, after considering the influence of temperature and humidity on the counting efficiency of the CZT detector 2, the estimated radon concentration is 48 Bq / m³, which more accurately reflects the actual radon concentration compared to the original measured value of 50 Bq / m³.

[0044] The intermediate concentration data, after temperature and humidity compensation, obtained through neural network calculation, is output from the output layer of the neural network and stored in the storage medium of the detection device as input data for the next level of correction, so as to obtain a more accurate final measurement value of radon concentration for the current period.

[0045] This invention employs an artificial neural network temperature and humidity correction model to perform a primary correction on the raw measurement data for the current period, successfully eliminating the impact of temperature and humidity variations on the counting efficiency of the CZT detector and obtaining more accurate intermediate concentration data compensated for temperature and humidity. This process improves the accuracy and reliability of radon concentration measurement, providing a solid foundation for subsequent secondary corrections and the final generation of an accurate comprehensive test report. This allows indoor radon detection results to more accurately reflect the actual indoor radon concentration, providing strong data support for ensuring indoor air quality and human health.

[0046] S5. Based on the final concentration data of the previous measurement cycle and the intermediate concentration data, a second-level correction is performed using a preset iterative correction factor algorithm to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. Based on the final concentration data from the previous measurement period and the intermediate concentration data, a secondary correction is performed using a preset iterative correction factor algorithm, including: S51, read the final concentration data of the previous measurement cycle and obtain the intermediate concentration data after temperature and humidity compensation; S52, calculate the difference between the intermediate concentration data of the current period and the final concentration data of the previous measurement period, and determine the residual memory effect influence coefficient of the current period based on the difference. S53, Substitute the residual memory effect influence coefficient into the preset iterative correction factor algorithm model to perform secondary correction on the intermediate concentration data of the current period, so as to eliminate the measurement deviation caused by the accumulation of residual radon progeny in the detection chamber. S54 outputs the corrected result as the final concentration data for the current cycle and stores it as a historical basis for secondary correction in the next measurement cycle.

[0047] The iterative correction factor algorithm model specifically includes: ; in, This is the final concentration data for the current period. This is the intermediate concentration data for the current period. The final concentration data for the previous measurement period is given, and k is the residual memory effect coefficient, ranging from 0.1 to 0.3.

[0048] In one possible implementation, the data storage module of the detection device stores the final concentration data for each measurement cycle. The final concentration data for the previous measurement cycle is read from the storage module via the device's data read interface. Assume the final concentration data for the previous cycle is 850 Bq / m³. In this embodiment, the intermediate concentration data for the current cycle is 880 Bq / m³.

[0049] The difference between the intermediate concentration data of the current period and the final concentration data of the previous measurement period is calculated. In this embodiment, the difference is 30 Bq / m³. Based on this difference, the residual memory effect influence coefficient k of the current period is determined. The value of k is related to the physical characteristics of the probe cavity and is usually obtained through experimental calibration. In this embodiment, the influence coefficient is determined using a piecewise linear function: When the difference value is less than or equal to 0, it indicates that the current concentration is lower than or equal to the previous period, and the residual effect is small. Take k=0.1. When the difference value is less than or equal to 200 and greater than 0, take k=0.15; When the difference value is greater than 200, take k=0.2.

[0050] In this embodiment, k=0.15.

[0051] Substituting the residual memory effect coefficient *k* into the preset iterative correction factor algorithm model, a secondary correction is performed on the intermediate concentration data of the current cycle. The physical meaning of this formula is that the residual radon protons in the detection chamber mainly originate from the accumulation of the previous cycle, and their interference with the current cycle's measurement value is proportional to the concentration of the previous cycle. The interference amount is the product of *k* and the final concentration data of the previous measurement cycle. Therefore, subtracting this interference amount from the intermediate concentration data of the current cycle yields the true concentration after eliminating the residual memory effect. The calculated final concentration data for the current cycle is 752.5 Bq / m³. This invention solves the problem of the inability to quantify the accumulation state of residual radon progeny in the detection chamber during continuous measurement by reading the final concentration data of the previous measurement cycle and obtaining the intermediate concentration data of the current cycle after temperature and humidity compensation. This significantly improves the utilization rate of historical information and the reliability of the correction basis of the two-stage correction algorithm. By calculating the difference between the intermediate concentration data of the current cycle and the final concentration data of the previous measurement cycle, and dynamically determining the residual memory effect influence coefficient of the current cycle based on the difference, this invention solves the problem that a fixed correction coefficient cannot adapt to different concentration change scenarios, significantly improving the adaptive capability and scene matching accuracy of the correction algorithm. Substituting the response coefficient into the preset iterative correction factor algorithm model, a secondary correction is performed on the intermediate concentration data of the current period. This solves the problem of the current measurement result being systematically too high due to the accumulation of residual radon progeny in the detection chamber, significantly reduces the hysteresis effect and memory interference of the continuous measurement sequence, and makes the measurement result closer to the true concentration value. By outputting and storing the corrected result as the final concentration data of the current period, and using it as the historical basis for secondary correction in the next measurement period, the problem of the ineffective transmission of correction information in multi-period continuous measurement is solved. This significantly improves the self-consistency of the entire measurement sequence and the cumulative effect of the correction, and realizes accurate tracking of the dynamic change process of radon concentration.

[0052] S6. Based on the final concentration data from multiple consecutive periods, calculate the average radon concentration and confidence interval throughout the measurement period. Combined with synchronously acquired air pressure data and the rate of change of radon concentration over time, couple the ventilation conditions and radon concentration during the measurement period to generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations, including: S61. Extract the radon concentration change curve over time from the final concentration data sequence of multiple consecutive measurement periods, calculate the radon concentration change rate between adjacent periods, and generate a radon concentration change rate time series. S62, synchronously acquire the air pressure data sequence within the corresponding time period, and determine the potential impact of air pressure on radon concentration based on the air pressure data sequence; S63, based on the time series of radon concentration change rate, identifies the period of continuous concentration decrease, and determines the ventilation period during the measurement by analyzing the rate of concentration decrease; S64, Couple the potential impact of the air pressure on radon concentration with the ventilation period to generate a ventilation recommendation that includes the number of ventilations and the ventilation period; S65, integrate the ventilation recommendations, radon concentration values, and environmental parameter change curves into a comprehensive test report for output.

[0053] The potential impact of air pressure on radon concentration is coupled with the ventilation period to generate ventilation recommendations that include the number of ventilations and the ventilation period, including: S641, based on the ventilation periods during the determined measurement period, extract the start time, duration and corresponding radon concentration decrease rate of each ventilation period, and statistically obtain the user's ventilation habit characteristic parameters. The ventilation habit characteristic parameters include at least the distribution of typical ventilation periods, average ventilation duration and average ventilation air exchange rate. S642, Obtain meteorological forecast data for the next preset number of days, wherein the meteorological forecast data includes at least an hourly sequence of predicted air pressure values; S643, Based on the ventilation habit characteristic parameters, the air pressure prediction value sequence, and the potential impact of air pressure on radon concentration, generate an hourly variation curve of indoor radon concentration under the current ventilation conditions; S644, compare the hourly change curve of the indoor radon concentration with the preset safe concentration threshold to determine whether there is a risk of radon concentration exceeding the standard; S644a: If there is a risk of radon concentration exceeding the standard, then based on the air pressure prediction value sequence and user habits, the recommended ventilation period and recommended ventilation frequency that can effectively reduce radon concentration are determined through an optimization algorithm; return to S643, replace the ventilation habit feature parameters with the recommended ventilation period and recommended ventilation frequency, and re-execute S643 and S644 until it is determined that there is no risk of radon concentration exceeding the standard, and output the current recommended ventilation period and recommended ventilation frequency; S644b If there is no risk of radon concentration exceeding the standard, the current ventilation habit characteristic parameters will be output as ventilation recommendations.

[0054] In one possible implementation, the final concentration values ​​of multiple consecutive measurement cycles are extracted from the final concentration data sequence output in step S5. The radon concentration variation curve over time is constructed with the measurement cycle number as the x-axis and the final concentration as the y-axis, and the radon concentration variation rate between adjacent cycles is calculated.

[0055] Calculate the rate of change of air pressure between adjacent cycles. If the rate of change of air pressure for multiple consecutive cycles is less than... A pressure change rate of 0.5 hPa / h is considered a continuous depressurization process; if the rate of change of air pressure exceeds +0.5 hPa / h for multiple consecutive cycles, it is considered a continuous pressure increase process. The lowest air pressure value during the entire measurement period is recorded, and periods with air pressure below a certain threshold, such as 95% of the local average air pressure, are marked as low-pressure influence periods. Based on experience or regional geological data, an influence coefficient Q (unit: Bq / m³ / hPa) is set to represent the potential increase in radon concentration for every 1 hPa decrease. In this embodiment, Q = 5 Bq / m³ / hPa.

[0056] Based on the time series of radon concentration change rates, periods of continuous concentration decrease are identified. This embodiment uses the following criteria: The rate of change of radon concentration was less than [a certain value] for at least three consecutive periods. 10 Bq / m³ / h, meaning a decrease of more than 10 Bq / m³ per hour; The average rate of change of radon concentration over the period was negative, and the downward trend was significant.

[0057] A set of consecutive periods that meet the above conditions is labeled as a ventilation period. Each ventilation period records the start period, end period, duration, and average rate of descent during that period.

[0058] In this embodiment, based on the data from step S61: Periods 2-4: Radon concentration change rates were -18.3, -16.4, and -13.7, respectively, all less than -10. These three consecutive periods constituted the first ventilation period, lasting for 3 hours, with an average decrease rate of approximately -16.1 Bq / m³ / h.

[0059] Periods 25-27: Radon concentration change rates were -30.3, -27.6, and -22.3, respectively, all less than -10. These three consecutive periods constituted the second ventilation period, lasting for 3 hours, with an average decrease rate of approximately -26.7 Bq / m³ / h.

[0060] For each ventilation period, calculate the average pressure change rate within that period. If the average pressure change rate is less than -0.5, it indicates a significant pressure drop during that period, which enhances radon release, resulting in a measured concentration decrease rate lower than the actual ventilation effect. If the average pressure change rate is greater than 0.5, it indicates that pressure increase inhibits radon release, and the measured decrease rate may be too high. Therefore, the corrected actual ventilation decrease rate = average radon concentration change rate - Q × average pressure change rate.

[0061] In this embodiment: During the first ventilation period (cycles 2-4), the average rate of change in radon concentration was -16.1 Bq / m³ / h, and the average rate of change in air pressure during this period was -0.5 hPa / h. Therefore, the corrected actual rate of decrease in radon concentration during ventilation was -13.6 Bq / m³ / h, indicating that the actual rate of decrease in radon concentration caused by ventilation was approximately 13.6 Bq / m³ / h, which is lower than the measured value because the pressure drop increased radon release.

[0062] During the second ventilation period (cycles 25-27), the average rate of change of radon concentration was -26.7 Bq / m³ / h, and the average rate of change of air pressure during this period was -0.6 hPa / h. Therefore, the corrected actual ventilation radon concentration decrease rate was -23.7 Bq / m³ / h.

[0063] The number of air changes per minute can be roughly estimated based on the corrected rate of decrease. An indoor radon mass balance model is used: in the absence of new radon sources, the number of air changes per minute = the negative of the ratio of the corrected actual radon concentration decrease rate to the average radon concentration during the current period.

[0064] In this embodiment: The average concentration during the first ventilation period was 816.7 Bq / m³, resulting in an air exchange rate of 0.0167 / h (approximately 0.4 times / day), indicating a weak ventilation effect. The average concentration during the second ventilation period was 864.3 Bq / m³, resulting in an air exchange rate of 0.0274 / h (approximately 0.66 times / day), indicating a moderate ventilation effect.

[0065] Based on the multiple ventilation periods identified in the steps, the start time, duration, and corresponding radon concentration decrease rate of each ventilation period were extracted. For example, in a continuous monitoring period of 7 days, 5 effective ventilation periods were identified, occurring daily from 01:00-03:00 and 14:00-15:00. By statistically analyzing the data from these periods, the user's ventilation habit characteristics were obtained: typical ventilation periods are 01:00-03:00 and 14:00-15:00, with an average ventilation duration of 1.5 hours and an average air exchange rate of approximately 0.6 times / day.

[0066] Connecting to the meteorological service API via a built-in wireless communication module, it obtains hourly air pressure forecasts for the next 3 days (configurable as needed). For example, it can obtain hourly air pressure values ​​for the next 72 hours with a data accuracy of 0.1 hPa. Assume that the air pressure will continue to drop from night to early morning on the first day, with a decrease of 3 hPa; and that the air pressure will rise rapidly in the afternoon of the second day, with an increase of 2 hPa.

[0067] The principle of radon mass balance is used for hourly iterative calculations to predict the dynamic changes in indoor radon concentration over future periods. The specific calculation method is as follows: Initial conditions are set, using the final concentration data of the most recent measurement cycle as the radon concentration value C0 at the start of the calculation. In this embodiment, C0 is set to 180 Bq / m³. 3 ; Divide the future prediction period (e.g., 72 hours) into discrete time points t=1, 2, ..., N with a step size of 1 hour; For each time step, the radon concentration C at the next time step is calculated according to the following recursive formula. t+1 : ; Among them, C t+1 This represents the current radon concentration. R is the radon release rate (Bq / m³ / h), which can be determined by inverting historical data, for example, by estimating the radon concentration increase rate during periods without ventilation (such as when doors and windows are closed late at night); in this embodiment, R is calibrated to 30 Bq / m³ / h based on previous monitoring data. This is a correction function for the radon release rate based on gas pressure, reflecting the influence of gas pressure changes on the radon release rate; based on the gas pressure influence coefficient Q determined in step S62 and the deviation of the current gas pressure from the standard gas pressure. Calculate the correction factor: ;when When it is negative, A value greater than 1 indicates precipitation enhancement; when When it is a positive value, A value less than 1 indicates reduced precipitation; Let be the decay constant of radon, taken as 0.00756h. 1 ; The ventilation rate at time t is set according to the user's ventilation habits: during the user's typical ventilation periods (e.g., 01:00-03:00 and 14:00-15:00). Take the average number of air changes per hour (e.g., 0.6 h). -1 ); The outdoor radon concentration can be taken as a constant value, such as 5 Bq / m³.

[0068] Hourly iterative calculations yield the curve of indoor radon concentration changing over the next 72 hours. For example, calculations based on the above parameters show that on the first night of the future, due to a drop in air pressure, the concentration will be higher at this time. The value was negative, indicating increased radon release. Even with ventilation in the early morning, the radon concentration dropped to approximately 190 Bq / m³ at 02:00, before rapidly rising again after doors and windows were closed. The following afternoon, due to increased air pressure, radon release was suppressed, and the concentration remained at a low level. The overall curve reflects the periodic fluctuation of radon concentration, decreasing during ventilation periods and increasing during non-ventilation periods, while also incorporating the effects of air pressure changes.

[0069] The calculated hourly radon concentration variation curve is compared point-by-point with a preset safe concentration threshold, such as 200 Bq / m³. If the radon concentration exceeds this threshold at any given time, a risk of exceeding the limit is identified. In this embodiment, the simulation curve shows that the radon concentration remains above 200 Bq / m³ from 22:00 on the first day to 01:00 on the next day, reaching a maximum of 250 Bq / m³, thus indicating a risk of exceeding the limit.

[0070] If there is a risk of exceeding the standard, an optimization process will be initiated to find a ventilation solution that can eliminate the risk by adjusting the ventilation time and frequency. Specifically, the following methods can be used: Use the user's preferred ventilation time as the initial plan, and set an adjustable range (e.g., allow the ventilation time to be brought forward or delayed by 1-2 hours, and allow 1-2 additional ventilations). An enumeration or heuristic search algorithm (such as genetic algorithm or particle swarm optimization) is used to find a set of ventilation periods and ventilation times in the feasible solution space, so that the hourly radon concentration curve recalculated according to the S643 method does not exceed the safety threshold at all times, while being as close as possible to the user's original habits (e.g., minimal changes). For each candidate scheme, return to S643 to recalculate the radon concentration change curve and determine whether it meets the safety requirements.

[0071] In this embodiment, the algorithm iteratively searches and suggests advancing the early morning ventilation period to 00:00-02:00, and adding one ventilation session (lasting 1 hour) between 09:00-10:00 AM. Substituting this set of suggested ventilation periods and frequencies into S643 for recalculation, it is found that the radon concentration at all times is below 200 Bq / m³, with a maximum concentration of 185 Bq / m³, meeting safety requirements. The iteration then terminates, outputting the currently suggested ventilation periods (00:00-02:00, 09:00-10:00) and suggested ventilation frequency (2 times / day). If there is no risk of exceeding the standard in the initial calculation, the current ventilation habit characteristic parameters are directly output as the ventilation suggestion. In this embodiment, due to the risk of exceeding the standard, the optimized suggestion is output.

[0072] The above analysis results will be integrated into a comprehensive testing report. The report will include: Basic information, including testing time, location, testing cycle, and instrument information; Radon concentration statistics include the average radon concentration, standard deviation, and 95% confidence interval for the entire measurement period; Environmental parameter change curves, including radon concentration change curve over time, temperature change curve, humidity change curve, and air pressure change curve; Ventilation status assessment, including a list of identified ventilation periods, pressure impact analysis, and estimated air exchange rate; Ventilation recommendations will be provided based on the assessment results.

[0073] This invention addresses the problem that static statistical indicators cannot reflect the dynamic evolution of radon concentration by extracting the radon concentration variation curve over time from the final concentration data sequence of multiple consecutive measurement periods, calculating the radon concentration change rate between adjacent periods, and generating a change rate time series. This significantly improves the characterization ability and analytical dimensions of indoor radon temporal distribution characteristics. Furthermore, by simultaneously acquiring the air pressure data sequence within the corresponding time period and determining the potential impact of air pressure on radon concentration based on the air pressure data sequence, this invention solves the problem of neglecting the interference of meteorological factors on the radon release process, leading to inaccurate ventilation assessments, and significantly improves the environmental adaptability correction accuracy of ventilation effect evaluation. By identifying periods of continuous concentration decline based on the radon concentration change rate time series and analyzing the rate of concentration decline to determine the ventilation period during the measurement period, this invention solves the inconvenience and subjectivity issues of ventilation events not being automatically identified and relying on manual recording, significantly reducing the subjective error and operational complexity of ventilation assessments. Finally, by coupling the potential impact of air pressure on radon concentration with the ventilation period, this invention generates ventilation recommendations including the number of ventilations and the ventilation period, solving the problem of a lack of effective connection between detection results and actual radon reduction measures, and significantly improving the user guidance and practical value of the detection report.

[0074] This invention addresses the technical problem of existing detection methods that focus solely on radon concentration values ​​while neglecting user behavior pattern analysis by statistically analyzing ventilation periods identified in historical measurement cycles. This allows for the quantitative representation and personalized modeling of user ventilation habits. This process enables subsequent ventilation recommendations to be based on users' actual lifestyles, avoiding the pitfalls of vague suggestions that are disconnected from user habits and difficult to implement, significantly improving the adaptability of detection results to user living scenarios. Furthermore, by incorporating predicted air pressure sequences from future weather forecasts, it solves the technical problem of traditional radon detection relying solely on historical data and failing to predict the impact of future environmental changes on radon concentrations. As a key meteorological factor influencing radon release rates, air pressure provides environmental input conditions for dynamically predicting indoor radon concentration evolution by obtaining its changing trends in advance. This enables the detection system to possess forward-looking environmental perception capabilities, overcoming the limitations of static detection reports. By performing hourly iterative calculations based on the radon mass balance principle, user ventilation habit characteristics are coupled with future air pressure prediction sequences to generate an hourly indoor radon concentration change curve considering the influence of air pressure under current ventilation conditions. This solves the technical problem that existing technologies cannot quantitatively assess the coupling effect of user habits and meteorological factors. This calculation method dynamically simulates the accumulation and dissipation process of radon concentration through recursive formulas, and can intuitively present the fluctuation trend and peak level of radon concentration in future periods, providing a scientific basis for the quantitative assessment of exceedance risks. By comparing the simulated curve with safe concentration thresholds, and using optimization algorithms to iteratively adjust ventilation periods and frequencies when there is an exceedance risk, the optimal ventilation plan is output after meeting safety requirements. This solves the technical problems of existing ventilation recommendations lacking quantitative basis and effect verification, making it difficult to ensure the effectiveness of radon reduction. This closed-loop optimization mechanism can not only accurately identify periods of risk exceeding standards, but also automatically generate targeted ventilation strategies to ensure that the recommended ventilation plan can control future radon concentrations within safe thresholds. It realizes intelligent decision-making throughout the entire process from risk warning to intervention measures, significantly improving the practical value and user guidance of the test report.

[0075] Example 2 like Figure 2 As shown, the present invention also provides an indoor radon gas detection system, the system being used to implement the indoor radon gas detection method described in any of Embodiment 1, the system comprising: The raw data acquisition module is used to collect radon progeny in indoor air onto the surface of CZT detector 2 by electrostatic enrichment, and to use CZT detector 2 to perform α-energy spectrum analysis to continuously acquire raw measurement data of radon concentration for a first predetermined period. An environmental data acquisition module, connected to the original data acquisition module, is used to synchronously acquire environmental parameter data, including temperature, humidity and air pressure, through a built-in environmental sensor at the first predetermined cycle. The data preprocessing module, connected to the environmental data acquisition module, is used to clean the raw measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbances, and align the cleaned multi-source data on the time axis to generate a synchronous data frame sequence. The temperature and humidity compensation module, connected to the data preprocessing module, is used to input the synchronous data frame sequence into the pre-built and trained artificial neural network temperature and humidity correction model, and to perform a first-level correction on the original measurement data of the current period to eliminate the influence of temperature and humidity changes on the counting efficiency of CZT detector 2, and generate intermediate concentration data after temperature and humidity compensation. The iterative correction module, connected to the temperature and humidity compensation module, is used to perform secondary correction based on the final concentration data of the previous measurement cycle and the intermediate concentration data, using a preset iterative correction factor algorithm, in order to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. The comprehensive report output module, connected to the iterative correction module, is used to calculate the average concentration and confidence interval of radon during the entire measurement period based on the final concentration data of multiple consecutive periods. It also combines the synchronously collected air pressure data and the rate of change of radon concentration over time to couple the ventilation status and radon concentration during the measurement period, and finally generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations.

[0076] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for detecting radon gas indoors, characterized in that, include: S1, radon progeny in indoor air is collected onto the surface of a CZT detector by electrostatic enrichment, and α-energy spectrum analysis is performed using the CZT detector to continuously acquire raw measurement data of radon concentration for a first predetermined period. S2, through the built-in environmental sensor, synchronously collects environmental parameter data including temperature, humidity and air pressure at the first predetermined cycle; S3, clean the original measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbance, and align the cleaned multi-source data on the time axis to generate a synchronous data frame sequence. S4, the synchronization data frame sequence is input into the pre-built and trained artificial neural network temperature and humidity correction model to perform a first-level correction on the original measurement data of the current period, so as to eliminate the influence of temperature and humidity changes on the counting efficiency of the CZT detector and generate intermediate concentration data compensated for temperature and humidity. S5. Based on the final concentration data of the previous measurement cycle and the intermediate concentration data, a second-level correction is performed using a preset iterative correction factor algorithm to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. S6. Based on the final concentration data of multiple consecutive periods, calculate the average concentration and confidence interval of radon during the entire measurement period. Combine the synchronously collected air pressure data and the rate of change of radon concentration over time, couple the ventilation status and radon concentration during the measurement period, and finally generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations. The second-level correction, based on the final concentration data from the previous measurement period and the intermediate concentration data, is performed using a preset iterative correction factor algorithm, including: S51, read the final concentration data of the previous measurement cycle and obtain the intermediate concentration data after temperature and humidity compensation; S52, calculate the difference between the intermediate concentration data of the current period and the final concentration data of the previous measurement period, and determine the residual memory effect influence coefficient of the current period based on the difference. S53, Substitute the residual memory effect influence coefficient into the preset iterative correction factor algorithm model to perform secondary correction on the intermediate concentration data of the current period, so as to eliminate the measurement deviation caused by the accumulation of residual radon progeny in the detection chamber. S54 outputs the corrected result as the final concentration data for the current cycle and stores it as a historical basis for secondary correction in the next measurement cycle.

2. The indoor radon gas detection method according to claim 1, characterized in that, The process of collecting radon progeny from indoor air onto the surface of a CZT detector via electrostatic enrichment, and then performing alpha spectrum analysis using the CZT detector, includes: S11 introduces indoor air into the measurement chamber at a constant flow rate through a built-in air pump. Existing radon progeny and aerosol particles in the air are filtered out by a filter membrane, allowing only newly generated radon progeny to enter the detection area. S12, a high-voltage electrostatic field is applied in the measurement chamber, causing polonium produced by radon decay to migrate directionally under the action of the electric field force and be enriched and deposited on the sensitive surface of the CZT detector; S13 uses a CZT detector under high pressure to continuously detect alpha particles released during the decay of deposited polonium. The detected signal is converted into a voltage pulse by a charge-sensitive preamplifier, and after being shaped by the main amplifier, the alpha energy spectrum is collected by a multichannel analyzer. S14 identifies and extracts the characteristic peak count of polonium near the 6.00 MeV energy. Based on the net count rate of the characteristic peaks and the preset efficiency calibration coefficient, it calculates and outputs the raw measurement data of radon concentration in the current cycle.

3. The indoor radon gas detection method according to claim 1, characterized in that, The process of cleaning the original measurement data and environmental parameter data, removing abnormal data points caused by instrument disturbances, and aligning the cleaned multi-source data along the time axis to generate a synchronized data frame sequence includes: S31, perform outlier detection on the collected raw measurement data and environmental parameter data respectively, identify and remove abrupt data points; S32 uses linear interpolation to repair and fill in the data gaps caused by the removal of mutation data points, ensuring the continuity and integrity of the data. S33 uses the sampling time point of the original measurement data as a reference, and aligns the repaired temperature, humidity and air pressure data to the same time coordinate system through a time matching algorithm to generate a synchronous data frame sequence containing the original radon concentration measurement value and its corresponding temperature, humidity and air pressure values ​​in each measurement cycle.

4. The indoor radon gas detection method according to claim 1, characterized in that, The first-level correction of the original measurement data for the current period includes: S41, extract the original measurement data, temperature data and humidity data of the current measurement cycle from the synchronous data frame sequence, and use them as the input feature vector of the artificial neural network temperature and humidity correction model; S42, the input feature vector is input into a pre-constructed and trained multi-layer feedforward neural network, the neural network comprising an input layer, a hidden layer and an output layer, wherein the hidden layer uses a non-linear activation function to map and transform the input features layer by layer; S43, through forward propagation calculation of neural network, obtains intermediate concentration data after temperature and humidity compensation in the output layer, so as to characterize the estimated value of radon concentration in the current period after eliminating the influence of temperature and humidity. S44, output and store the intermediate concentration data as input data for the next level of correction.

5. The indoor radon gas detection method according to claim 1, characterized in that, The method calculates the average radon concentration and confidence interval throughout the measurement period based on the final concentration data from multiple consecutive periods. It then couples the ventilation conditions and radon concentration during the measurement period with simultaneously acquired air pressure data and the rate of change of radon concentration over time. Finally, it generates a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations, including: S61. Extract the radon concentration change curve over time from the final concentration data sequence of multiple consecutive measurement periods, calculate the radon concentration change rate between adjacent periods, and generate a radon concentration change rate time series. S62, synchronously acquire the air pressure data sequence within the corresponding time period, and determine the potential impact of air pressure on radon concentration based on the air pressure data sequence; S63, based on the time series of radon concentration change rate, identifies the period of continuous concentration decrease, and determines the ventilation period during the measurement by analyzing the rate of concentration decrease; S64, Couple the potential impact of the air pressure on radon concentration with the ventilation period to generate a ventilation recommendation that includes the number of ventilations and the ventilation period; S65, integrate the ventilation recommendations, radon concentration values, and environmental parameter change curves into a comprehensive test report for output.

6. The indoor radon gas detection method according to claim 5, characterized in that, The process of coupling the potential impact of air pressure on radon concentration with the ventilation period to generate ventilation recommendations including the number of ventilations and the ventilation period includes: S641, based on the ventilation periods during the determined measurement period, extract the start time, duration and corresponding radon concentration decrease rate of each ventilation period, and statistically obtain the user's ventilation habit characteristic parameters. The ventilation habit characteristic parameters include at least the distribution of typical ventilation periods, average ventilation duration and average ventilation air exchange rate. S642, Obtain meteorological forecast data for the next preset number of days, wherein the meteorological forecast data includes at least an hourly sequence of predicted air pressure values; S643, Based on the ventilation habit characteristic parameters, the air pressure prediction value sequence, and the potential impact of air pressure on radon concentration, generate an hourly variation curve of indoor radon concentration under the current ventilation conditions; S644, compare the hourly change curve of the indoor radon concentration with the preset safe concentration threshold to determine whether there is a risk of radon concentration exceeding the standard; S644a: If there is a risk of radon concentration exceeding the standard, then based on the air pressure prediction value sequence and user habits, the recommended ventilation period and recommended ventilation frequency that can effectively reduce radon concentration are determined through an optimization algorithm; return to S643, replace the ventilation habit feature parameters with the recommended ventilation period and recommended ventilation frequency, and re-execute S643 and S644 until it is determined that there is no risk of radon concentration exceeding the standard, and output the current recommended ventilation period and recommended ventilation frequency; S644b If there is no risk of radon concentration exceeding the standard, the current ventilation habit characteristic parameters will be output as ventilation recommendations.

7. The indoor radon gas detection method according to claim 1, characterized in that, The iterative correction factor algorithm model specifically includes: ； in, This is the final concentration data for the current period. This is the intermediate concentration data for the current period. The final concentration data for the previous measurement period is given, and k is the residual memory effect coefficient, ranging from 0.1 to 0.

3.

8. The indoor radon gas detection method according to claim 3, characterized in that, The abrupt change data points are abnormal data points that exceed a preset threshold range due to instrument vibration, electromagnetic interference, or momentary power outage. The preset threshold range is the distribution interval of normal values ​​determined based on the statistical distribution of historical data.

9. An indoor radon gas detection system, characterized in that, The system is used to implement the indoor radon gas detection method according to any one of claims 1 to 8, the system comprising: The raw data acquisition module is used to collect radon progeny in indoor air onto the surface of the CZT detector by electrostatic enrichment, and to perform alpha spectrum analysis using the CZT detector to continuously acquire raw measurement data of radon concentration for a first predetermined period. An environmental data acquisition module, connected to the original data acquisition module, is used to synchronously acquire environmental parameter data, including temperature, humidity and air pressure, through a built-in environmental sensor at the first predetermined cycle. The data preprocessing module, connected to the environmental data acquisition module, is used to clean the raw measurement data and the environmental parameter data, remove abnormal data points caused by instrument disturbances, and align the cleaned multi-source data on the time axis to generate a synchronous data frame sequence. The temperature and humidity compensation module, connected to the data preprocessing module, is used to input the synchronous data frame sequence into the pre-built and trained artificial neural network temperature and humidity correction model, and to perform a first-level correction on the original measurement data of the current period to eliminate the influence of temperature and humidity changes on the counting efficiency of the CZT detector, and generate intermediate concentration data after temperature and humidity compensation. The iterative correction module, connected to the temperature and humidity compensation module, is used to perform secondary correction based on the final concentration data of the previous measurement cycle and the intermediate concentration data, using a preset iterative correction factor algorithm, in order to eliminate the interference of the residual memory effect of the previous measurement in the detection cavity on the current measurement, thereby generating the final concentration data of the current cycle. The comprehensive report output module, connected to the iterative correction module, is used to calculate the average concentration and confidence interval of radon during the entire measurement period based on the final concentration data of multiple consecutive periods. It also combines the synchronously collected air pressure data and the rate of change of radon concentration over time to couple the ventilation status and radon concentration during the measurement period, and finally generate a comprehensive test report containing radon concentration values, environmental parameter change curves, and ventilation recommendations.

Images