Poisonous gas concentration detection method based on multi-sensor data fusion and acquisition system

Abstract

The invention discloses a toxic gas concentration detection method based on multi-sensor data fusion and a toxic gas concentration acquisition system based on multi-sensor data fusion. A plurality of different types of toxic gas sensors are used for collecting toxic gas original data in real time; filtering and noise reduction are carried out on the collected data, so that the data quality is improved; fusing the preprocessed data by adopting a weighted average method and a Kalman filtering algorithm to obtain more accurate gas concentration information; calculating the actual concentration of the toxic gas according to the fused data; and monitoring the calculated concentration value in real time by combining a detection method based on a threshold value and a detection method based on statistical analysis, and when the concentration value exceeds a normal range or deviates from normal distribution of a statistical model, judging that the concentration value is abnormal. The concentration of toxic gas is accurately and dynamically monitored in real time by using various sensors, the abnormal condition of the gas concentration can be detected in time, and a powerful guarantee is provided for environmental safety and industrial production.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental monitoring, and in particular relates to a toxic gas concentration detection method and acquisition system based on multi-sensor data fusion. Background Art

[0002] In industrial production, environmental protection, and daily life, the leakage and excessive emission of toxic gases can pose serious risks to human health and the ecological environment. Traditional toxic gas monitoring methods often rely on a single sensor, which suffers from low monitoring accuracy and poor reliability. Single sensors are easily affected by environmental factors (such as temperature, humidity, and electromagnetic interference), resulting in inaccurate measurement results and an inability to accurately reflect the true concentration of toxic gases in a timely manner. Furthermore, a single sensor cannot comprehensively monitor multiple toxic gases in complex environments, making it difficult to meet the needs of practical applications.

[0003] In recent years, multi-sensor data fusion technology has become a research hotspot. By fusing data from multiple different sensor types, this technology can fully leverage the strengths of each sensor and improve monitoring accuracy and reliability. However, existing multi-sensor data fusion methods still have some shortcomings in the field of toxic gas monitoring, such as complex fusion algorithms, high computational complexity, and poor real-time performance. These issues hinder the effective dynamic monitoring of toxic gas concentrations and anomaly detection. Therefore, the development of an efficient and accurate multi-sensor data fusion toxic gas concentration detection method and acquisition system is of great practical significance.

[0004] The defect of the existing technology is how to provide a toxic gas concentration detection method with multi-sensor data fusion, which aims to use multiple sensors to conduct real-time and accurate dynamic monitoring of toxic gas concentrations, and to detect abnormal gas concentrations in a timely manner, providing strong protection for environmental safety and industrial production. Summary of the Invention

[0005] In view of the above-mentioned deficiencies in the existing technology, the technical problem to be solved by the present invention is: how to provide a toxic gas concentration detection method with multi-sensor data fusion, which uses multiple sensors to perform real-time and accurate dynamic monitoring of toxic gas concentrations, and can promptly detect abnormal gas concentrations, providing strong protection for environmental safety and industrial production.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions: a multi-sensor data fusion toxic gas concentration detection method and acquisition system, comprising the following steps;

[0007] Step 1: Data acquisition, using multiple types of toxic gas sensors to collect the original data of toxic gases in real time; the data acquisition module performs data acquisition;

[0008] Step 2: Data preprocessing: filtering and noise reduction of the collected raw data to improve the data quality; this work is performed by the data preprocessing module;

[0009] Step 3: Data fusion: The pre-processed data is fused using a combination of weighted averaging and Kalman filtering to obtain more accurate gas concentration information. This is done by the data fusion module.

[0010] Step 4: Concentration calculation: Calculate the actual concentration of the toxic gas based on the fused data; this is done by the concentration calculation module;

[0011] Step 5: Anomaly Detection. This module uses a combination of threshold-based and statistical analysis-based detection methods to monitor the calculated concentration values ​​in real time. When the concentration value exceeds the normal range or deviates from the normal distribution of the statistical model, it is determined to be an anomaly and an alarm is issued. This task is performed by the anomaly detection module.

[0012] The key to the multi-sensor data fusion toxic gas concentration detection method and acquisition system is that the formula of the weighted average method in step 3 is: Suppose there are n sensors in total, and the measurement value of the i-th sensor is x i , and its corresponding weight is w i , then the weighted average fusion result x avg It can be expressed as:

[0013]

[0014] Among them, the weight w i Determined based on the reliability and accuracy of the sensor.

[0015] The key to the multi-sensor data fusion toxic gas concentration detection method and acquisition system is that the Kalman filter algorithm in step three includes two steps: prediction and update; the formula of the prediction step is as follows:

[0016]

[0017] in, is the estimated value of the predicted state at time k, F k is the state transition matrix, is the estimated value of the state at time k-1, B k is the control input matrix, u k is the control input vector, P k|k-1 is the prediction covariance matrix at time k, P k-1|k-1 is the forecast covariance matrix at time k-1, Q k is the process noise covariance matrix;

[0018] The formula for the update step is as follows:

[0019]

[0020] P k|k =(IK k H k )P k|k-1

[0021] Among them, K k is the Kalman gain, H k is the observation matrix, z k is the observed value at time k, R k is the observation noise covariance matrix, is the estimated state value at time k, P k|k is the covariance matrix at time k, and I is the identity matrix.

[0022] The key point of the multi-sensor data fusion toxic gas concentration detection method and acquisition system is that the threshold-based anomaly detection formula in step 5 is: abnormal or normal

[0023]

[0024] Among them, x concentration is the calculated toxic gas concentration value, x min and x max These are the lower and upper limits of the normal concentration range, respectively.

[0025] The key to the multi-sensor data fusion toxic gas concentration detection method and acquisition system is that the anomaly detection formula based on statistical analysis in step 5 determines whether it is abnormal by calculating the degree of deviation between the current concentration value and the mean μ and standard deviation σ of the statistical model; when |x concentration When -μ|>kσ, it is determined to be an abnormal situation, where k is a constant with a value of 2 or 3.

[0026] In step 1, a toxic gas sensor is arranged in a sensor device, which includes a cylindrical sensor environment control tube, in which at least one toxic gas sensor is fixedly arranged, and an internal temperature sensor is also fixedly arranged in the sensor environment control tube. The toxic gas sensor and the internal temperature sensor are led out of the sensor environment control tube through wires and connected to a microcontroller of a data preprocessing module. The microcontroller is connected to a wireless communication module. A conical cold air converging tube is connected to the top of the sensor environment control tube, and the small end of the cold air converging tube is connected to the upper end of the sensor environment control tube, and the large end of the cold air converging tube is provided with a A first sealing plate is provided with a first air hole in the center of the first sealing plate, at least two semiconductor refrigeration plates are provided on the outer wall of the cold air converging tube from top to bottom, the semiconductor refrigeration plates are connected to the microcontroller, a conical hot air converging tube is connected to the bottom of the sensor environment control tube, the small end of the hot air converging tube is connected to the lower end of the sensor environment control tube, a second sealing plate is provided on the large end of the hot air converging tube, a second air hole is provided in the center of the second sealing plate, an electric heating wire is provided on the inner wall of the hot air converging tube, and the electric heating wire is connected to the microcontroller; the microcontroller is also connected to an external ambient temperature sensor, which is used to detect the temperature of the on-site environment.

[0027] The first air hole is provided with an air intake fan, and the air intake fan is connected to the microcontroller.

[0028] Significant effect: The present invention provides a multi-sensor data fusion toxic gas concentration detection method and acquisition system, which uses multiple sensors to perform real-time and accurate dynamic monitoring of toxic gas concentrations, and can promptly detect abnormal gas concentrations, providing strong protection for environmental safety and industrial production. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a module structure diagram of the present invention;

[0030] Figure 2 It is a module structure diagram in a specific implementation process of the present invention;

[0031] Figure 3 is a flow chart of the method of the present invention;

[0032] Figure 4 is a structural diagram of the sensor device;

[0033] Figure 5 This is the connection circuit diagram of the microcontroller, semiconductor cooling chip, and electric heating wire. DETAILED DESCRIPTION

[0034] The present invention will be described in further detail below with reference to the accompanying drawings.

[0035] like Figure 1-Figure 5As shown, the present invention provides a toxic gas concentration detection method based on multi-sensor data fusion, which aims to use multiple toxic gas sensors to perform real-time and accurate dynamic monitoring of toxic gas concentrations, and can promptly detect abnormal gas concentrations, providing strong protection for environmental safety and industrial production.

[0036] A plurality of toxic gas sensors 12 are disposed in the sensor device 1 .

[0037] (1) Algorithm logic diagram

[0038] 1. Overall logical architecture

[0039] like Figure 1 As shown in the figure, the algorithm logic diagram of the present invention mainly includes five main modules: data acquisition, data preprocessing, data fusion, concentration calculation, and anomaly detection. The data acquisition module is responsible for acquiring raw data from multiple different types of toxic gas sensors; the data preprocessing module performs filtering and noise reduction on the collected raw data to improve data quality; the data fusion module uses advanced fusion algorithms to fuse the preprocessed data to obtain more accurate gas concentration information; the concentration calculation module calculates the actual concentration of toxic gases based on the fused data; and the anomaly detection module monitors the calculated concentration values ​​in real time and immediately issues an alarm if an anomaly is detected.

[0040] 2. Logical relationship between modules

[0041] The data acquisition module is the foundation of the entire algorithm, providing raw data for subsequent processing. The data preprocessing module performs preliminary processing on the collected data, removing noise and interference, and providing more reliable data input for the data fusion module. The data fusion module is the core of the algorithm, integrating data from multiple sensors to improve the accuracy of gas concentration monitoring. The concentration calculation module calculates the actual concentration of toxic gases based on the fused data, providing a basis for the anomaly detection module. The anomaly detection module monitors concentration values ​​in real time and triggers an alarm when the concentration value exceeds the normal range, prompting relevant personnel to take action.

[0042] (2) Algorithm principle diagram

[0043] 1. Principle of multi-sensor data fusion

[0044] The multi-sensor data fusion method employed in this invention combines a weighted averaging method with a Kalman filter algorithm. The weighted averaging method assigns different weights to each toxic gas sensor based on its reliability and accuracy, then performs a weighted average of the sensor data to produce a preliminary fusion result. The Kalman filter algorithm further processes this weighted average result, continuously optimizing the fusion result through two steps: prediction and updating, thereby improving the accuracy and stability of the data.

[0045] like Figure 2 As shown, the multi-sensor system includes an electrochemical toxic gas sensor. In addition to a toxic gas concentration sensor, it also includes a temperature sensor, a pressure sensor, an image sensor, and a flow sensor. The electrochemical toxic gas sensor collects the concentration of harmful gases and uploads it to the cloud through the edge service engine. The signal of the electrochemical toxic gas sensor is pre-processed by the signal conditioning circuit and microcontroller, and then uploaded to the cloud through the edge service engine. The cloud performs data fusion, concentration calculation, and anomaly detection. Among them, the electrochemical toxic gas sensor is used to detect the toxic gas concentration in the current environment, the temperature sensor is used to detect the current temperature of the environment, and the image sensor is used to detect the real-time image of the current environment; the pressure sensor is used to detect the air pressure or hydraulic pressure in the pipeline, and the flow sensor is used to detect the flow of toxic gas or liquid in the pipeline. The pipeline is used to transmit toxic gas or liquid, and the toxic gas or liquid may leak into the external environment.

[0046] The multiple sensors may include electrochemical harmful gas sensors and / or semiconductor toxic gas sensors to improve the accuracy of toxic gas detection.

[0047] Toxic gas sensors include ammonia sensors, methane sensors, sulfur dioxide sensors, chlorine sensors, sulfur trioxide sensors, hydrogen sulfide sensors, carbon monoxide sensors, nitrogen dioxide sensors, etc.

[0048] 2. Anomaly Detection Principle

[0049] The anomaly detection module combines threshold-based and statistical analysis-based detection methods. The threshold-based detection method sets a normal concentration range and identifies an anomaly when the calculated concentration value exceeds this range. The statistical analysis-based detection method analyzes historical data to establish a statistical model of concentration values. If the current concentration value deviates from the normal distribution of the statistical model, it is identified as an anomaly.

[0050] (3) Toxic gas sensor circuit diagram

[0051] 1. Toxic gas sensor circuit design

[0052] The present invention can utilize a variety of toxic gas sensors, such as electrochemical and semiconductor sensors. Each sensor is equipped with its own independent signal conditioning circuit, which amplifies and filters the weak signal output by the sensor to meet the input requirements of the subsequent data acquisition module. The detailed circuit diagram is omitted. The signal conditioning circuit comprises an amplifier, a filter, and an analog-to-digital converter; the detailed circuit diagram is omitted.

[0053] 2. Data processing circuit design

[0054] The data preprocessing module uses a high-performance microcontroller (such as an STM32F103 microcontroller or an ARM processor) as its core processing unit. It connects to the signal conditioning circuits of each sensor via serial or SPI communication, collecting sensor output data in real time. After preliminary processing of the collected data, the microcontroller transmits the data to a host computer or edge server via a wireless communication module (such as ZigBee or WiFi) for further processing. Alternatively, it uploads the data to the cloud via an edge service engine for further processing and analysis (data fusion, concentration calculation, and anomaly detection). The detailed circuit diagram is omitted.

[0055] (IV) Sensor signal diagram

[0056] 1. Sensor original signal characteristics

[0057] Different types of toxic gas sensors produce signals with varying characteristics. Electrochemical sensors typically produce a weak current signal, whose magnitude is proportional to the concentration of the toxic gas. Semiconductor sensors produce a resistance signal, whose resistance changes with increasing toxic gas concentration. By analyzing the raw sensor signals, we can understand the sensor's operating status and changes in gas concentration. Depending on the site's environmental requirements, we can flexibly utilize electrochemical and / or semiconductor toxic gas sensors.

[0058] 2. Features after signal preprocessing

[0059] After filtering and noise reduction in the data preprocessing module, the sensor signal quality is significantly improved. Filtering removes high-frequency noise from the signal, making it smoother; noise reduction reduces random interference and improves signal stability. This preprocessed signal more accurately reflects changes in toxic gas concentrations.

[0060] 3. Characteristics of the Fusion Signal

[0061] After processing by the data fusion module, the resulting fused signal integrates information from multiple sensors, resulting in higher accuracy and reliability. The fused signal has a smaller fluctuation range and can more accurately reflect the actual concentration changes of toxic gases. By analyzing the fused signal, abnormal gas concentrations can be detected more promptly and accurately.

[0062] (V) Algorithm formula

[0063] 1. Weighted average method formula

[0064] Assume there are n sensors in total, and the measurement value of the i-th sensor is x i , and its corresponding weight is wi , then the weighted average fusion result x avg It can be expressed as:

[0065]

[0066] Among them, the weight w i It can be determined based on factors such as the reliability and accuracy of the sensor.

[0067] In one toxic environment, the weight of the electrochemical toxic gas sensor is 0.4, and the weight of the semiconductor toxic gas sensor is 0.6. The weights of the electrochemical toxic gas sensor and the semiconductor toxic gas sensor are assigned according to the importance of different locations.

[0068] Furthermore, in one of the toxic usage environments, the electrochemical toxic gas sensor includes a sulfur dioxide sensor, a nitrogen dioxide sensor and a nitric oxide sensor, the sulfur dioxide sensor has a weight of 0.5, the nitrogen dioxide sensor has a weight of 0.3, and the nitric oxide sensor has a weight of 0.2.

[0069] 2. Kalman filter algorithm formula

[0070] The Kalman filter algorithm includes two steps: prediction and update. The formula for the prediction step is as follows:

[0071]

[0072] in, is the estimated value of the predicted state at time k, F k is the state transition matrix, is the estimated value of the state at time k-1, B k is the control input matrix, u k is the control input vector, P k|k-1 is the prediction covariance matrix at time k, P k-1|k-1 is the forecast covariance matrix at time k-1, Q k is the process noise covariance matrix.

[0073] The formula for the update step is as follows:

[0074]

[0075] P k|k =(IK k H k )P k|k-1

[0076] Among them, K k is the Kalman gain, H k is the observation matrix, z kis the observed value at time k, R k is the observation noise covariance matrix, is the estimated state value at time k, P k|k is the covariance matrix at time k, and I is the identity matrix.

[0077] 3. Anomaly Detection Formula

[0078] The formula for anomaly detection based on threshold is: abnormal or normal

[0079]

[0080] Among them, x concentration is the calculated toxic gas concentration value, x min and x max These are the lower and upper limits of the normal concentration range, respectively.

[0081] The anomaly detection formula based on statistical analysis can determine whether it is abnormal by calculating the degree of deviation between the current concentration value and the mean μ and standard deviation σ of the statistical model. concentration When -μ|>kσ, it is determined to be an abnormal situation, where k is a constant, usually 2 or 3.

[0082] like Figure 4 and Figure 5 As shown, the toxic gas sensor 12 is arranged in the sensor device 1, and the sensor device 1 includes a cylindrical sensor environment control tube 11. At least one toxic gas sensor 12 is fixedly arranged in the sensor environment control tube 11. An internal temperature sensor 13 is also arranged in the sensor environment control tube 11. The toxic gas sensor 12 and the internal temperature sensor 13 are led out of the sensor environment control tube 11 through wires to a microcontroller connected to a data preprocessing module 14 outside the sensor environment control tube 11. The microcontroller is connected to a wireless communication module. A conical cold air converging tube 15 is connected to the top of the sensor environment control tube 11. The small end of the cold air converging tube 15 is connected to the upper end of the sensor environment control tube 11. The large end of the cold air converging tube 15 is provided with a first sealing plate 151, with a first air hole 1511 defined in the center. At least two semiconductor cooling fins 16 are provided on the outer wall of the cold air converging tube 15 from top to bottom. The semiconductor cooling fins 16 are connected to the microcontroller. A conical hot air converging tube 17 is connected to the bottom of the sensor environment control tube 11. The small end of the hot air converging tube 17 is connected to the lower end of the sensor environment control tube 11. The large end of the hot air converging tube 17 is provided with a second sealing plate 171, with a second air hole 1711 defined in the center. An electric heating wire 172 is provided on the inner wall of the hot air converging tube 17, which is connected to the microcontroller. The microcontroller is also connected to an external ambient temperature sensor for detecting the temperature of the on-site environment. The external ambient temperature sensor is not shown in the figure.

[0083] The internal temperature sensor 13 and the external environment temperature sensor can adopt DHT11 temperature and humidity sensor, AD590 temperature sensor and the like.

[0084] The temperature of the on-site environment is detected by an external ambient temperature sensor. This temperature may be as high as 40 degrees Celsius or as low as -20 degrees Celsius. The detection accuracy of electrochemical toxic gas sensors and semiconductor toxic gas sensors is easily affected by the ambient temperature, resulting in a large deviation in the detection accuracy. In addition, the ambient temperature can easily cause the toxic gas in the detection environment to expand when heated or contract when cooled, resulting in a deviation in the actual concentration of the toxic gas detected. The sensor device 1 of the present invention is used to adjust the operating temperature of the toxic gas sensor 12 so that its operating temperature is at a normal temperature of 20 degrees Celsius. In addition, the toxic gas in the detection environment is sampled and its temperature is adjusted to a normal temperature of 20 degrees Celsius before its concentration is detected. This makes the detection result closer to the actual concentration of the toxic gas and more accurately determines the leakage amount.

[0085] The internal temperature sensor 13 is used to detect the temperature inside the sensor environment control tube 11, that is, the operating temperature of the toxic gas sensor 12 and the sampled toxic gas temperature, and send the detected temperature to the microcontroller of the data preprocessing module 14. If the temperature is higher than 20 degrees, the microcontroller sends a signal through one of the switch transistors Q1 to control the refrigeration relay J1 to close, and the semiconductor refrigeration plate 16 works. The cooling surface of the semiconductor refrigeration plate 16 is close to the outer wall of the cold air convergence tube 15, reducing the temperature of the gas in the cold air convergence tube 15 to a normal temperature of 20 degrees. The heat dissipation surface of the semiconductor refrigeration plate 16 faces outward. The first air hole 1511 is used to take in air to sample toxic gases. The cold air convergence tube 15 is used to gather cold air. After the gas becomes cold, it sinks into the sensor environment control tube 11. When the temperature in the sensor environment control tube 11 reaches 20 degrees, it maintains the set time T1, and the microcontroller controls the toxic gas sensor 12 to sample. On the other hand, if the temperature of the detection environment is too low, less than 20 degrees Celsius, the microcontroller sends a signal through another switching transistor Q2 to control the heating relay J2 to close, energizing the electric heating wire 172. At this time, the semiconductor refrigeration plate 16 stops working, and the heat generated by the electric heating wire 172 is gathered by the hot gas converging tube 17 and rises into the environment control tube 11. At this time, the second air hole 1711 is used for air intake. When the temperature in the sensor environment control tube 11 reaches 20 degrees Celsius, it is maintained for a set time T1, and the microcontroller controls the toxic gas sensor 12 to take samples. Through the above structural arrangement, the operating temperature and sampled gas temperature of the toxic gas sensor 12 are both at 20 degrees Celsius, which improves the detection accuracy of the toxic gas sensor 12. At the same time, because the temperature of the toxic gas is also 20 degrees Celsius, the actual concentration of the toxic gas in the environment can be detected more accurately, and the leakage amount can be more accurately determined. The set time T1 is usually greater than 5 minutes.

[0086] The cold air converging tube 15 can not only converge cold air, but also prevent the toxic gas sensor 12 from direct contact with the external environment, thereby reducing the interference of the external environment temperature on the toxic gas sensor 12; the hot air converging tube 17 can not only converge hot air, but also reduce the interference of the external environment temperature on the toxic gas sensor 12.

[0087] The first air hole 1511 is provided with an intake fan 18 , which is connected to the microcontroller.

[0088] Before each sampling test, the microcontroller first controls the air intake fan 18 to work for a time T2, which is usually more than 1 minute, to remove the original gas in the sensor device 1 and inhale new test gas, so that the detection concentration of the toxic gas is more accurate.

[0089] The first sealing plate 151 and the second sealing plate 171 are both provided with a bracket 19, and the bracket 19 is provided with a mounting hole (the mounting hole is not shown in the figure). The bracket 19 can be installed on the ground, roof, and wall. Preferably, in a room where toxic gas may leak, the sensor device 1 is installed around the ground, roof, and wall to perform all-round detection of the room and detect toxic gas leakage in time.

[0090] The sensor environment control tube 11 is wrapped with an aluminum foil layer 111 to reduce the interference of the external electromagnetic environment on the toxic gas sensor 12.

[0091] In one embodiment, the model of the toxic gas sensor 12 is QT-MQ-135, the operating voltage is DC5V, the operating current is 150mA, the product type is a semiconductor sensitive component, the detection gas is ammonia, benzene vapor, sulfide, smoke and other toxic gases, the size is 32mm×20mm×22mm; the output supports switch digital signal and concentration analog signal output.

[0092] Preferably, the microcontroller can also adjust the ambient humidity through a corresponding humidity adjustment device after obtaining the humidity signal through the DHT11 temperature and humidity sensor (not shown).

[0093] IV. Innovation

[0094] (1) Multi-sensor data fusion innovation

[0095] This multi-sensor data fusion method combines weighted averaging and the Kalman filter algorithm, leveraging the strengths of each sensor and improving the accuracy and reliability of toxic gas concentration monitoring. The weighted averaging method assigns different weights to sensors based on their reliability and accuracy, effectively reducing the impact of unreliable sensor data on the fusion results. The Kalman filter algorithm continuously optimizes the fusion results through two steps: prediction and updating, improving data stability and real-time performance.

[0096] (2) Innovation in anomaly detection methods

[0097] This invention uses a combined threshold-based and statistical analysis-based anomaly detection method to more comprehensively and accurately detect abnormalities in toxic gas concentrations. The threshold-based detection method is simple and intuitive, quickly identifying obvious anomalies; the statistical analysis-based detection method, on the other hand, can identify potential anomalies by analyzing historical data, improving the sensitivity and accuracy of anomaly detection.

[0098] (3) The structure of the novel sensor device 1 makes the detection accuracy of the toxic gas sensor 12 more accurate and is less affected by the temperature of the external environment.

[0099] 5. Beneficial Effects

[0100] (1) Improving monitoring accuracy

[0101] By leveraging multi-sensor data fusion technology, the present invention fully leverages information from multiple sensors, effectively reducing the impact of environmental factors on monitoring results and improving the accuracy of toxic gas concentration monitoring. Compared to traditional single-sensor monitoring methods, the present invention provides more reliable monitoring results, providing stronger protection for environmental safety and industrial production.

[0102] (2) Enhance anomaly detection capabilities

[0103] The anomaly detection method employed in this invention combines threshold-based detection with statistical analysis, enabling timely and accurate detection of abnormal toxic gas concentrations. Both obvious and potential anomalies are promptly detected and an alarm is issued, providing ample time for relevant personnel to take action.

[0104] The above are only preferred embodiments of the present invention. It should be pointed out that various modifications and improvements made by those skilled in the art without departing from the present technical solution should also be deemed to fall within the scope of protection required by the claims.

Claims

1. A method for detecting toxic gas concentration by multi-sensor data fusion, characterized in that: The method includes the following steps: Step 1: Data collection, using multiple types of toxic gas sensors (12) to collect raw data of toxic gases in real time; Step 2: Data preprocessing, filtering and noise reduction of the collected raw data to improve the data quality; Step 3: Data fusion: The pre-processed data is fused by combining the weighted average method and the Kalman filter algorithm to obtain more accurate gas concentration information; Step 4: Concentration calculation: calculate the actual concentration of toxic gas based on the fused data; Step 5: Anomaly detection: A combination of threshold-based and statistical analysis-based detection methods is used to monitor the calculated concentration values ​​in real time. When the concentration value exceeds the normal range or deviates from the normal distribution of the statistical model, it is determined to be an abnormal situation and an alarm is issued.

2. The method for detecting toxic gas concentration by multi-sensor data fusion according to claim 1, characterized in that: The formula of the weighted average method in step 3 is: Assume there are n sensors in total, and the measurement value of the i-th sensor is x i , and its corresponding weight is w i , then the weighted average fusion result x avg It can be expressed as: Among them, the weight w i Determined based on the reliability and accuracy of the sensor.

3. The method for detecting toxic gas concentration by multi-sensor data fusion according to claim 1, characterized in that: The Kalman filter algorithm in step 3 includes two steps: prediction and update; the formula for the prediction step is as follows: in, is the estimated value of the predicted state at time k, F k is the state transition matrix, is the estimated value of the state at time k-1, B k is the control input matrix, u k is the control input vector, P k|k-1 is the prediction covariance matrix at time k, P k-1|k-1 is the forecast covariance matrix at time k-1, Q k is the process noise covariance matrix; The formula for the update step is as follows: P k|k =(I-K k H k )P k|k-1 Among them, K k is the Kalman gain, H k is the observation matrix, z k is the observed value at time k, R k is the observation noise covariance matrix, is the estimated state value at time k, P k|k is the covariance matrix at time k, and I is the identity matrix.

4. The method for detecting toxic gas concentration by multi-sensor data fusion according to claim 1, characterized in that: The threshold-based anomaly detection formula in step 5 is: abnormal or normal Among them, x concentration is the calculated toxic gas concentration value, x min and x max These are the lower and upper limits of the normal concentration range, respectively.

5. The method for detecting toxic gas concentration by multi-sensor data fusion according to claim 1, characterized in that: In step 5, the anomaly detection formula based on statistical analysis determines whether it is abnormal by calculating the degree of deviation between the current concentration value and the mean μ and standard deviation σ of the statistical model; when |x cocentration When -μ|>kσ, it is determined to be an abnormal situation, where k is a constant with a value of 2 or 3.

6. A multi-sensor data fusion toxic gas concentration detection and acquisition system, used for collecting toxic gas concentration in the toxic gas concentration detection method according to any one of claims 1 to 5, characterized in that: The toxic gas sensor (12) is arranged in the sensor device (1). The sensor device (1) comprises a cylindrical sensor environment control tube (11). At least one toxic gas sensor (12) is fixedly arranged in the sensor environment control tube (11). An internal temperature sensor (13) is also fixedly arranged in the sensor environment control tube (11). The toxic gas sensor (12) and the internal temperature sensor (13) are led out of the sensor environment control tube (11) through wires to a microcontroller connected to a data preprocessing module (14). The microcontroller is connected to a wireless communication module. A conical cold air converging tube (15) is connected above the sensor environment control tube (11). The small end of the cold air converging tube (15) is connected to the upper end of the sensor environment control tube (11). The large end of the cold air converging tube (15) is provided with a first sealing plate (14). 51), a first air hole (1511) is provided in the center of the first sealing plate (151), at least two semiconductor refrigeration plates (16) are provided on the outer wall of the cold air converging tube (15) from top to bottom, the semiconductor refrigeration plates (16) are connected to the microcontroller, a conical hot air converging tube (17) is connected below the sensor environment control tube (11), the small end of the hot air converging tube (17) is connected to the lower end of the sensor environment control tube (11), a second sealing plate (171) is provided on the large end of the hot air converging tube (17), a second air hole (1711) is provided in the center of the second sealing plate (171), an electric heating wire (172) is provided on the inner wall of the hot air converging tube (17), and the electric heating wire (172) is connected to the microcontroller; the microcontroller is also connected to an external environment temperature sensor, and the external environment temperature sensor is used to detect the temperature of the on-site environment.

7. A multi-sensor data fusion toxic gas concentration detection and acquisition system according to claim 6, characterized in that: The first air hole (1511) is provided with an air intake fan (18), and the air intake fan (18) is connected to the microcontroller.

Images