A high-precision toxic gas detector signal processing method and storage medium

Abstract

This invention relates to the fields of gas detection technology and digital signal processing technology, specifically a high-precision signal processing method for a toxic gas detector, comprising the following steps: S1: Weak signal acquisition and ultra-high precision analog front-end preprocessing: Acquiring the microvolt-level weak analog signal output from the toxic gas sensor, and sequentially performing high-precision I-V conversion, high common-mode rejection ratio differential amplification, high-order low-pass filtering, adaptive power frequency notch filtering, impedance matching, and bias voltage regulation to obtain a low-noise, highly stable, and amplitude-adaptive analog signal. This high-precision toxic gas detector signal processing method and storage medium employs high-precision solutions at every step, from analog front-end preprocessing, analog-to-digital conversion, digital noise reduction to error compensation and concentration calculation, effectively suppressing various types of noise and interference to meet the needs of high-end applications.

Description

Technical Field

[0001] This invention relates to the fields of gas detection technology and digital signal processing technology, specifically to a high-precision signal processing method and storage medium for a toxic gas detector. Background Technology

[0002] Toxic gas detectors are widely used in petrochemical, mining, scientific research laboratories, and environmental monitoring fields. Their detection accuracy is directly related to personnel safety, environmental monitoring accuracy, and the reliability of scientific research data. With industrial upgrading and increasing scientific research needs, the accuracy requirements for toxic gas detection have risen from traditional high precision to ultra-high precision, with a resolution of 0.001 ppm.

[0003] Existing signal processing methods for toxic gas detectors have the following shortcomings: 1. The simulation front-end preprocessing stage has insufficient processing accuracy for the weak microvolt-level signals output by the sensor, resulting in low common-mode rejection ratio and susceptibility to power frequency interference and electromagnetic interference, leading to high signal noise. 2. In the analog-to-digital conversion stage, the effective number of bits is insufficient and the quantization noise is high, which cannot meet the digital requirements of ultra-high precision detection. 3. Digital noise reduction methods are limited, often employing simple averaging or median filtering, which makes it difficult to simultaneously suppress spike interference, random noise, and signal drift. 4. The error compensation is not comprehensive enough. It mostly uses single temperature compensation or linear compensation, which cannot accurately fit the nonlinear effect of temperature and humidity coupling on the sensor output, and does not fully consider the errors caused by air pressure deviation and cross interference. 5. The concentration calculation model is not accurate enough, and the calculation error is large in the low concentration range, which cannot meet the requirements of ultra-low concentration detection.

[0004] To address the aforementioned issues, we propose an improvement: a high-precision signal processing method and storage medium for toxic gas detectors. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a high-precision signal processing method for a toxic gas detector, comprising the following steps: S1: Weak signal acquisition and ultra-high precision analog front-end preprocessing: Acquire the microvolt-level weak analog signal output by the toxic gas sensor, and sequentially perform high-precision IV conversion, high common-mode rejection ratio differential amplification, high-order low-pass filtering, adaptive power frequency notch filtering, impedance matching and bias voltage regulation to obtain a low-noise, high-stability, amplitude-adaptive analog signal. S2: High-precision analog-to-digital conversion: A 24-bit Σ-Δ ADC is used to synchronously sample the pre-processed analog signal, while simultaneously acquiring auxiliary signals such as temperature, humidity, and air pressure; through oversampling and decimation filtering, the effective resolution of the ADC is improved, quantization noise is reduced, and a high-resolution, low-noise digital signal is output; S3: Multi-stage noise reduction processing: Employs a four-stage filtering combination: a. Power frequency notch filtering; b. Moving weighted average; c. Adaptive median filtering; d. Improved adaptive Kalman filtering; Suppresses spike interference, random noise, and signal drift to improve signal stability; S4: Multi-dimensional error compensation: Two-dimensional tri-term temperature and humidity compensation is adopted, combined with air pressure compensation, ultra-high precision zero-point automatic calibration and cross-interference decoupling compensation, to eliminate various system errors and obtain ultra-high precision compensation signals. S5: Concentration calculation based on high-order calibration model: A cubic polynomial concentration calibration model is established through multi-point high-order calibration, and combined with piecewise high-precision interpolation, the compensation signal is calculated into an ultra-high precision toxic gas concentration value. S6: Intelligent Judgment, Fault Diagnosis and Stable Output: Performs hysteresis judgment and trend prediction on the calculated concentration value, monitors the detector's operating status in real time, and outputs standardized concentration values, status and alarm information.

[0006] As a preferred embodiment of the present invention, the high-precision IV conversion in step S1 employs a low-temperature drift, low-offset precision operational amplifier, with a conversion linearity error ≤0.05%; The differential amplifier uses an ultra-high precision instrumentation amplifier with a common-mode rejection ratio ≥140dB and an adjustable gain of 100-1000 times; the high-order low-pass filter uses a 4th-order Butterworth low-pass filter circuit with an adjustable cutoff frequency of 1-10Hz. The adaptive power frequency notch filter module dynamically tracks 50 / 60Hz power frequency interference to ensure that the signal-to-noise ratio of the preprocessed signal is ≥80dB.

[0007] As a preferred technical solution of the present invention, the ADC sampling clock accuracy in step S2 is ≤1ppm, the multi-channel sampling timing deviation is ≤1μs, the oversampling rate is ≥256 times, and the effective number of bits of the ADC is increased to more than 20 bits by combining the CIC decimation filter and the FIR low-pass filter, and the quantization noise is ≤0.001ppm.

[0008] As a preferred technical solution of the present invention, the power frequency notch filter in step S3 uses a dual-T type active filter circuit or a digital notch filter algorithm to accurately filter out 50Hz mains interference; the sliding weighted average filter uses a 15-point sliding window and performs smoothing processing according to the principle that the closer to the current sampling point, the greater the weight; the adaptive median filter window is dynamically adjusted within a range of 3-7 points to accurately remove peak interference. The improved adaptive Kalman filter dynamically adjusts the Kalman gain (0.01-0.1) to suppress system noise and slow drift, with signal stability ≤ ±0.01%FS / 24h.

[0009] As a preferred technical solution of the present invention, the two-dimensional temperature and humidity compensation model formula in step S4 is as follows: ; Current temperature ( ),humidity( The sensor output voltage correction amount under ( ) :coefficient, Real-time ambient temperature, in units of , Real-time ambient relative humidity, in units of ; The pressure compensation adopts a linear correction model with a correction error ≤0.003ppm; the zero-point automatic calibration cycle is adjustable from 1 to 24 hours with a calibration error ≤0.005ppm; the cross-interference decoupling adopts a multi-sensor array with an interference suppression ratio ≥50dB.

[0010] As a preferred embodiment of the present invention, the cubic polynomial concentration calibration model in step S5 is as follows: ; Gas concentration; For calibration coefficients, This is a pressure correction signal.

[0011] A storage medium for computer-readable storage, wherein the storage medium is an embedded Flash memory or an SD card, thereon storing computer program instructions that are executed by an embedded microprocessor such as an STM32 or ESP32.

[0012] The beneficial effects of this invention are: I. The signal processing method and storage medium of this high-precision toxic gas detector adopt a high-precision scheme in every step, from analog front-end preprocessing, analog-to-digital conversion, digital noise reduction to error compensation and concentration calculation. This effectively suppresses various noises and interferences, ensuring that the detection accuracy reaches ±0.5%FS and the resolution is ≤0.001ppm, meeting the needs of high-end scenarios.

[0013] II. The signal processing method and storage medium of this high-precision toxic gas detector adopts a two-dimensional temperature and humidity compensation model of cubic polynomial, which is more in line with the nonlinear relationship between temperature and humidity and sensor output compared with traditional linear or two-dimensional compensation, and effectively eliminates the influence of temperature drift and humidity drift.

[0014] III. The signal processing method and storage medium of this high-precision toxic gas detector integrate multi-dimensional compensation for temperature and humidity, air pressure, zero drift, and cross-interference, completely eliminating various systematic errors and improving detection stability.

[0015] IV. The signal processing method and storage medium of this high-precision toxic gas detector effectively suppress electromagnetic interference, power frequency interference, and spike interference through high common-mode rejection ratio differential amplification, adaptive power frequency notch filtering, and multi-level digital noise reduction technologies, with a signal-to-noise ratio ≥80dB. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall flowchart of the signal processing method for the ultra-high precision toxic gas detector of the present invention; Figure 2 This is a schematic diagram of the signal acquisition and front-end preprocessing module of the present invention; Figure 3 This is a schematic diagram of the calibration surface of the two-dimensional temperature and humidity compensation model of the present invention; Figure 4 This is a flowchart illustrating the multi-level digital noise reduction processing module of the present invention; Figure 5 This is a logical schematic diagram of the multi-dimensional error compensation of the present invention; Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0018] Example: Figures 1-5 As shown, a high-precision toxic gas detector signal processing method includes the following steps: S1: Weak signal acquisition and ultra-high precision analog front-end preprocessing: Acquire the microvolt-level weak analog signal output by the toxic gas sensor, and sequentially perform high-precision IV conversion, high common-mode rejection ratio differential amplification, high-order low-pass filtering, adaptive power frequency notch filtering, impedance matching and bias voltage regulation to obtain a low-noise, high-stability, amplitude-adaptive analog signal. The specific processing steps are as follows: First, an electrochemical CO sensor is used to collect its weak analog output signal. This signal is susceptible to electromagnetic interference, temperature drift, and humidity drift, requiring ultra-high precision analog front-end conditioning. The specific operation is as follows: High-precision I / V conversion: Employs an ultra-low noise, zero-drift operational amplifier (model: ADA4528) to convert the weak current signal (1-10nA) output by the sensor into a voltage signal, with the conversion linearity error controlled within 0.05%, ensuring no signal distortion during the conversion process.

[0019] High common-mode rejection ratio differential amplification: An ultra-high precision instrumentation amplifier (model: AD8421) is selected to differentially amplify the converted voltage signal. The gain is adjusted to 500 times. The common-mode rejection ratio of this amplifier is ≥140dB, which can effectively suppress external common-mode interference and avoid the influence of environmental electromagnetic noise on weak signals.

[0020] High-order low-pass filtering and adaptive power frequency notch filtering: A 4th-order Butterworth low-pass filter circuit is used with a cutoff frequency of 5Hz to filter out high-frequency noise (frequency > 5Hz); at the same time, an adaptive 50Hz power frequency notch filtering module is added to dynamically track power frequency interference signals and further filter out power frequency interference and harmonics, so that the signal-to-noise ratio of the pre-processed signal is improved to over 80dB.

[0021] Impedance matching and bias regulation: Impedance matching between the sensor and the amplifier is achieved through a precision resistor network (matching impedance is 1kΩ), while providing a stable bias voltage (2.5V) to ensure that the signal is attenuated and distorted before entering the ADC, laying the foundation for subsequent ultra-high precision acquisition.

[0022] After the above preprocessing, a low-noise, highly stable, amplitude-adaptive (0.005-0.05V) analog signal is obtained, which then enters the next step of analog-to-digital conversion.

[0023] S2: High-precision analog-to-digital conversion: A 24-bit Σ-Δ ADC (model: ADS1248) is used to synchronously sample the pre-processed analog signal, while simultaneously acquiring auxiliary signals such as temperature, humidity, and air pressure; through oversampling and decimation filtering, the effective resolution of the ADC is improved, quantization noise is reduced, and a high-resolution, low-noise digital signal is output; The specific processing steps are as follows: The ADC opens 4 acquisition channels to synchronously acquire CO sensor signals, temperature signals (acquisition module: DS18B20, accuracy ±0.05℃), humidity signals (acquisition module: SHT30, accuracy ±1%RH) and air pressure signals (acquisition module: BMP280, accuracy ±0.1kPa). The sampling timing deviation of each channel is ≤1μs to ensure that the compensation data is strictly synchronized with the gas signals and to avoid compensation deviation caused by timing errors.

[0024] High-level oversampling and decimation filtering: The ADC oversampling rate is set to 256x. Combined with a CIC decimation filter (decimation rate 16) and an FIR low-pass filter (order 128), noise reduction and decimation processing are performed, increasing the effective bit depth of the ADC to over 20 bits and reducing quantization noise to below 0.001ppm, effectively improving the resolution and purity of the digital signal. The ADC outputs a high-resolution, low-noise digital signal, which then enters the subsequent digital noise reduction processing stage.

[0025] S3: Multi-stage noise reduction processing: Employs a four-stage filtering combination: a. Power frequency notch filtering; b. Moving weighted average; c. Adaptive median filtering; d. Improved adaptive Kalman filtering; Suppresses spike interference, random noise, and signal drift to improve signal stability; The specific processing steps are as follows: Adaptive median filtering: Set the dynamic adjustment range of the filtering window to 3-7 points. When pulse interference, spike noise or abnormal jump points are detected, the window size is automatically increased to accurately remove interference points. When the signal is stable, the window size is reduced to avoid damaging the real concentration change signal and ensure the integrity of the filtered signal.

[0026] Sliding weighted average filtering: Select 10 consecutive sampling points and perform weighted smoothing calculation according to the principle that the closer to the current sampling point, the greater the weight (the weight coefficients are 0.2, 0.18, 0.16, 0.14, 0.12, 0.08, 0.06, 0.04, 0.02, 0.0). This effectively suppresses random white noise and improves signal stability.

[0027] Improved adaptive Kalman filter: Using real-time acquired temperature and humidity data and sensor baseline drift as observed variables, a state equation and an observation equation are established, and the Kalman gain is dynamically adjusted (adjustment range 0.01-0.1) to optimally estimate and suppress system noise and slow drift, providing reliable data support for subsequent ultra-high precision compensation.

[0028] S4: Multi-dimensional error compensation: Two-dimensional tri-term temperature and humidity compensation is adopted, combined with air pressure compensation, ultra-high precision zero-point automatic calibration and cross-interference decoupling compensation, to eliminate various system errors and obtain ultra-high precision compensation signals. The formula for the two-dimensional temperature and humidity compensation model in step S4 is as follows: ; Current temperature ( ),humidity( The sensor output voltage correction amount under ( ) :coefficient, Real-time ambient temperature, in units of , Real-time ambient relative humidity, in units of ; The actual value of the sensor output and the theoretical standard value are obtained through high-precision calibration experiments. The correction amount is calculated, and the compensation coefficients (a0~a9) are obtained by least squares fitting. The set of coefficients is stored in the on-chip high-precision EEPROM for real-time retrieval.

[0029] Real-time compensation calculation: The detector collects the current ambient temperature in real time. ) and humidity ( Substituting these values ​​into the aforementioned two-dimensional temperature and humidity compensation model, the correction amount is calculated. Then, the true signal for eliminating temperature and humidity drift is calculated using the following formula: ; In the formula, The voltage value is the result of converting the original digital signal acquired by the ADC. This is a true signal without temperature or humidity drift after temperature and humidity compensation.

[0030] Pressure compensation: A pressure correction term is introduced, and a linear correction model is used to correct for gas diffusion and concentration deviations under different altitudes and pressures. The correction formula is as follows: ; In the formula, This is the pressure-corrected signal. The pressure is standard atmosphere (101.325 kPa), and P is the real-time collected air pressure value. This is the pressure correction factor.

[0031] Ultra-high precision zero-point automatic calibration: The calibration cycle is set to 12 hours. High-purity clean air with a purity of ≥99.999% is introduced at regular intervals to perform zero-point calibration. The sensor baseline drift is automatically detected and the zero-point parameters are updated, effectively eliminating the error caused by long-term baseline drift of the sensor.

[0032] S5: Concentration calculation based on high-order calibration model: A cubic polynomial concentration calibration model is established through multi-point high-order calibration, and combined with piecewise high-precision interpolation, the compensation signal is calculated into an ultra-high precision toxic gas concentration value. The cubic polynomial concentration calibration model in step S5: ; Gas concentration; For calibration coefficients, This is a pressure correction signal.

[0033] The full-range calibration was performed using standard CO gas with five concentration gradients. After each concentration gradient stabilized, the compensated signal value was collected, and the cubic polynomial concentration calibration model was established by fitting the data using the least squares method.

[0034] Piecewise high-precision interpolation: For the ultra-low concentration range of 0–10 ppm, a piecewise linear interpolation algorithm is used, dividing this range into 10 sub-segments. Linear interpolation calculations are performed on each sub-segment to further improve the resolution and accuracy at low concentrations. At the same time, the concentration values ​​are limited to ensure the reasonableness of the output concentration.

[0035] S6: Intelligent Judgment, Fault Diagnosis and Stable Output: Performs hysteresis judgment and trend prediction on the calculated concentration value, monitors the detector's operating status in real time, and outputs standardized concentration values, status and alarm information.

[0036] The high-precision IV conversion in step S1 uses a low-temperature drift, low-offset precision operational amplifier, with a conversion linearity error ≤0.05%; The differential amplifier uses an ultra-high precision instrumentation amplifier with a common-mode rejection ratio ≥140dB and an adjustable gain of 100-1000 times; the high-order low-pass filter uses a 4th-order Butterworth low-pass filter circuit with an adjustable cutoff frequency of 1-10Hz. The adaptive power frequency notch filter module dynamically tracks 50 / 60Hz power frequency interference to ensure that the signal-to-noise ratio of the pre-processed signal is ≥80dB.

[0037] In step S2, the ADC sampling clock accuracy is ≤1ppm, the multi-channel sampling timing deviation is ≤1μs, the oversampling rate is ≥256 times, and the effective number of bits of the ADC is increased to more than 20 bits by combining the CIC decimation filter and the FIR low-pass filter, and the quantization noise is ≤0.001ppm.

[0038] In step S3, the power frequency notch filter uses a dual-T active filter circuit or a digital notch filter algorithm to accurately filter out 50Hz mains interference; the sliding weighted average filter uses a 15-point sliding window, and performs smoothing according to the principle that the closer to the current sampling point, the greater the weight; the adaptive median filter window dynamically adjusts the range of 3-7 points to accurately remove peak interference. The improved adaptive Kalman filter dynamically adjusts the Kalman gain (0.01-0.1) to suppress system noise and slow drift, with signal stability ≤ ±0.01%FS / 24h.

[0039] The air pressure compensation adopts a linear correction model with a correction error ≤0.003ppm; the zero-point automatic calibration cycle is adjustable from 1 to 24 hours with a calibration error ≤0.005ppm; the cross-interference decoupling adopts a multi-sensor array with an interference suppression ratio ≥50dB.

[0040] A storage medium for computer-readable storage, wherein the storage medium is an embedded Flash memory or an SD card, on which computer program instructions are stored, which are executed by an embedded microprocessor such as an STM32 or ESP32.

[0041] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high-precision method for signal processing of a toxic gas detector, characterized in that, Includes the following steps: S1: Weak signal acquisition and ultra-high precision analog front-end preprocessing: Acquire the microvolt-level weak analog signal output by the toxic gas sensor, and sequentially perform high-precision IV conversion, high common-mode rejection ratio differential amplification, high-order low-pass filtering, adaptive power frequency notch filtering, impedance matching and bias voltage regulation to obtain a low-noise, high-stability, amplitude-adaptive analog signal. S2: High-precision analog-to-digital conversion: A 24-bit Σ-Δ ADC is used to synchronously sample the pre-processed analog signal, while simultaneously acquiring auxiliary signals such as temperature, humidity, and air pressure; through oversampling and decimation filtering, the effective resolution of the ADC is improved, quantization noise is reduced, and a high-resolution, low-noise digital signal is output; S3: Multi-stage noise reduction processing: Employs a four-stage filtering combination: a. Power frequency notch filtering; b. Moving weighted average filtering; c. Adaptive median filtering; d. Improved adaptive Kalman filter; Suppress spike interference, random noise, and signal drift to improve signal stability; S4: Multi-dimensional error compensation: Two-dimensional tri-term temperature and humidity compensation is adopted, combined with air pressure compensation, ultra-high precision zero-point automatic calibration and cross-interference decoupling compensation, to eliminate various system errors and obtain ultra-high precision compensation signals. S5: Concentration calculation based on high-order calibration model: A cubic polynomial concentration calibration model is established through multi-point high-order calibration, and combined with piecewise high-precision interpolation, the compensation signal is calculated into an ultra-high precision toxic gas concentration value. S6: Intelligent Judgment, Fault Diagnosis and Stable Output: Performs hysteresis judgment and trend prediction on the calculated concentration value, monitors the detector's operating status in real time, and outputs standardized concentration values, status and alarm information.

2. The signal processing method of claim 1, wherein, The high-precision IV conversion in step S1 uses a low-temperature drift, low-offset precision operational amplifier, with a conversion linearity error ≤0.05%; The differential amplifier uses an ultra-high precision instrumentation amplifier with a common-mode rejection ratio ≥140dB and an adjustable gain of 100-1000 times; the high-order low-pass filter uses a 4th-order Butterworth low-pass filter circuit with an adjustable cutoff frequency of 1-10Hz. The adaptive power frequency notch filter module dynamically tracks 50 / 60Hz power frequency interference to ensure that the signal-to-noise ratio of the preprocessed signal is ≥80dB.

3. The signal processing method of a high-precision toxic gas detector according to claim 1, characterized in that, In step S2, the ADC sampling clock accuracy is ≤1ppm, the multi-channel sampling timing deviation is ≤1μs, the oversampling rate is ≥256 times, and the effective number of bits of the ADC is increased to more than 20 bits by combining the CIC decimation filter and the FIR low-pass filter, and the quantization noise is ≤0.001ppm.

4. The signal processing method of a high-precision toxic gas detector according to claim 1, characterized in that, The power frequency notch filtering in step S3 uses a dual-T active filter circuit or a digital notch filtering algorithm to accurately filter out 50Hz mains interference; the sliding weighted average filtering uses a 15-point sliding window and performs smoothing processing according to the principle that the closer to the current sampling point, the greater the weight. The adaptive median filter window dynamically adjusts within a range of 3-7 points to accurately eliminate spike interference. The improved adaptive Kalman filter dynamically adjusts the Kalman gain (0.01-0.1) to suppress system noise and slow drift, with signal stability ≤ ±0.01%FS / 24h.

5. The signal processing method of a high-precision toxic gas detector according to claim 1, wherein The temperature and humidity two-dimensional compensation model formula in step S4: ; : sensor output voltage correction amount under current temperature (T), humidity (H), : sensor output voltage correction amount under current temperature (T), humidity (H), : coefficient, : real-time ambient temperature, unit: °C , : real-time ambient relative humidity, unit: % ;​ The air pressure compensation adopts a linear correction model, the correction error is less than or equal to 0.003 ppm; the zero automatic calibration period is adjustable between 1-24 hours, the calibration error is less than or equal to 0.005 ppm; the cross interference decoupling adopts a multi-sensor array, and the interference suppression ratio is greater than or equal to 50 dB.

6. The signal processing method of a high-precision toxic gas detector according to claim 1, wherein The cubic polynomial concentration calibration model in step S5: ; is a gas concentration, is a calibration coefficient, is a gas pressure correction signal.

7. A storage medium for computer-readable storage, characterized in that, The storage medium is an embedded Flash memory or an SD card, and computer program instructions are stored on the storage medium. When the program instructions are executed by an embedded microprocessor such as an STM32 or an ESP32, the signal processing method of the high-precision toxic gas detector in any one of claims 1-6 can be implemented.

Images