Suspended particulate concentration calibration method

Abstract

This invention relates to a method for calibrating the concentration of suspended particles, applicable to air quality monitoring, environmental detection devices, and cleanroom monitoring systems. It combines linear and quadratic regression models, using linear regression for calibration in the low concentration range (<100 µg / m³) and quadratic regression for calibration in the high concentration range (≥100 µg / m³), and automatically switches between the models to output calibration data. Furthermore, through a "dual calibration" and "dynamic update" mechanism, it effectively corrects the nonlinear deviation and long-term aging error of the optical sensor under high concentration conditions, improving the accuracy and reliability of concentration detection.

Claims

1. A method for calibrating the concentration of suspended particulate matter, comprising: data acquisition, wherein a sensor module detects the concentration of suspended particulate matter and synchronously records data with a standard detection instrument, and transmits the data to a regression calibration module after timestamp alignment and filtering; primary regression calibration, wherein the regression calibration module performs primary regression calibration when the concentration is less than 100 µg / m³, for data calibration using a primary regression model; and secondary regression calibration, wherein the regression calibration module performs secondary regression calibration when the concentration is greater than or equal to 100 µg / m³, for data calibration using a secondary regression model. Automatic switching: Based on the concentration of suspended particles, a threshold is determined, and the primary regression model and the secondary regression model are automatically switched, and the corrected data of the suspended particle concentration are output. The automatic switching further includes a set concentration threshold range. When the concentration value falls within the set concentration threshold range, the current model is maintained to avoid switching jitter. Dual calibration and dynamic update: After completing the primary and secondary regression calibrations, a dual calibration step is further performed. This includes performing a first calibration using the primary regression model and using its output value as input to refit the secondary regression model for a second calibration to correct residual bias. The data is dynamically updated through a data output module, which periodically compares the data from the sensor module with the data from the standard detection instrument and sends the output results back to the regression correction module for dynamic parameter updates and automatic correction.

2. The method for calibrating the concentration of suspended particles as described in claim 1, wherein the primary regression calibration model is Y=aX+b, wherein the sensor module outputs the original scattering signal or the converted concentration value X, and the standard detection instrument outputs the corresponding reference concentration value Y, so as to perform primary regression calibration, thereby maintaining the linearity and stability of the response in the low concentration range.

3. The method for calibrating the concentration of suspended particles as described in claim 1, wherein the quadratic regression model is Y=bX2+aX, is used for quadratic regression calibration to compensate for the nonlinear saturation effect of optical scattering signals under high concentrations, so that the calibration results are closer to the data of standard testing instruments.

4. The method for calibrating the concentration of suspended particles as described in claim 1, wherein the double calibration state update performs a double calibration step, performing a first calibration with a linear regression model Y=bX2+aX, and then using its output value Y as input to fit a second quadratic regression model Y1=b1Y2+a1 for a second calibration, so that the expanded model contains quartic and cubic terms to correct residual bias.

5. The suspended particulate concentration calibration method as described in claim 4, wherein the dual calibration and dynamic update are dynamically updated by the data output module, and the sensor module data is compared with the standard detection instrument data periodically. If the error exceeds the threshold, the parameters a, b, a₁, b₁ are re-estimated and updated to ensure long-term accuracy.

6. The particulate matter concentration calibration method as described in claim 1, wherein the dual calibration and dynamic update, under extremely high concentration conditions, calculates the ratio of a reference parameter (PMref) to an instantaneous parameter (dPMref), and corrects the output particulate matter concentration data according to the ratio to avoid underestimation caused by signal overlap.

Images