A gas concentration detection method based on gas sensor response overshoot compensation

Abstract

The present invention provides a gas concentration detection method based on gas sensor response overshoot compensation, which relates to the technical field of gas sensors. The present invention obtains the response characteristic curve of a semiconductor gas sensor and performs smoothing preprocessing on it to determine whether an overshoot signal exists. If an overshoot signal exists, the overshoot signal is reconstructed and smoothed based on the Eley-Rideal model to obtain a final response characteristic curve. Therefore, the present invention reconstructs and smoothes the response characteristic curve with the overshoot signal through reaction kinetic modeling and signal reconstruction technology, effectively compensating for the interference of the overshoot signal on the response characteristic curve of the semiconductor gas sensor, thereby improving the accuracy of gas concentration detection.

Description

Technical Field

[0001] The present invention relates to the technical field of gas sensors, in particular to a gas concentration detection method based on gas sensor response overshoot compensation. Background Art

[0002] Semiconductor gas sensors utilize the change in resistance caused by gas adsorption and reaction on the semiconductor material's surface. This change in resistance is converted into a usable output signal, enabling the detection of gas type and concentration. Specifically, when a gas sensor is continuously exposed to clean air at a constant operating temperature, its resistance remains stable. This resistance is also known as the gas sensor's baseline value. When the gas sensor is exposed to the gas being detected, its resistance gradually decreases (or increases) until it reaches a minimum (or maximum) value and persists for a period of time. This process is known as the gas sensor's response characteristic. When the gas sensor is removed from the gas being detected, its resistance gradually returns to its initial baseline value. This process is known as the gas sensor's recovery characteristic. However, when performing gas sensing tests to obtain the response and recovery characteristic curves of a gas sensor, difficult-to-eliminate overshoot signals may sometimes be observed in the response characteristic curve. An overshoot signal is a temporary, sudden change in the output signal before the gas sensor's response characteristic curve reaches a steady-state value, sometimes exceeding the expected steady-state value due to certain factors. Specifically, an overshoot signal manifests as a sharp peak in the gas sensor's response characteristic curve for a short period of time, followed by a gradual return to equilibrium. Overshoot signals can cause signal distortion, waveform errors, or parameter misjudgment in gas sensors and their associated sensing systems. The reason for overshoot signals in the gas sensor's response characteristic curve is that before the chemical reaction between the gas to be detected and the adsorbed oxygen ions on the surface of the semiconductor material based on oxygen adsorption theory reaches a steady state, the transfer of electrons to the semiconductor material surges, which is a transient enhancement of the chemical reaction. The products of the chemical reaction between the gas to be detected and the adsorbed oxygen ions on the surface of the semiconductor material include water molecules. The large number of water molecules generated instantly accumulate on the surface of the semiconductor material, delaying their escape. This temporarily prevents oxygen in the air from colliding with and re-adsorbing the semiconductor material surface, preventing the rapidly consumed adsorbed oxygen ions from being replenished in a timely manner, causing an overshoot in the sensor signal.

[0003] Regarding the problem of correcting the overshoot signal of the gas sensor response, the smoothing processing methods in the existing technology, including but not limited to the adjacent averaging method, moving average smoothing, median filter and percentile filter, cannot effectively deal with it: First, the existing technology cannot achieve selective processing of the overshoot signal area and the non-overshoot signal area in the response characteristic curve when used, that is, it cannot effectively eliminate the overshoot signal area while retaining the non-overshoot signal area; Second, the existing technology has low robustness and is highly sensitive to parameter selection, which can easily cause distortion of the response characteristic curve, causing the response characteristic curve to be distorted and deviate from the true trend. Since the interference of the overshoot signal on the response characteristic curve of the gas sensor cannot be effectively compensated, when detecting the gas concentration based on the response characteristic curve, the gas concentration detection data will inaccurately deviate from the true value, resulting in measurement deviation, and will also introduce false peak signals, affecting the accuracy of the analysis results. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the present invention aims to provide a gas concentration detection method based on gas sensor response overshoot compensation, comprising:

[0005] Step 1: Obtain a response and recovery characteristic curve of a semiconductor gas sensor for a gas to be detected, wherein the response and recovery characteristic curve is constructed based on at least one of the resistance value, conductance value, current value, voltage value, response rate, and sensitivity of the semiconductor gas sensor; and obtain a curve of a response characteristic interval after the semiconductor gas sensor contacts the gas to be detected from the response and recovery characteristic curve as the response characteristic curve. ,in, i Indicates the i sampling points, n is the total number of sampling points, x i Indicates the response characteristic curve i The horizontal coordinate of the sampling point, that is, i The time of the sampling point, y i Indicates the response characteristic curve i The vertical coordinate of the sampling point, that is, i The characteristic value of each sampling point;

[0006] Step 2: Perform smooth preprocessing on the response characteristic curve to obtain the response characteristic curve after smooth preprocessing ;

[0007] Step 3: Determine whether the smoothed preprocessed response characteristic curve has an overshoot signal. If the smoothed preprocessed response characteristic curve has an overshoot signal, determine the first characteristic point, the second characteristic point, and the third characteristic point of the overshoot signal, and then execute step 4. If the smoothed preprocessed response characteristic curve does not have an overshoot signal, use the smoothed preprocessed response characteristic curve as the final response characteristic curve.

[0008] Step 4: Based on the first characteristic point, the second characteristic point, and the third characteristic point, the response characteristic curve after smoothing preprocessing is reconstructed and smoothed using the Eley-Rideal model to obtain a smoothed reconstructed characteristic curve, which is used as the final response characteristic curve.

[0009] Step 5: Determine the concentration of the gas to be detected based on the final response characteristic curve.

[0010] Optionally, step 3 specifically includes:

[0011] Step 3.1: Calculate the first-order difference of the response characteristic curve after smoothing preprocessing. The first-order difference is expressed as: ;

[0012] Step 3.2: Calculate the product of the first-order differences between two adjacent sampling points , expressed as: ;

[0013] Step 3.3: Determine the interval Is there a sampling point in j , making , in the interval There are no sampling points in j , making In the case of , it is characterized that the response characteristic curve after smooth preprocessing does not have an overshoot signal, and the response characteristic curve after smooth preprocessing is used as the final response characteristic curve;

[0014] In the interval There are sampling points in j , making In the case of j Corresponding is an extreme point, and then multiple extreme points are determined, and it is determined whether the multiple extreme points include both the maximum point and the minimum point. If the multiple extreme points include both the maximum point and the minimum point, step 3.4 is executed. If the multiple extreme points do not include both the maximum point and the minimum point, it is indicated that the response characteristic curve after smoothing preprocessing does not have an overshoot signal, and the response characteristic curve after smoothing preprocessing is used as the final response characteristic curve;

[0015] Step 3.4: Based on the maximum and minimum points, calculate the difference between the overshoot signal peak value and the specified threshold value. I ,judge I Is it greater than or equal to 0? I If it is greater than or equal to 0, it indicates that there is an overshoot signal in the response characteristic curve after smooth preprocessing. Then execute step 3.5. I When it is less than 0, it indicates that there is no overshoot signal in the response characteristic curve after smooth preprocessing, and the response characteristic curve after smooth preprocessing is taken as the final response characteristic curve;

[0016] Step 3.5: Determine the first characteristic point, the second characteristic point, and the third characteristic point of the overshoot signal based on the timing relationship between the maximum point and the minimum point, and then execute step 4.

[0017] Optionally, in step 3.3, determining whether the multiple extreme points include both a maximum point and a minimum point may include:

[0018] For each extreme point, As the center, get the width on the horizontal axis interval;

[0019] If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the maximum point ;

[0020] If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the minimum point ;

[0021] Thus, each extreme point is determined to be a maximum point or a minimum point. After all extreme points are determined, it is determined whether the multiple extreme points include both maximum points and minimum points.

[0022] Optionally, in step 3.4, the difference between the peak value of the overshoot signal and the specified threshold is calculated based on the maximum and minimum points. I , which is specifically achieved through the following formula:

[0023] ;

[0024] in, T is the threshold coefficient, After smoothing preprocessing The characteristic value of the sampling point, is the characteristic value of the 0th sampling point after smoothing preprocessing.

[0025] Optionally, step 3.5 specifically includes:

[0026] exist In this case, As the first feature point, As the second feature point, As the third characteristic point;

[0027] exist In this case, As the first feature point, As the second feature point, As the third characteristic point;

[0028] Step 4 is performed after the first feature point, the second feature point, and the third feature point are determined.

[0029] Optionally, step 4 specifically includes:

[0030] Step 4.1: In the response characteristic curve after smoothing preprocessing, the curve corresponding to the first characteristic point to the third characteristic point is used as the overshoot signal region curve, and the curves other than the overshoot signal region curve in the response characteristic curve after smoothing preprocessing are used as the non-overshoot signal region curve;

[0031] Step 4.2: Based on the Eley-Rideal model, perform reaction kinetic fitting on the non-overshoot signal region curve, determine the fitting coefficient, and then obtain the fitting curve;

[0032] Step 4.3: Substitute the time corresponding to all sampling points in the interval of the overshoot signal area curve into the fitting curve to obtain multiple fitting data points , and for each sampling point in the interval of the overshoot signal region curve, calculating the absolute value of the difference between the characteristic value of the fitting data point and the characteristic value on the overshoot signal region curve, thereby obtaining the absolute values ​​of multiple differences;

[0033] Step 4.4: Obtain the minimum absolute value of all differences, and then determine the sampling point corresponding to the minimum absolute value. Use this sampling point as the intersection of the overshoot signal curve and the fitting curve, and then determine the horizontal and vertical coordinates of the intersection to obtain the intersection point. ;

[0034] Step 4.5: retain the curve between the first characteristic point and the intersection point of the overshoot signal region curve, replace the sampling points between the intersection point and the third characteristic point of the overshoot signal region curve with fitting data points, retain the curve of the non-overshoot signal region, and obtain the reconstructed characteristic curve;

[0035] Step 4.6: Smoothing the reconstructed characteristic curve to obtain a smoothed reconstructed characteristic curve, and using the smoothed reconstructed characteristic curve as the final response characteristic curve.

[0036] Optionally, the Eley-Rideal model described in step 4.2 is expressed as:

[0037] ;

[0038] in, a , b and c represents the fitting coefficient.

[0039] The beneficial effects of adopting the above technical solution are:

[0040] The present invention obtains the response characteristic curve of the semiconductor gas sensor and performs smoothing preprocessing on it, and then determines whether there is an overshoot signal. If so, the overshoot signal is reconstructed and smoothed based on the Eley-Rideal model to obtain the final response characteristic curve. Therefore, the present invention reconstructs and smoothes the response characteristic curve with the overshoot signal through reaction kinetic modeling and signal reconstruction technology, effectively compensating for the interference of the overshoot signal on the response characteristic curve of the semiconductor gas sensor, thereby improving the accuracy of gas concentration detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 1 is a flow chart of a gas concentration detection method based on gas sensor response overshoot compensation in an embodiment of the present invention;

[0042] Figure 2 is the response and recovery characteristic curve of the gas sensor in the embodiment of the present invention;

[0043] Figure 3 Response characteristic curves of the gas sensor in the embodiment of the present invention, wherein (a) is the response characteristic curve of the AB segment, and (b) is the response characteristic curve of the CD segment;

[0044] Figure 4 The results of natural logarithm processing of the current value of the gas sensor and its reaction kinetics fitting curve in the embodiment of the present invention are shown;

[0045] Figure 51 and 2 are the final response characteristic curves in the embodiment of the present invention, wherein (a) is the response characteristic curve of the AB segment after smoothing preprocessing, and (b) is the response characteristic curve of the CD segment after reconstruction and smoothing. DETAILED DESCRIPTION

[0046] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0047] In view of the problems existing in the prior art, the present invention provides a gas concentration detection method based on gas sensor response overshoot compensation, combined with Figure 1 , which may include the following steps:

[0048] Step 1: Obtain a response and recovery characteristic curve of a semiconductor gas sensor for a gas to be detected, wherein the response and recovery characteristic curve is constructed based on at least one of the resistance value, conductance value, current value, voltage value, response rate, and sensitivity of the semiconductor gas sensor; and obtain a curve of a response characteristic interval after the semiconductor gas sensor contacts the gas to be detected from the response and recovery characteristic curve as the response characteristic curve. ,in, i Indicates the i sampling points, n is the total number of sampling points, x i Indicates the response characteristic curve i The horizontal coordinate of the sampling point, that is, i The time of the sampling point, y i Indicates the response characteristic curve i The vertical coordinate of the sampling point, that is, i The characteristic value of each sampling point.

[0049] In the specific implementation process, the present invention takes the SnO2 semiconductor gas sensor detecting ethanol gas under the conditions of 300℃ working temperature, 5V test voltage and 2Hz sampling frequency as an example, and combines Figure 2 , is the response and recovery characteristic curve of the gas sensor to 500 ppm ethanol gas, Figure 2 The vertical coordinate of the characteristic curve is the current value, wherein the AB segment and the CD segment are curves of the response characteristic interval of the gas sensor after contacting the gas to be detected in two independent tests. In the embodiment of the present invention, the AB segment can be selected as the response characteristic curve, that is, Figure 3 (a), we can also select the CD segment as the response characteristic curve, that is, Figure 3 (b). The total number of sampling points of the response characteristic curve is equal to the sampling time (240 s) multiplied by the sampling frequency (2 Hz), that is, nThe value is 480, which gives , let point A of segment AB (similarly, point C of segment CD) be , point B of segment AB (similarly, point D of segment CD) is recorded as .

[0050] It should be noted that, in the embodiment of the present invention Figures 2 to 5 The curves in the graph are composed of discrete sampling points. Because there are many sampling points, the final Figures 2 to 5 It is a continuous curve.

[0051] Step 2: Perform smooth preprocessing on the response characteristic curve to filter out noise interference and obtain the response characteristic curve after smooth preprocessing. ;

[0052] The smoothing preprocessing may be performed by using at least one of Savitzky-Golay filter, local weighted regression, wavelet transform, Gaussian filter and spline interpolation.

[0053] Step 3: Determine whether there is an overshoot signal in the response characteristic curve after smoothing preprocessing. If there is an overshoot signal in the response characteristic curve after smoothing preprocessing, determine the first characteristic point, second characteristic point and third characteristic point of the overshoot signal, and then execute step 4. If there is no overshoot signal in the response characteristic curve after smoothing preprocessing, use the response characteristic curve after smoothing preprocessing as the final response characteristic curve.

[0054] Step 3.1: Calculate the first-order difference of the response characteristic curve after smoothing preprocessing. The first-order difference is expressed as: ;

[0055] Step 3.2: Calculate the product of the first-order differences between two adjacent sampling points , expressed as: ;

[0056] Step 3.3: Determine the interval Is there a sampling point in j , making , in the interval There are no sampling points in j , making In the case of , it is characterized that the response characteristic curve after smooth preprocessing does not have an overshoot signal, and the response characteristic curve after smooth preprocessing is used as the final response characteristic curve;

[0057] In the interval There are sampling points in j , making In the case of j Corresponding is an extreme point, and then multiple extreme points are determined, and it is determined whether the multiple extreme points include both the maximum point and the minimum point. If the multiple extreme points include both the maximum point and the minimum point, step 3.4 is executed. If the multiple extreme points do not include both the maximum point and the minimum point, it is indicated that the response characteristic curve after smoothing preprocessing does not have an overshoot signal, and the response characteristic curve after smoothing preprocessing is used as the final response characteristic curve;

[0058] The process of determining whether multiple extreme points include both maximum and minimum points includes:

[0059] For each extreme point, As the center, get the width on the horizontal axis interval;

[0060] If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the maximum point , which can be expressed as:

[0061] ;

[0062] If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the minimum point , which can be expressed as:

[0063] ;

[0064] Thus, each extreme point is determined to be a maximum point or a minimum point. After all extreme points are determined, it is determined whether the multiple extreme points include both maximum points and minimum points.

[0065] In the specific implementation process, set k Value 7: Combined Figure 3 (a) According to the above calculation, the AB segment does not include both the maximum and minimum points. Therefore, there is no overshoot signal in the response characteristic curve after smooth preprocessing. The response characteristic curve after smooth preprocessing is taken as the final response characteristic curve. Figure 5 (a);

[0066] Combine Figure 3(b) After the above calculation, we can get that the maximum point of CD segment is point E. , The value is 28, and the minimum point is point F , The value is 51, so the response characteristic curve after smoothing preprocessing has both maximum and minimum points. Go to step 3.4.

[0067] Step 3.4: Based on the maximum and minimum points, calculate the difference between the overshoot signal peak value and the specified threshold value. I , which is specifically achieved through the following formula:

[0068] ;

[0069] in, T is the threshold coefficient, After smoothing preprocessing The characteristic value of each sampling point, is the characteristic value of the 0th sampling point after smoothing preprocessing.

[0070] judge I Is it greater than or equal to 0? I If it is greater than or equal to 0, it indicates that there is an overshoot signal in the response characteristic curve after smooth preprocessing. Then execute step 3.5. I When it is less than 0, it indicates that there is no overshoot signal in the response characteristic curve after smooth preprocessing, and the response characteristic curve after smooth preprocessing is taken as the final response characteristic curve;

[0071] Combine Figure 3 (b) In the specific implementation process of the present invention, it is set T The value is 0.1, The value is 3.79×10 -4 , The value is 6.77×10 -4 , I The value is 3.12×10 -4 , I is greater than 0, then proceed to step 3.5.

[0072] Step 3.5: Determine the first characteristic point, the second characteristic point, and the third characteristic point of the overshoot signal based on the timing relationship between the maximum point and the minimum point, and then execute step 4.

[0073] Specifically, in In this case, As the first feature point, As the second feature point, As the third characteristic point;

[0074] exist In this case, As the first feature point, As the second feature point, As the third characteristic point;

[0075] Step 4 is performed after the first feature point, the second feature point, and the third feature point are determined.

[0076] in, The value is 28, The value is 51, which satisfies , then the response characteristic curve will show local maximum points and local minimum points in turn, such as Figure 3 As shown in (b), point C is determined at this time is the first characteristic point, point E The second feature point, point F is the third feature point.

[0077] Step 4: Based on the first characteristic point, the second characteristic point, and the third characteristic point, the response characteristic curve after smoothing preprocessing is reconstructed and smoothed using the Eley-Rideal model to obtain a smoothed reconstructed characteristic curve, which is used as the final response characteristic curve.

[0078] Step 4.1: In the response characteristic curve after smooth preprocessing, the curve corresponding to the first characteristic point to the third characteristic point is used as the overshoot signal area curve, and the curves other than the overshoot signal area curve in the response characteristic curve after smooth preprocessing are used as the non-overshoot signal area curve; Figure 3 (b) The CF segment is the overshoot signal region curve, and the FD segment is the non-overshoot signal region curve.

[0079] Step 4.2: Based on the Eley-Rideal model, perform reaction kinetic fitting on the non-overshoot signal region curve, determine the fitting coefficient, and then obtain the fitting curve;

[0080] The Eley-Rideal model is expressed as:

[0081] ;

[0082] in, a , b and c represents the fitting coefficient.

[0083] That is to say, the present invention combines the Eley-Rideal model with the curve fitting in the non-overshoot signal region to obtain a , b and c The value of Figure 4 The fitting curve shown is aThe value is -7.2834, b The value is -0.7973, c The value is 0.0219.

[0084] Step 4.3: Substitute the time corresponding to all sampling points in the interval of the overshoot signal area curve into the fitting curve to obtain multiple fitting data points , and for each sampling point in the interval of the overshoot signal area curve, calculate the absolute value of the difference between the characteristic value of the fitting data point and the characteristic value on the overshoot signal area curve, and then obtain the absolute values ​​of multiple differences, which can be specifically expressed by the following formula:

[0085] .

[0086] Specifically, The value is 51, that is:

[0087] .

[0088] Step 4.4: Obtain the minimum absolute value of all differences, and then determine the sampling point corresponding to the minimum absolute value. Use this sampling point as the intersection of the overshoot signal curve and the fitting curve, and then determine the horizontal and vertical coordinates of the intersection to obtain the intersection point. , combined with Figure 4 , p Value is 12, G point is the intersection of the overshoot signal and the fitting curve.

[0089] Step 4.5: retain the curve between the first characteristic point and the intersection point of the overshoot signal region curve, replace the sampling points between the intersection point and the third characteristic point of the overshoot signal region curve with fitting data points, retain the curve of the non-overshoot signal region, and obtain the reconstructed characteristic curve;

[0090] Specifically, combined Figure 4 , retain the CG segment curve, replace all sampling points of the GF segment with the corresponding fitting data points, retain the FD segment curve, complete the reconstruction process, and obtain the reconstructed characteristic curve.

[0091] Step 4.6: Smooth the reconstructed characteristic curve to obtain the smoothed reconstructed characteristic curve, and use the smoothed reconstructed characteristic curve as the final response characteristic curve. Figure 5 (b).

[0092] The smoothing process may be performed by using at least one of a neighboring average method and a moving average smoothing method.

[0093] Step 5: Determine the concentration of the gas to be detected based on the final response characteristic curve.

[0094] Specifically, a support vector regression (SVR) model was used to predict gas concentrations. The response characteristic curves of the SnO2 semiconductor gas sensor to 50, 100, 200, 300, 400, and 500 ppm ethanol gas were obtained under the conditions of an operating temperature of 300 °C, a test voltage of 5 V, and a sampling frequency of 2 Hz. Ten repeated experiments were performed for each concentration gradient, and the model was trained using a ten-fold cross-validation method. The kernel function of the SVR model was RBF, the regularization parameter was 200, the insensitive loss parameter was 2, and the mean absolute error of the gas concentration prediction on the test set was 2.42. Figure 3 (b) The response characteristic curve of the gas to be detected with overshoot signal is predicted to be 424.97 ppm; Figure 5 The response characteristic curve in (b) shows a predicted value of 501.86 ppm, which is closer to the true value of 500 ppm. The results show that the reconstruction and smoothing processing of the present invention can effectively improve the accuracy of gas concentration detection.

[0095] Specifically, a back propagation (BP) neural network can also be used as a regression model to predict gas concentration. The response characteristic curves of the SnO2 semiconductor gas sensor to 50, 100, 200, 300, 400 and 500 ppm ethanol gas at an operating temperature of 300 ° C, a test voltage of 5 V and a sampling frequency of 2 Hz are obtained. The experiment is repeated 10 times for each concentration gradient, and the model is trained by a ten-fold cross-validation method. The BP neural network model includes two hidden layers, with 48 and 24 neurons respectively, and the activation function is ReLU. The average absolute error of the gas concentration prediction on the test set is 8.42. For the embodiment of the present invention, Figure 3 (b) The response characteristic curve of the gas to be detected with overshoot signal is predicted to be 463.41 ppm; Figure 5 The response characteristic curve in (b) shows a predicted value of 505.31 ppm, which is closer to the true value of 500 ppm. The results show that the reconstruction and smoothing processing of the present invention can effectively improve the accuracy of gas concentration detection.

[0096] The above description is merely an illustration of the preferred embodiments of the present disclosure and the technical principles employed. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present disclosure is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A gas concentration detection method based on gas sensor response overshoot compensation, characterized in that: include: Step 1: Obtaining a response and recovery characteristic curve of a semiconductor gas sensor for a gas to be detected, wherein the response and recovery characteristic curve is constructed based on at least one of a resistance value, a conductance value, a current value, a voltage value, a response rate, and a sensitivity of the semiconductor gas sensor; In the response and recovery characteristic curve, a curve of the response characteristic interval after the semiconductor gas sensor contacts the gas to be detected is obtained as a response characteristic curve ,in, i Indicates the i sampling points, n is the total number of sampling points, x i Indicates the response characteristic curve i The horizontal coordinate of the sampling point, that is, i The time of the sampling point, y i Indicates the response characteristic curve i The vertical coordinate of the sampling point, that is, i characteristic values ​​of sampling points; Step 2: Perform smooth preprocessing on the response characteristic curve to obtain the response characteristic curve after smooth preprocessing ; Step 3: Determine whether the smoothed preprocessed response characteristic curve has an overshoot signal. If the smoothed preprocessed response characteristic curve has an overshoot signal, determine the first characteristic point, the second characteristic point, and the third characteristic point of the overshoot signal, and then execute step 4. If the smoothed preprocessed response characteristic curve does not have an overshoot signal, use the smoothed preprocessed response characteristic curve as the final response characteristic curve. Step 4: Based on the first characteristic point, the second characteristic point, and the third characteristic point, the response characteristic curve after smoothing preprocessing is reconstructed and smoothed using the Eley-Rideal model to obtain a smoothed reconstructed characteristic curve, which is used as the final response characteristic curve. Step 5: Determine the concentration of the gas to be detected based on the final response characteristic curve; Among them, step 4 specifically includes: Step 4.1: In the response characteristic curve after smoothing preprocessing, the curve corresponding to the first characteristic point to the third characteristic point is used as the overshoot signal region curve, and the curves other than the overshoot signal region curve in the response characteristic curve after smoothing preprocessing are used as the non-overshoot signal region curve; Step 4.2: Based on the Eley-Rideal model, perform reaction kinetic fitting on the non-overshoot signal region curve, determine the fitting coefficient, and then obtain the fitting curve; Step 4.3: Substitute the time corresponding to all sampling points in the interval of the overshoot signal area curve into the fitting curve to obtain multiple fitting data points , and for each sampling point in the interval of the overshoot signal region curve, calculating the absolute value of the difference between the characteristic value of the fitting data point and the characteristic value on the overshoot signal region curve, thereby obtaining the absolute values ​​of multiple differences; Step 4.4: Obtain the minimum absolute value of all differences, and then determine the sampling point corresponding to the minimum absolute value. Use this sampling point as the intersection of the overshoot signal curve and the fitting curve, and then determine the horizontal and vertical coordinates of the intersection to obtain the intersection point. ; Step 4.5: retain the curve between the first characteristic point and the intersection point of the overshoot signal region curve, replace the sampling points between the intersection point and the third characteristic point of the overshoot signal region curve with fitting data points, retain the curve of the non-overshoot signal region, and obtain the reconstructed characteristic curve; Step 4.6: Smoothing the reconstructed characteristic curve to obtain a smoothed reconstructed characteristic curve, and using the smoothed reconstructed characteristic curve as the final response characteristic curve.

2. A gas concentration detection method based on gas sensor response overshoot compensation according to claim 1, characterized in that: Step 3 specifically includes: Step 3.1: Calculate the first-order difference of the response characteristic curve after smoothing preprocessing. The first-order difference is expressed as: ; Step 3.2: Calculate the product of the first-order differences between two adjacent sampling points , expressed as: ; Step 3.3: Determine the interval Is there a sampling point in j , making , in the interval There are no sampling points in j , making In the case of , it is characterized that the response characteristic curve after smooth preprocessing does not have an overshoot signal, and the response characteristic curve after smooth preprocessing is used as the final response characteristic curve; In the interval There are sampling points in j , making In the case of j Corresponding is an extreme point, and then multiple extreme points are determined, and it is determined whether the multiple extreme points include both the maximum point and the minimum point. If the multiple extreme points include both the maximum point and the minimum point, step 3.4 is executed. If the multiple extreme points do not include both the maximum point and the minimum point, it is indicated that the response characteristic curve after smoothing preprocessing does not have an overshoot signal, and the response characteristic curve after smoothing preprocessing is used as the final response characteristic curve; Step 3.4: Based on the maximum and minimum points, calculate the difference between the overshoot signal peak value and the specified threshold value. I ,judge I Is it greater than or equal to 0? I If it is greater than or equal to 0, it indicates that there is an overshoot signal in the response characteristic curve after smooth preprocessing. Then execute step 3.

5. I When it is less than 0, it indicates that there is no overshoot signal in the response characteristic curve after smooth preprocessing, and the response characteristic curve after smooth preprocessing is taken as the final response characteristic curve; Step 3.5: Determine the first characteristic point, the second characteristic point, and the third characteristic point of the overshoot signal based on the timing relationship between the maximum point and the minimum point, and then execute step 4.

3. The gas concentration detection method based on gas sensor response overshoot compensation according to claim 2, characterized in that: In step 3.3, it is determined whether the multiple extreme points include both maximum and minimum points, specifically including: For each extreme point, As the center, get the width on the horizontal axis interval; If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the maximum point ; If in the interval Any sampling point i ,satisfy , and in the interval Any sampling point i ,satisfy , then the extreme point is the minimum point ; Thus, each extreme point is determined to be a maximum point or a minimum point. After all extreme points are determined, it is determined whether the multiple extreme points include both maximum points and minimum points.

4. The gas concentration detection method based on gas sensor response overshoot compensation according to claim 2, characterized in that: In step 3.4, based on the maximum and minimum points, the difference between the overshoot signal peak value and the specified threshold is calculated. I , which is specifically achieved through the following formula: ； in, T is the threshold coefficient, After smoothing preprocessing The characteristic value of the sampling point, is the characteristic value of the 0th sampling point after smoothing preprocessing.

5. The gas concentration detection method based on gas sensor response overshoot compensation according to claim 2, characterized in that: Step 3.5 specifically includes: exist In this case, As the first feature point, As the second feature point, As the third characteristic point; exist In this case, As the first feature point, As the second feature point, As the third characteristic point; Step 4 is performed after the first feature point, the second feature point, and the third feature point are determined.

6. The gas concentration detection method based on gas sensor response overshoot compensation according to claim 1, characterized in that: The Eley-Rideal model described in step 4.2 is expressed as: ； in, a , b and c represents the fitting coefficient.

Images