An eddy current displacement measurement adaptive calibration method

Abstract

The application discloses an electric eddy current displacement measurement adaptive calibration method, and relates to the technical field of precision measurement.The technical problem of interval division fixation of segmented linear interpolation calibration and error caused by temperature change on overall measurement is solved.The application realizes adaptive partitioning through error pre-scanning, matches the interval division with actual nonlinear distribution, combines boundary fusion zones and linear weight smoothing mechanism, completely eliminates interval boundary jump error, reduces overall nonlinear error, separates the temperature influence of a probe and a measured body by using double-frequency excitation technology, combines GA-BP neural network to construct a nonlinear temperature model, adapts to nonlinear drift in a wide temperature range, is compatible with the temperature characteristics of the conductivity of measured bodies of different materials, reduces temperature compensation error, predicts fast temperature fluctuation trend through Kalman filtering, and loads compensation parameters in advance, so that aging drift compensation of a sensor in a whole life cycle can be realized without manual intervention, and temperature calibration cycle and response time are shortened.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of precision measurement, in particular to an eddy current displacement measurement adaptive calibration method. BACKGROUND

[0002] Eddy current displacement measurement technology is widely used in industrial rotating machinery monitoring, aerospace equipment detection, precision manufacturing and assembly due to its advantages of non-contact, fast response, strong anti-pollution ability, etc. The core principle is to induce eddy current on the surface of the measured metal through the alternating magnetic field generated by the probe coil, to use the coil impedance change caused by the eddy current counter magnetic field to establish the mapping relationship between displacement and output voltage, and then to realize displacement measurement.

[0003] To solve the inherent nonlinearity of eddy current displacement measurement, the existing technology often uses piecewise linear interpolation calibration method, which divides the measurement interval and fits a linear model respectively, simplifies the complexity of nonlinearity correction, and considers the real-time performance and engineering practicability. At the same time, for the problems of probe coil parameter drift caused by temperature change and measured body electromagnetic characteristic variation, a linear temperature compensation model is often used to correct the interval parameters by presetting temperature coefficients; for the aging drift of the sensor after long-term use, it usually relies on regular manual recalibration.

[0004] However, the existing technology still has many limitations in practical application: the interval division of piecewise linear interpolation is mostly fixed segmentation, which does not combine with the actual nonlinearity distribution law of the sensor, resulting in too large local error, lack of smooth transition mechanism for interval boundary, and easy to appear displacement jump;

[0005] Temperature compensation only targets the probe, ignores the temperature characteristic differentiation caused by the difference in measured body material, and the linear temperature model cannot adapt to the nonlinearity drift of wide temperature range, the aging drift of the sensor lacks online compensation scheme and needs frequent manual calibration. SUMMARY

[0006] In view of the shortcomings of the prior art, the present application provides an eddy current displacement measurement adaptive calibration method, which solves the problems of fixed interval division of piecewise linear interpolation calibration and error caused by temperature change on the whole measurement.

[0007] To achieve the above purpose, the present application realizes the following technical scheme: an eddy current displacement measurement adaptive calibration method, which specifically comprises the following steps:

[0008] Step one, obtain the calibration points and the number of partitions, and use the uniform distribution and interval boundary point encryption mode to calibrate the calibration points, collect the calibration point data and pre-process to obtain the pre-processed data;

[0009] Step two, based on the obtained pre-processing data, using piecewise linear interpolation method for calibration processing, the linear weight coefficient is calculated, and the fused displacement is calculated according to the linear weight coefficient;

[0010] Step three, based on the fused displacement after piecewise linear interpolation calibration, temperature compensation calibration is carried out combined with the environment temperature, and the temperature displacement correction amount is calculated;

[0011] Step four, the sensor aging is compensated and corrected, and the total displacement is calculated by comprehensively fusing the fused displacement and the temperature displacement correction amount.

[0012] As a further scheme of the present application, the determination method of the number of partitions is:

[0013] The probe of the eddy current sensor is perpendicular to the surface of the standard measured body, and the calibration points are recorded as i, and i=1, 2, …, j, wherein j is the number of calibration points, and the number of calibration points is set by the calibration method, the total number of calibration points j is [3×partition number, 5×partition number], and j is a positive integer;

[0014] The calibration platform uniformly collects 31 pre-scanning points from 0mm to 5mm with a pre-scanning point interval of 0.17mm, and records the real displacement d i -Output voltage X i Data pairs, according to the formula =(X i -X i理想 ) / 5×100%, the original error is calculated, wherein X i理想 =2.0+0.6d i , according to k i =| - | / 0.17, the error rate is calculated, and a reference error rate k0 is set. In the pre-scanning data, the point position with k i >k0 is recorded as a split point, and the final interval is divided by the split point.

[0015] As a further scheme of the present application, the method for collecting calibration point data is:

[0016] The full range is uniformly divided into basic calibration points, and 1 encryption point is added before and after each interval boundary point to form a complete calibration point sequence. 20 groups of data are collected at each calibration point, and the sampling time is 2 seconds. According to the planned calibration point, move in turn, and start collecting after staying at each point for 3 seconds;

[0017] First, forward collection is carried out, from zero point to full scale, the output voltage of each calibration point is collected in turn, and the real displacement d i -Output voltage X i Data pairs are recorded, then reverse collection is carried out, and calibration point data is obtained.

[0018] As a further scheme of the present application, the acquisition method of the pre-processing data is:

[0019] Calculate the output voltage X of each calibration point i The average value of the forward and reverse acquisition X i =(X i正 +X i反 ) / 2, calculate the standard deviation σ of 20 groups of original data of each calibration point, and determine the data deviating from the average value within the group by more than 3σ as an abnormal value, if the proportion of abnormal values is ≤5%, then use the remaining valid data to recalculate X i ; if the proportion of abnormal values is >5%, reacquire the data of the calibration point.

[0020] As a further scheme of the present application, the calculation method of the linear weight coefficient is:

[0021] For each interval of d i -X i Data pair, fit the linear equation d=a k X+b k by the least square method, where a k is the slope and b k is the intercept, according to the formula =0.1×interval voltage range, the fusion bandwidth is calculated and the demarcation point voltage X k of each fusion band is determined, and the linear weight coefficient w is calculated according to the linear weight formula w=(X k + -X) / (2 ).

[0022] As a further scheme of the present application, the calculation method of the fused displacement according to the linear weight coefficient is:

[0023] According to the formula =w×d k +(1-w)d k+1 , the fused displacement is calculated, where d k is the displacement calculation value of the current interval, and d k+1 is the displacement calculation value of the adjacent interval.

[0024] When X is close to the demarcation point X k of the current interval, the linear weight coefficient w≈1, ≈d k , when X is close to the adjacent interval, the linear weight coefficient w≈0, ≈d k+1 , when X=X kAt that time, the linear weighting coefficient w≈0.5, =0.5d k +0.5d k+1 .

[0025] As a further aspect of the present invention, the specific method for the temperature displacement correction amount is as follows:

[0026] Obtain the partitioned interval and the corresponding baseline interval coefficient 'a'. 00 =4.917, b 00 =9.834, the test object is placed in a constant temperature chamber and stabilized to the target temperature. A dual-frequency excitation probe is used, and the displacement stage is moved to the full-range calibration point. The high-frequency signal output voltage X is simultaneously acquired. f1 Low-frequency signal output voltage X f2 and probe temperature T probe Temperature T of the measured body work and generate T probe -T work -X f1 -X f2 - Dataset of the actual displacement d;

[0027] The dataset is mapped to temperature-interval coefficients, and the interval coefficients at each temperature are fitted using the least squares method. The GA-BP model is then used for training to obtain a nonlinear mapping model of temperature-interval coefficients for the test body at different temperatures.

[0028] For test subjects made of different materials, a temperature coefficient of conductivity that fits the test subject is constructed. Establish a material-temperature coefficient comparison table and apply the formula. The corresponding temperature displacement correction amount is calculated. , where X f2-20 This indicates the low-frequency voltage of the tested object at 20℃.

[0029] As a further aspect of the present invention, the total displacement is calculated as follows:

[0030] According to formula d 总 = ·X+b k (T)+ ,in The slope after sensor aging compensation is calculated as follows:

[0031] Multi-point calibration is performed at the reference temperature to obtain the reference slope 'a'. k0 At the same time, obtain the actual slope 'a' corresponding to the actual usage time. kt According to the formula =a kt -a k0 The aging slope offset was calculated. Then, the output of the GA-BP nonlinear model is a. k (T) Calculate the aging compensation slope according to the formula. =a k (T)+ The aging compensation slope was calculated.

[0032] As a further aspect of the present invention, the slope calculation method after sensor aging compensation also includes:

[0033] At the reference temperature, which is set to 20°C, the reference slope 'a' is obtained through multi-point calibration. k0 At different usage time points (t1, t2, t3..., t... n Recalibrate to obtain the corresponding actual slope a k (t1), a k (t2), ..., according to the formula =a k (t) n )-a k0 And fit using the least squares method =C×t m The parameters C and m are obtained by solving, where C is the aging constant, m is the aging exponent, and t is the sensor usage time. Then, the obtained parameters are substituted into the formula. =C×t m .

[0034] This invention provides an adaptive calibration method for eddy current displacement measurement. Compared with existing technologies, it has the following advantages:

[0035] This invention achieves adaptive partitioning through error pre-scanning, ensuring that the interval division matches the actual nonlinear distribution. Combined with boundary fusion bands and linear weight smoothing mechanisms, it completely eliminates interval boundary jump errors and reduces full-range nonlinear errors. It employs dual-frequency excitation technology to separate the temperature influence of the probe and the measured object, and combines a GA-BP neural network to construct a nonlinear temperature model that adapts to nonlinear drift over a wide temperature range. It is also compatible with the conductivity and temperature characteristics of measured objects of different materials, reducing temperature compensation errors. By predicting rapid temperature fluctuation trends through Kalman filtering and pre-loading compensation parameters, it can achieve aging drift compensation throughout the sensor's entire life cycle without manual intervention, shortening the temperature calibration cycle and response time. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating the steps of the adaptive calibration method for eddy current displacement measurement according to the present invention. Detailed Implementation

[0037] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0038] Please refer to Figure 1 The present application provides an electric eddy current displacement measurement adaptive calibration method, which specifically comprises the following steps:

[0039] Step one, the probe of the electric eddy current sensor is perpendicular to the surface of the standard measured object, and the calibration points are recorded as i, and i = 1, 2, …, j, wherein j is the number of calibration points, and the number of calibration points is set by the calibration method, the total number of calibration points j is [3×partition number, 5×partition number], and j is a positive integer, wherein the partition number is obtained by uniformly collecting 31 pre-scanning points from 0 mm to 5 mm with the calibration platform, and each pre-scanning point is spaced 0.17 mm apart, and the real displacement d i -Output voltage X i Data pairs, according to the formula = (X i -X i理想 ) / 5×100%, the original error is calculated , wherein X i理想 = 2.0 + 0.6d i , and the error change rate k i =| - | / 0.17, and a reference error rate k0 is set, the point with k i >k0 is recorded as a split point, and the final interval is divided based on the split point;

[0040] Then, a uniform distribution + interval split encryption mode is adopted, and the basis calibration points are uniformly divided in the full range, and 1 encryption point is added before and after each interval split point to form a complete calibration point sequence, for example: 0-5mm range, 3 segment partition, calibration points are 0, 0.85, 1.7, 2.55, 3.3, 4.15, 5mm, wherein 1.7mm and 3.3mm are interval split points, and 0.85mm, 2.55mm and 4.15mm are encryption points, 20 groups of data are collected for each calibration point, the sampling time is 2 seconds, and the planned calibration points are moved in turn, each point stays for 3 seconds before starting to collect, and first, the output voltage of each calibration point is collected in turn from zero point to full range, and the real displacement d i -Output voltage X iData pairs, then reverse collection, from full scale to zero, and repeat the operation, calculate the output voltage X of each calibration point i Average value of forward and reverse collection X i =(X i正 +X i反 ) / 2, using 3σ criterion to eliminate outliers, get the pretreatment data, and the specific processing method is as follows:

[0041] Calculate the standard deviation σ of 20 groups of original data of each calibration point, and determine the data deviating from the average value of the group by more than 3σ as an outlier. If the proportion of outliers is ≤5%, recalculate X with the remaining valid data i ; If the proportion of outliers is > 5%, re-collect the data of the calibration point;

[0042] Step two, based on the obtained pretreatment data, use piecewise linear interpolation method for calibration processing, for each interval of d i -X i Data pairs are fitted with linear equation d=a k X+b k , where a k is the slope, b k is the intercept, for example, interval [0, 1.2mm], the corresponding fitting data points (0, 2.0000), (0.5998, 2.3002), (1.1999, 2.7201), the slope is 4.917, the intercept is-9.834, interval [1.2mm, 2.8mm], the corresponding fitting data points (1.1999, 2.7201), (1.7997, 2.9803), (2.3998, 3.2402), (2.7999, 3.4801), the slope is 4.892, the intercept is-9.756, interval [2.8mm, 5mm], the corresponding fitting data points (2.7999, 3.4801), (3.5997, 3.7603), (4.2998, 4.0802), (4.9999, 4.4801), the slope is 4.868, the intercept is-9.682, then set the boundary fusion zone, according to the formula =0.1×interval voltage range, the fusion zone width is calculated , at the same time, the fusion zone is divided, interval 0 and 1 fusion zone, X [2.62V, 2.82V], the dividing point X0=2.7201V, interval 1 and 2 fusion zone, X [3.38V, 3.58V], the dividing point X1=3.4801V, at the same time, according to the linear weight formula w=(X k + -X) / (2 The linear weighting coefficients are calculated, where X k The voltage value corresponding to the interval boundary point, where X is the current real-time measured voltage value;

[0043] Next, the fused displacement is calculated based on the obtained linear weighting coefficients, according to the formula. =w×d k +(1-w)d k+1 The displacement after fusion was calculated. , where d k d represents the calculated displacement value for the current interval. k+1 This is the displacement calculation value between adjacent intervals, specifically when X is close to the boundary point X of the current interval. k At that time, the linear weighting coefficient w≈1, ≈d k When X is close to an adjacent interval, the linear weighting coefficient w≈0. ≈d k+1 When X=X k At that time, the linear weighting coefficient w≈0.5, =0.5d k +0.5d k+1 ;

[0044] Based on practical analysis, taking the 0-5mm range as an example, the voltage X at the boundary between interval 0 and 1 is... k =2.7201V, fusion bandwidth =0.1V, the current measured voltage X=2.77V (within the fusion band), according to the linear weighting formula w=(X k + -X) / (2 The calculation yields w≈0.2505. Assuming the displacement d in the current interval is 0... k =1.25mm, displacement d of adjacent interval 1 k+1 =1.30mm, then further according to the formula =w×d k +(1-w)d k+1 The displacement after fusion was calculated. ≈1.287mm.

[0045] Step 3: Based on the linearly calibrated displacement, and combined with the ambient temperature, the displacement is calibrated. First, piecewise linear interpolation calibration is performed at the reference temperature to obtain the divided intervals and the corresponding reference interval coefficients 'a'. 00 =4.917, b 00=-9.834, taking the interval 0 as an example, data is collected on the test object at different temperatures. Specifically, the test object is placed in a constant temperature chamber and stabilized to the target temperature, which includes 10℃ / 20℃ / 30℃ / 40℃ / 50℃. A dual-frequency excitation probe is used, and the displacement stage is moved to the full-range calibration point, and the high-frequency signal output voltage X is collected simultaneously. f1 Low-frequency signal output voltage X f2 and probe temperature T probe Temperature T of the measured body work and generate T probe -T work -X f1 -X f2 - Dataset of the actual displacement d;

[0046] The dataset is mapped to a temperature-interval coefficient, and the interval coefficient at each temperature is fitted using the least squares method. The model is then trained using a GA-BP model, with the probe temperature T input to the input layer. probe Temperature T of the measured body work The output layer outputs interval coefficients, resulting in a nonlinear mapping model of temperature-interval coefficients for the test object at different temperatures. For test objects of different materials, a temperature coefficient fitting the conductivity is constructed. Establish a material-temperature coefficient comparison table and apply the formula. The corresponding temperature displacement correction amount is calculated. , where X f2-20 This indicates the low-frequency voltage of the tested object at 20℃.

[0047] Based on practical analysis, taking the real-time measurement of a 45# steel workpiece, with a current probe temperature of 32℃, a measured body temperature of 30℃, and a measurement voltage X=2.3002V as an example, the high-frequency signal output voltage X was acquired. f1 =2.302V, low-frequency signal output voltage X f2 =2.298V, probe temperature T probe =32℃, the temperature of the measured object T work =30℃, based on nearly 10 temperature sampling points, predict T after 50ms. probe-pred =33℃, the obtained T probe-pred Input the trained GA-BP model and output the coefficients in the interval 0, a0(33)=GA-BP(T probe-pred )=5.021, b0(33)=GA-BP(T probe-pred = -9.412;

[0048] Calculate the displacement correction caused by the temperature of the measured object based on the low-frequency signal output voltage and the material of the measured object. K f2 To represent the sensitivity to low-frequency signals, substitute the above parameters to obtain... ≈-0.004mm.

[0049] Step four, based on the linear calibration displacement and temperature calibration displacement, calculate the total displacement amount, according to the formula d 总 = ·X+b k (T)+ , wherein is the slope of the sensor aging compensation, the specific calculation method is as follows:

[0050] At the reference temperature, multi-point calibration is carried out to obtain the reference slope a k0 , and the actual slope a kt corresponding to the actual use time is obtained at the same time, according to the formula =a kt -a k0 , the aging slope offset a is calculated, and then the a k (T) output by the GA-BP nonlinear model is integrated to calculate the aging compensation slope, according to the formula =a k (T) + , the aging compensation slope is calculated.

[0051] Second embodiment

[0052] As a second embodiment of the application, it is implemented on the basis of the first embodiment, and the difference from the first embodiment is that the calculation method of the aging compensation slope can also be calculated by the following method:

[0053] At the reference temperature, and the reference temperature is set to 20℃, the reference slope a k0 is obtained by multi-point calibration, and the actual slope a n (t1), a k (t2), … is obtained by recalibration at different use time points (t1, t2, t3…, t k ), according to the formula =a k (t n )-a k0 , and through the least square fitting =C×t m , the parameters C and m are solved, wherein C is the aging constant, m is the aging index, t is the use time of the sensor, and then the obtained parameters are substituted into the formula =C×t m .

[0054] Third embodiment

[0055] As a third embodiment of the present application, the implementation process of the first and second embodiments is combined.

[0056] The data in the above formula are all dimensionless for numerical calculation, and the contents not described in detail in the specification all belong to the prior art known by those skilled in the art.

[0057] The above embodiments are only used to illustrate the technical method of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. An electroviscous displacement measurement adaptive calibration method, characterized by, The method specifically comprises the following steps: Step one, obtaining calibration points and the number of partitions, and using uniform distribution and interval boundary point encryption mode for calibration point calibration, while collecting calibration point data and preprocessing to obtain preprocessed data; Step two, based on the obtained preprocessed data, using piecewise linear interpolation method for calibration processing, calculating the linear weight coefficient, and calculating the fused displacement according to the linear weight coefficient; Step three, based on the fused displacement after piecewise linear interpolation calibration, combining the environmental temperature for temperature compensation calibration, and calculating the temperature displacement correction amount; Step four, compensating and correcting the sensor aging, and calculating the total displacement amount by comprehensively fusing the fused displacement and the temperature displacement correction amount.

2. The method of claim 1, wherein the method further comprises: The determination method of the number of partitions is: The probe of the eddy current sensor is perpendicular to the surface of the standard measured body, and the calibration points are obtained as i, and i=1, 2, …, j, wherein j is the number of calibration points, and the number of calibration points is set by the calibration method, the total number of calibration points j is [3×partition number, 5×partition number], and j is a positive integer; The calibration platform uniformly collects 31 pre-scanning points from 0mm to 5mm with a pre-scanning point interval of 0.17mm, and records the real displacement d i -Output voltage X i Data pairs, according to the formula = (X i - X i理想 ) / 5x100%, the original error is calculated, wherein X i理想 = 2.0+0.6d i , the error rate is calculated according to k i =| - | / 0.17, and the reference error rate k0 is set. In the pre-scanning data, the point position with k i >k0 is recorded as a split point, and the final interval is divided by the split point.

3. The method of claim 1, wherein the method further comprises: The data acquisition method of the calibration point is: The full range is uniformly divided into basic calibration points, and one encryption point is added before and after each interval boundary point to form a complete calibration point sequence, 20 groups of data are collected for each calibration point, the sampling time is 2 seconds, and the calibration points are moved in turn according to the planned calibration points, and the total displacement amount is calculated by comprehensively fusing the fused displacement and the temperature displacement correction amount; First, forward acquisition is performed, from zero point to full scale, output voltage of each calibration point is acquired in turn, and real displacement d is recorded i - output voltage X i Data pairs, then reverse acquisition is performed, and calibration point data is obtained.

4. The method of claim 1, wherein the method further comprises: The preprocessed data acquisition method is: Calculate the output voltage X of each calibration point i Average value of forward and reverse collection X i = (X i正 + X i反 ) / 2, calculate the standard deviation σ of 20 groups of original data for each calibration point, and determine the data deviating from the average value within the group by more than 3σ as abnormal value, if the proportion of abnormal value is ≤5%, then use the remaining valid data to recalculate X i ; if the proportion of abnormal value is > 5%, re-collect the data of the calibration point.

5. The method of claim 1, wherein the method further comprises: The calculation method of the linear weight coefficient is: For each interval d i -X i The data pairs were fitted to the linear equation d=a using the least squares method. k X+b k , where a k Let b be the slope. k The intercept is calculated according to the formula. =0.1 × interval voltage range, the fusion band width is calculated. And determine the boundary voltage X of each fusion band. k Meanwhile, according to the linear weight formula w = (X k + -X) / (2 The linear weighting coefficient w is calculated.

6. The method of claim 1, wherein the method further comprises: The method for calculating the fused displacement according to the linear weight coefficient is: According to the formula = w x d k + (1 - w) d k+1 , the displacement after fusion is calculated , wherein d k is the displacement calculation value of the current interval, and d k+1 is the displacement calculation value of the adjacent interval; When X is close to the boundary point X k of the current interval, the linear weight coefficient w ≈ 1, ≈ d k When X is close to the adjacent interval, the linear weight coefficient w ≈ 0, ≈ d k+1 When X = X k , the linear weight coefficient w ≈ 0.5, = 0.5d k + 0.5d k+1 .

7. The method of claim 1, wherein the method further comprises: The specific method of the temperature displacement correction amount is: Acquiring the division interval and acquiring the corresponding reference interval coefficient a 00 = 4.917, b 00 = -9.834, the measured body into the constant temperature chamber, stable to the target temperature, using double frequency excitation probe, and move the displacement table to the full range calibration point, synchronous acquisition high frequency signal output voltage X f1 , low frequency signal output voltage X f2 and probe temperature T probe , the measured body temperature T work , and generate T probe -T work -X f1 -X f2 - the data set of real displacement d; The data set is mapped according to the temperature-interval coefficient, the interval coefficient at each temperature is fitted by the least square method, and the GA-BP model is trained to obtain the nonlinear mapping model of the measured body temperature-interval coefficient at different temperatures; For test subjects made of different materials, a temperature coefficient of conductivity that fits the test subject is constructed. Establish a material-temperature coefficient comparison table and apply the formula. The corresponding temperature displacement correction amount is calculated. , where X f2-20 This indicates the low-frequency voltage of the tested object at 20℃.

8. The method of claim 1, wherein the method further comprises: The calculation method of the total displacement amount is: According to the formula d 总 = ·X+b k (T) + wherein is the slope after sensor aging compensation, which is calculated as follows: Multi-point calibration is performed at the reference temperature to obtain the reference slope 'a'. k0 At the same time, obtain the actual slope 'a' corresponding to the actual usage time. kt According to the formula =a kt -a k0 The aging slope offset was calculated. Then, the output of the GA-BP nonlinear model is a. k (T) Calculate the aging compensation slope according to the formula. =a k (T)+ The aging compensation slope was calculated.

9. An electroviscous displacement measurement self-adapting calibration method according to claim 8, characterized in that, The calculation method of the slope after sensor aging compensation further comprises: At a reference temperature, and the reference temperature is set to 20℃, the reference slope a is obtained by multi-point calibration k0 , at different use time points (t1, t2, t3…, t n ) re-calibration, the corresponding actual slope a is obtained k (t1), a k (t2), …, according to the formula =a k (t n ) - a k0 , and the least square method is fitted =C×t m , the parameters C and m are solved, wherein C is an aging constant, m is an aging index, t is the use time of the sensor, and then the obtained parameters are substituted into the formula =C×t m .

Images