Trace oxygen sampling control method

Abstract

This invention discloses a trace oxygen sampling control method, including: calculating the complex impedance modulus, phase angle, and time derivative based on real and imaginary impedance data; obtaining the double-layer perturbation index by weighted fusion of the impedance change rate and phase stability factor; adaptively determining the pressure fluctuation amplitude based on the double-layer perturbation index to protect sensor stability and improve replacement efficiency; and generating a micro-pressure differential modulation control parameter based on the pressure fluctuation amplitude. This invention solves the technical problem of traditional fixed-parameter control methods that cannot balance sensor protection and rapid response, and achieves intelligent and precise control of the trace oxygen sampling process.

Description

Technical Field

[0001] The invention belongs to the field of fluid dynamics, in particular to a trace oxygen sampling control method. Background Art

[0002] Trace oxygen analysis plays a vital role in high-end manufacturing fields such as semiconductor manufacturing, high-purity gas production, and metal heat treatment. During semiconductor manufacturing, even trace amounts of oxygen contamination can lead to serious problems such as wafer oxidation and device performance degradation, necessitating the control of the oxygen content in process gases below 1 ppm. Electrochemical oxygen sensors, due to their high sensitivity and selectivity, have become the mainstream technology for trace oxygen detection. However, in practical applications, achieving rapid response while maintaining measurement accuracy has always been a technical challenge in this field. In particular, in online monitoring systems with multiple gas path switching, the efficiency of residual gas replacement within the sensor cavity directly impacts the accuracy and real-time nature of the measurement.

[0003] Existing trace oxygen sampling control technologies mainly adopt fixed parameter control strategies. In terms of gas path switching, a constant flow rate carrier gas purge method is usually adopted, and the purge time is extended to ensure sufficient replacement of the gas in the cavity. Some advanced systems have introduced pulse purge technology, which applies a pulsed airflow of fixed frequency and intensity at the initial stage of switching to accelerate the replacement of dead zone gas. In terms of sensor protection, existing technologies mostly use passive protection measures such as current limiting and pressure limiting to avoid impact on the sensor by setting fixed flow and pressure upper limits. In terms of control parameter optimization, the optimal purge time, pulse frequency and other parameters are mainly determined based on empirical values ​​or offline calibration, and remain unchanged during actual operation.

[0004] However, existing technologies have obvious technical defects in practical applications. First, although pulse purging with a fixed intensity can accelerate gas replacement, high-intensity pressure pulses will disturb the double-layer structure of the electrochemical sensor, causing sensor baseline drift and increased noise. This disturbance is particularly prominent in trace oxygen measurement, because small changes in the double layer will cause measurement deviations at the ppm level. Secondly, existing methods fail to monitor the electrochemical state of the sensor in real time and cannot dynamically adjust control parameters according to the stability of the double layer. As a result, the same control strategy is used when the sensor is in different working states, which may cause excessive disturbances and insufficient replacement. Summary of the Invention

[0005] The purpose of the invention is to provide a trace oxygen sampling control method to solve at least one technical problem existing in the prior art.

[0006] The technical solution, trace oxygen sampling control method, includes:

[0007] Obtaining the real impedance data and imaginary impedance data of the sensor;

[0008] Calculate the complex impedance modulus and phase angle time derivative based on the real impedance data and the imaginary impedance data;

[0009] Calculating the rate of change of the complex impedance modulus relative to a reference value, and calculating the phase stability factor based on the time derivative of the phase angle;

[0010] The double layer perturbation index is calculated by weighted fusion of the change rate and phase stability factor;

[0011] Based on the double layer disturbance index, the pressure fluctuation amplitude is dynamically determined, and the pressure fluctuation amplitude is adaptively adjusted according to the disturbance degree within a preset range;

[0012] The micro-pressure difference modulation control parameters are generated according to the pressure fluctuation amplitude and used to control the pressure modulation during the sampling process.

[0013] Beneficial effect: The present invention avoids permanent damage caused by traditional high-intensity pulses, solves the technical problem that traditional fixed parameter control methods cannot balance sensor protection and rapid response, and realizes intelligent and precise control of the trace oxygen sampling process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flowchart of the steps of a trace oxygen sampling control method provided in an embodiment of the present application.

[0015] Figure 2 A flowchart of the steps for calculating the phase stability factor provided in an embodiment of the present application.

[0016] Figure 3 A flowchart of the steps for calculating the double layer perturbation index provided in an embodiment of the present application.

[0017] Figure 4 A flow chart of the steps for dynamically determining the pressure fluctuation amplitude provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0019] It should be noted that to clearly illustrate the steps of this application, serial numbers are assigned to each step in the specification. These serial numbers are for illustrative purposes only and do not limit the order in which the steps must be executed. In actual operation, depending on the technical requirements of the specific implementation scenario, the steps may be executed in a different order than shown in the specification, and in some cases, parallel processing between steps may be implemented.

[0020] The study found that fixed pulse timing cannot adapt to the replacement requirements under different concentration gradients. When switching from high concentration to ultra-low concentration, the diffusion rate of residual gas is closely related to the concentration gradient. The existing technology ignores this dynamic characteristic, resulting in longer response time or decreased measurement accuracy.

[0021] like Figure 1 As shown, a trace oxygen sampling control method is proposed, comprising the following steps:

[0022] Obtaining the real impedance data and imaginary impedance data of the sensor;

[0023] Based on the real and imaginary impedance data, the complex impedance modulus and phase angle time derivative are calculated. The rate of change of the complex impedance modulus relative to the reference value is calculated based on the complex impedance modulus, and the phase stability factor is calculated based on the phase angle time derivative. The double layer perturbation index is calculated by weighted fusion of the rate of change and the phase stability factor.

[0024] Dynamically determine the pressure fluctuation amplitude based on the double layer perturbation index;

[0025] Generate micro-pressure difference modulation control parameters according to the pressure fluctuation amplitude.

[0026] According to one aspect of the present application, obtaining real impedance data and imaginary impedance data of a sensor includes:

[0027] Obtain the original oxygen concentration voltage signal, impedance real part data, impedance imaginary part data and system timestamp from the system, perform timestamp alignment correction on the multi-channel data to compensate for acquisition delay, ensure data synchronization for subsequent processing, and output synchronized sensor data set;

[0028] Signal preprocessing was performed on the synchronized sensor data set: the oxygen concentration signal was processed using an adaptive Kalman filter, with the filter parameters adjusted in real time based on the signal change rate; the impedance data was filtered using a moving window median filter (window size: 5 sampling points) to remove pulse interference, obtaining the filtered oxygen concentration signal and purified impedance data;

[0029] The filtered oxygen concentration signal is converted into a three-point calibration interpolation through the pre-stored sensor calibration coefficient matrix. The calculation formula is C_O2 = Cal_Matrix[0] + Cal_Matrix[1] × V_O2_filtered + Cal_Matrix[2] × V_O2_filtered 2 , and obtain the standardized oxygen concentration signal C_O2_std, while passing the purification impedance data to the next processing link. Among them, Cal_Matrix[0], Cal_Matrix[1], and Cal_Matrix[2] are the first, second, and third coefficients of the sensor calibration coefficient matrix respectively, and V_O2_filtered is the filtered oxygen concentration signal.

[0030] like Figure 2 As shown, according to one aspect of the present application, calculating the phase stability factor includes:

[0031] The absolute value of the time derivative of the phase angle is divided by a preset phase change threshold to obtain a normalized phase change rate;

[0032] Calculate the difference between 1 and the normalized phase change rate to obtain the phase stability factor;

[0033] Among them, when the absolute value of the time derivative of the phase angle is larger, the phase stability factor is smaller, which means that the double layer stability is worse.

[0034] like Figure 3 As shown, according to one aspect of the present application, calculating the double layer perturbation index includes:

[0035] Multiplying the absolute value of the change rate by the first weight coefficient to obtain a first weighted result;

[0036] Multiplying the difference between 1 and the phase stability factor by the second weight coefficient to obtain a second weighted result;

[0037] Adding the first and second weighted results to obtain the double layer perturbation index;

[0038] The first weight coefficient is greater than the second weight coefficient, so that the impedance amplitude change plays a dominant role in the disturbance assessment.

[0039] Specifically, the complex impedance modulus |Z|= sqrt(Z_real_clean) is calculated based on the clean impedance data {Z_real_clean, Z_imag_clean} 2 + Z_imag_clean 2) and phase angle φ = arctan(Z_imag_clean / Z_real_clean). The time derivative of the phase angle dφ / dt is calculated using a numerical differentiation algorithm, and an impedance feature vector Z_feature = [|Z|, φ, dφ / dt] is constructed, which contains the impedance amplitude, phase information, and phase change rate. Where Z_real_clean is the cleaned real impedance data, Z_imag_clean is the cleaned imaginary impedance data, t is time, and d is the differential sign. To address the problem that traditional impedance analysis methods ignore phase stability, an improved double-layer disturbance detection algorithm is developed: the impedance change rate ΔZ_rate = (|Z|_current - |Z|_baseline) / |Z|_baseline within the sliding window is calculated, where |Z|_current is the complex impedance modulus at the current moment and |Z|_baseline is the baseline impedance modulus. Then, a phase stability factor Phase_stability = 1 - |dφ / dt| / φ_threshold is introduced, where φ_threshold is the preset threshold and |dφ / dt| / φ_threshold is the normalized phase change rate. Finally, a weighted fusion method is used to calculate the comprehensive disturbance index Disturbance_Index = α× |ΔZ_rate| +β× (1 - Phase_stability), where α=0.6 and β=0.4 are weight coefficients. The algorithm is optimized and determined based on a large amount of experimental data to output the double-layer disturbance index Disturbance_Index.

[0040] like Figure 4 As shown, according to one aspect of the present application, dynamically determining the pressure fluctuation amplitude includes:

[0041] Obtaining the time derivative of the oxygen concentration signal as concentration gradient data;

[0042] Determine the sensor stability state according to the double layer disturbance index and set the corresponding state correction factor;

[0043] Based on the concentration gradient data, the gradient response factor is calculated to respond to the concentration change rate;

[0044] The disturbance compensation factor is calculated based on the double layer disturbance index. The larger the double layer disturbance index, the smaller the disturbance compensation factor.

[0045] The pressure fluctuation amplitude is obtained by multiplying the preset basic pressure difference value with the state correction factor, gradient response factor and disturbance compensation factor.

[0046] According to one aspect of the present application, determining the sensor stability state according to the double layer perturbation index includes:

[0047] When the double layer perturbation index is less than the first threshold, it is determined to be in a stable state;

[0048] When the double layer disturbance index is greater than or equal to the first threshold and less than the second threshold, it is determined to be a slight disturbance state;

[0049] When the double layer disturbance index is greater than or equal to the second threshold, it is determined to be a significant disturbance state;

[0050] Among them, the second threshold is 2 to 4 times the first threshold, and the state correction factors corresponding to different states decrease in sequence.

[0051] Specifically, the double layer disturbance index Disturbance_Index is judged to have a three-level threshold value: when Disturbance_Index < 0.05, it is judged to be in a stable state (Status = 1); when 0.05 ≤ Disturbance_Index < 0.15, it is judged to be a slight disturbance (Status = 2); when Disturbance_Index ≥ 0.15, it is judged to be a significant disturbance (Status = 3), and the sensor stability status code Sensor_Status is output.

[0052] According to one aspect of the present application, calculating the gradient response factor comprises:

[0053] The absolute value of the concentration gradient data is divided by the preset concentration change threshold to obtain a normalized gradient value;

[0054] The normalized gradient value is multiplied by the gradient response coefficient and then added to 1 to obtain the initial response factor;

[0055] An upper limit value is set for the initial response factor to obtain a gradient response factor;

[0056] Among them, the faster the oxygen concentration changes, the larger the gradient response factor, which increases the pressure fluctuation amplitude accordingly to accelerate the replacement process.

[0057] According to one aspect of the present application, calculating the disturbance compensation factor includes:

[0058] Multiply the double layer perturbation index by the compensation coefficient to obtain the compensation amount;

[0059] Calculate the difference between 1 and the compensation amount to obtain the disturbance compensation factor;

[0060] Among them, the compensation coefficient is less than 1, so that when the double layer perturbation index is larger, the disturbance compensation factor is smaller, thereby reducing the pressure fluctuation amplitude to protect the stability of the sensor.

[0061] According to one aspect of the present application, calculating the gradient response factor based on the concentration gradient data further includes introducing dynamic compensation for the acceleration of concentration change:

[0062] Numerically differentiate the time derivative of the oxygen concentration signal to obtain the second-order derivative of the oxygen concentration;

[0063] The absolute value of the second-order derivative is divided by a preset acceleration threshold to obtain a normalized acceleration value;

[0064] The change acceleration factor is calculated according to the normalized acceleration value. When the absolute value of the second-order derivative is greater than zero, the change acceleration factor is greater than 1.

[0065] The gradient response factor is multiplied by the change acceleration factor to obtain the enhanced gradient response factor that takes into account the concentration change trend.

[0066] Specifically, the standardized oxygen concentration signal C_O2_std is subjected to gradient analysis, and the first-order derivative dC / dt = (C_O2_std(t) - C_O2_std(t-Δt)) / Δt and the second-order derivative d 2 C / dt 2 = (dC / dt(t) -dC / dt(t-Δt)) / Δt, construct the concentration gradient feature Gradient_Feature = [dC / dt, d 2 C / dt 2 , |d 2 C / dt 2|], where Δt is the time interval. Based on the significant differences in the effectiveness of the traditional fixed differential pressure method under different sensor states, a multi-factor coupled adaptive differential pressure algorithm is constructed: the base differential pressure P_base = 0.8% is set as the sensor protection benchmark. The state correction factor Status_Factor is determined based on the sensor stability status code Sensor_Status. The corresponding coefficients for stable, slightly disturbed, and significantly disturbed states are 1.0, 0.7, and 0.4, respectively. Status_Factor = {1.0, 0.7, 0.4}, corresponding to Status = {1, 2, 3}; the gradient response factor Gradient_Factor = min(2.0, 1.0 + 0.5 × |dC / dt| / C_threshold) is calculated to respond to the concentration change rate, where C_threshold is the concentration change threshold and 0.5 is the gradient response coefficient; and a disturbance compensation factor Disturbance_Factor = 1.0 - 0.3×Disturbance_Index, where Disturbance_Index is the double layer disturbance index and 0.3 is the compensation coefficient. The final pressure difference amplitude P_amplitude = P_base × Status_Factor × Gradient_Factor × Disturbance_Factor, and is limited to the range of 0.5% to 2.0%. P_amplitude = max(0.5%, min(2.0%, P_amplitude)), where P_base is the basic pressure difference value, outputs the optimal pressure fluctuation amplitude P_amplitude.

[0067] According to one aspect of the present application, generating a micro-pressure differential modulation control parameter according to the pressure fluctuation amplitude includes:

[0068] The initial pulse interval is calculated based on the absolute value of the concentration gradient data. The larger the concentration gradient data, the smaller the initial pulse interval.

[0069] Based on the initial pulse interval, a pulse time series is generated using a non-uniform interval increasing rule, where the interval of the nth pulse is the initial pulse interval multiplied by (1+αn) β , α and β are preset increasing coefficients;

[0070] The pressure fluctuation amplitude is used as the initial pulse intensity, and the intensity of subsequent pulses decays exponentially while maintaining the minimum intensity threshold;

[0071] Each time point in the pulse time series is combined with the corresponding pulse intensity to form the micro-voltage difference modulation control parameter.

[0072] According to one aspect of the present application, the exponential decay of the intensity of subsequent pulses includes:

[0073] The pressure fluctuation amplitude was taken as the intensity of the first pulse;

[0074] For the nth pulse, its intensity is equal to the sum of the pressure fluctuation amplitude multiplied by the exponential decay term and the pressure fluctuation amplitude multiplied by the holding coefficient;

[0075] The exponential decay term decreases exponentially with the pulse number, and the holding coefficient ensures that the pulse intensity is not lower than a preset proportion of the initial intensity.

[0076] According to one aspect of the present application, calculating the initial pulse interval includes:

[0077] Add 1 to the absolute value of the concentration gradient data to obtain the first normalized gradient value;

[0078] Dividing the preset reference interval value by the first normalized gradient value to obtain a dynamic adjustment factor;

[0079] The dynamic adjustment factor is compared with the minimum interval threshold, and the larger value is taken as the initial pulse interval.

[0080] Specifically, a non-equally spaced pulse sequence generation algorithm is developed to replace the traditional equal-interval method: the initial pulse interval Interval_initial = max(1ms, 50ms / (1 + |dC / dt|)) is calculated based on the concentration gradient feature Gradient_Feature, and the increasing rule Interval(n) = Interval_initial × (1 + 0.2n) is adopted. 0.8 Generate a pulse time sequence, where n is the pulse number and 0.8 is the empirical optimization index. The sequence termination condition is: cumulative time ≥ 0.8 × T_required. Sequence generation terminates when the cumulative time reaches 0.8 times the predicted replacement time T_required, and the pulse interval sequence Pulse_Sequence is output. Construct an adaptive intensity decay strategy: the initial pulse intensity is set to the optimal pressure fluctuation amplitude P_amplitude. Subsequent pulse intensities decay according to the rule Intensity(n) = P_amplitude × exp(-0.15n) + 0.3 × P_amplitude, where 0.15 is the decay rate and 0.3 is the minimum intensity maintenance factor. The time series and intensity data are combined to generate the micro-pressure differential modulation parameters Control_Params and pulse control commands Pulse_Commands, which the system uses to perform pressure modulation operations.

[0081] According to one aspect of the present application, before generating the pulse time sequence, the method further includes predicting the replacement completion time based on the pressure fluctuation amplitude to determine the generation length of the pulse time sequence:

[0082] The pressure differential acceleration factor is calculated based on the ratio of the pressure fluctuation amplitude to the reference pressure differential;

[0083] Multiply the basic attenuation coefficient by the pressure difference acceleration factor to obtain the enhanced attenuation coefficient;

[0084] Based on the exponential decay model and the enhanced decay coefficient, the predicted displacement time required to reduce the residual concentration to a preset proportion of the initial concentration is calculated;

[0085] The predicted replacement time is multiplied by a preset coefficient as the generation length threshold of the pulse time series.

[0086] Specifically, in view of the limitation that the traditional fixed time series method cannot adapt to different replacement scenarios, a residual concentration prediction model based on diffusion dynamics is established: the exponential decay basic model C_residual(t) = C_initial × exp(-λt) is adopted, where C_initial is the initial concentration and λ is the decay coefficient; the pressure difference acceleration factor Acceleration_Factor = 1 +γ× P_amplitude / 2.0% is calculated in combination with the optimal pressure fluctuation amplitude P_amplitude; the enhanced attenuation coefficient λ_enhanced = λ_base ×Acceleration_Factor, where γ = 1.5 is the acceleration coefficient and λ_base = 0.1s -1 Based on the basic attenuation coefficient, the predicted replacement time T_required is estimated by T_required = -ln(Target_ratio) / λ_enhanced, where Target_ratio = 0.05 means reducing the residual concentration to below 5%.

[0087] According to one aspect of the present application, the preset range of the pressure fluctuation amplitude is 0.5% to 2.0%;

[0088] The first weight coefficient is 0.5 to 0.7, and the second weight coefficient is 0.3 to 0.5;

[0089] The first threshold is 0.03 to 0.08, and the second threshold is 0.12 to 0.20;

[0090] The compensation factor is 0.2 to 0.4;

[0091] The parameter range is determined based on the balance between sensor protection requirements and replacement efficiency.

[0092] In summary, this application discloses a trace oxygen sampling control method, including calculating the complex impedance modulus, phase angle, and time derivative based on real and imaginary impedance data, and obtaining the double-layer perturbation index by weighted fusion of the impedance change rate and phase stability factor. Based on the perturbation index and oxygen concentration gradient, the pressure fluctuation amplitude is adaptively determined to protect sensor stability and improve replacement efficiency. A non-uniformly spaced pulse sequence is employed, with the initial interval dynamically calculated based on the concentration gradient, and the pulse intensity decays exponentially. By establishing a residual concentration prediction model based on diffusion dynamics, an enhanced attenuation coefficient is calculated based on the pressure fluctuation amplitude to accurately predict the replacement completion time. This method overcomes the technical difficulty of traditional fixed-parameter control methods in balancing sensor protection and rapid response, achieving intelligent and precise control of the trace oxygen sampling process.

[0093] In a specific embodiment of the present application, the oxygen content monitoring of protective gas in the steel metallurgical process is used as an application scenario. In the metal heat treatment process protected by high-purity nitrogen, the oxygen content needs to be controlled below 0.5ppm. The system uses an electrochemical oxygen sensor for online monitoring. The specific process of the trace oxygen sampling control method is as follows:

[0094] Step 1: Sensor data preprocessing and state recognition.

[0095] 1.1. When switching from standard gas (oxygen content 10 ppm) to high-purity nitrogen (target oxygen content <0.5 ppm), the system collects sensor data every 100 milliseconds. Two seconds after the switch, the acquired raw data is as follows: oxygen concentration voltage signal V_O2 = 0.825 V (acquisition timestamp T1 = 2.000s), real impedance Z_real = 1250 Ω (acquisition timestamp T2 = 2.002s), and imaginary impedance Z_imag = -320 Ω (acquisition timestamp T3 = 2.002s). Due to the 2ms delay in the impedance data, timestamp alignment is performed to align all data to T = 2.000s, and the synchronized sensor data set {V_O2(2.0) = 0.825 V, Z_real(2.0) = 1250 Ω, Z_imag(2.0) = -320 Ω}_sync is output.

[0096] 1.2. An adaptive Kalman filter is applied to the oxygen concentration signal. The current signal rate of change is 0.015 V / s. The noise covariance during the filter automatic adjustment process is Q = 0.001 × (1 + 10 × 0.015) = 0.00115. After filtering, V_O2_filtered = 0.824 V. A five-point median filter is applied to the impedance data. The processing sequence is [1248, 1249, 1250, 1251, 1252]Ω, and the output Z_real_clean is 1250Ω. The processing sequence is [-318, -319, -320, -321, -322]Ω, and the output Z_imag_clean is -320Ω.

[0097] 1.3. Use the pre-stored sensor calibration coefficient matrix Cal_Matrix = [2.3, 12.5, 0.08] and perform three-point calibration interpolation conversion: C_O2 = 2.3 + 12.5×0.824 + 0.08×0.824 2 = 2.3 + 10.3 + 0.054 = 7.95ppm, obtaining the standardized oxygen concentration signal C_O2_std = 7.95ppm, while transmitting the purification impedance data {Z_real_clean = 1250Ω, Z_imag_clean = -320Ω}.

[0098] Step 2: Real-time monitoring of double layer stability.

[0099] 2.1、Calculate the complex impedance modulus based on the purified impedance data: |Z| = sqrt(Z_real 2 + Z_imag 2 )=sqrt(1250 2 + (-320) 2 ) = sqrt(1664900) = 1290.3Ω. Calculate the phase angle: φ = arctan(Z_imag / Z_real) = arctan(-320 / 1250) = -0.2506 radians. Using numerical differentiation, using the phase angle φ_prev = -0.2485 radians at the previous moment (t = 1.9s), calculate the time derivative of the phase angle: dφ / dt = (φ_current -φ_prev) / Δt = (-0.2506 - (-0.2485)) / 0.1 = -0.021 radians / second. Construct the impedance feature vector Z_feature = [1290.3Ω, -0.2506 rad, -0.021 rad / s]. φ_current is the phase angle at the current moment.

[0100] 2.2. Calculate the baseline impedance modulus within the sliding window (the past 10 sampling points): |Z|_baseline = 1285Ω. The impedance change rate is: ΔZ_rate = (|Z|_current - |Z|_baseline) / |Z|_baseline = (1290.3 - 1285) / 1285 = 0.00412. Introduce a phase stability factor, set the phase change threshold φ_threshold = 0.05 rad / s, and the normalized phase change rate is: |dφ / dt| / φ_threshold = 0.021 / 0.05 = 0.42, resulting in a phase stability factor, Phase_stability = 1 - 0.42 = 0.58. The comprehensive disturbance index is calculated by weighted fusion (the first weight coefficient is optimized to 0.6 and the second weight coefficient is 0.4 through experiments): Disturbance_Index = α× |ΔZ_rate| + β× (1 - Phase_stability) = 0.6×|0.00412| + 0.4×(1 - 0.58) = 0.00247 + 0.168 = 0.171.

[0101] 2.3. A three-level threshold determination is performed on the double layer disturbance index of 0.171, with the first threshold set at 0.05 and the second threshold set at 0.15. Since 0.171 > 0.15, a significant disturbance is determined, and the sensor stability status code Sensor_Status = 3 is output, corresponding to the status correction factor Status_Factor = 0.4.

[0102] Step 3: Calculation of dynamic micro-pressure difference modulation parameters.

[0103] 3.1. Perform gradient analysis on the standardized oxygen concentration signal. The first derivative is: dC / dt = (7.95 - 7.79) / 0.1 = 1.6ppm / s (where 7.79ppm is the concentration at t = 1.9s). The second derivative is: d 2 C / dt 2 = (1.6 - 1.45) / 0.1 = 1.5ppm / s 2 (Where 1.45ppm / s is the first derivative at t = 1.9s). Construct the concentration gradient feature Gradient_Feature = [1.6ppm / s, 1.5ppm / s 2 , 1.5ppm / s 2 ].

[0104] 3.2. Set the base pressure differential, P_base, to 0.8%. Based on Sensor_Status = 3 (significant disturbance), determine the status correction factor, Status_Factor, to 0.4. Calculate the gradient response factor: C_threshold = 2.0 ppm / s, the normalized gradient value is 1.6 / 2.0 = 0.8, and the gradient response factor, Gradient_Factor = min(2.0, 1.0 + 0.5×|1.6| / 2.0) = min(2.0, 1.4) = 1.4. Introduce a disturbance compensation factor: Set the compensation coefficient to 0.3, and the disturbance compensation factor, Disturbance_Factor, is 1.0 - 0.3×0.171 = 0.949. If the base pressure differential P_base is 0.8%, the final pressure differential amplitude P_amplitude = P_base × Status_Factor × Gradient_Factor × Disturbance_Factor = 0.8% × 0.4 × 1.4 × 0.949 = 0.425%. Since it is lower than the minimum value of 0.5%, the output optimal pressure fluctuation amplitude P_amplitude = 0.5%.

[0105] Step 4: Adaptive pulse timing optimization.

[0106] 4.1. Using the exponential decay base model, C_residual(t) = C_initial × exp(-λt), where C_initial = 7.95 ppm, calculate the enhanced attenuation coefficient: λ_enhanced = λ_base × (1 + 1.5 × P_amplitude / 2.0%) = 0.1 × (1 + 1.5 × 0.5% / 2.0%) = 0.1 × 1.375 = 0.1375 / s, with the base attenuation coefficient λ_base = 0.1 / s. Predict the time required to reduce the residual concentration to below 5%: T_required = -ln(0.05) / 0.1375 = 2.996 / 0.1375 = 21.8 seconds.

[0107] 4.2. Calculate the initial pulse interval based on the concentration gradient characteristics: The absolute value of the oxygen concentration gradient, |dC / dt|, is 1.6 ppm / s, and the normalized gradient value, 1 + |dC / dt|, is 1.6 + 1 = 2.6. The baseline interval is set to 50 ms, and the dynamic adjustment factor is 50 / 2.6 = 19.2 ms. Compared to the minimum interval threshold of 1 ms, the larger value is taken. Therefore, Interval_initial = max(1 ms, 50 ms / (1 + |1.6|)) = max(1 ms, 19.2 ms) = 19.2 ms. Generate the pulse time series using an increasing rule, setting α = 0.2 and β = 0.8. The interval of the nth pulse is: Interval(n) = 19.2 × (1 + 0.2n) 0.8 The specific sequence is: First pulse: t = 0ms, Interval(1) = 19.2ms; Second pulse: t = 19.2ms, Interval(2) = 19.2×(1 + 0.2×2) 0.8 = 22.3ms; the third pulse: t = 41.5ms, Interval(3) = 19.2×(1 + 0.2×3) 0.8 = 25.2ms. When the accumulated time reaches 0.8×21.8 = 17.4 seconds, it terminates and outputs the pulse interval sequence Pulse_Sequence.

[0108] 4.3. The initial pulse intensity is set to P_amplitude = 0.5%. Subsequent pulse intensity calculations are: Intensity(1) = 0.5%; Intensity(2) = 0.5%×exp(-0.15×1) + 0.3×0.5% = 0.431%; Intensity(3) = 0.5%×exp(-0.15×2) + 0.3×0.5% = 0.370%. The time series and intensity data are combined to generate the micro-voltage differential modulation parameters Control_Params = {(0ms, 0.5%), (19.2ms, 0.431%), (41.5ms, 0.370%)...} and the pulse control instructions Pulse_Commands.

[0109] To further optimize the control effect, calculate the second-order derivative of oxygen concentration: d 2 C / dt 2 = (1.6 - 1.45) / 0.1 = 1.5ppm / s 2 . Set the acceleration threshold to 2.0ppm / s 2, the normalized acceleration value is 1.5 / 2.0 = 0.75. The variable acceleration factor is 1 + 0.3×0.75 = 1.225. The enhanced gradient response factor is 1.4×1.225 = 1.715.

[0110] This embodiment shortens system response time and reduces sensor baseline drift. Especially under significant disturbances, it effectively protects the sensor by reducing pressure fluctuations, avoiding permanent damage caused by traditional high-intensity pulses. Actual operational data shows improved sensor sensitivity retention after 1,000 hours of continuous operation, fully demonstrating the superiority of this embodiment in balancing fast response and sensor protection.

[0111] The preferred embodiments of the present invention are described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the scope of protection of the present invention.

Claims

1. A trace oxygen sampling control method, characterized in that: include: Obtain the real impedance data and the imaginary impedance data of the sensor; and calculate the complex impedance modulus and the phase angle time derivative based on the data; Calculating the rate of change of the complex impedance modulus relative to a reference value, and calculating the phase stability factor based on the time derivative of the phase angle; The double layer perturbation index is calculated by weighted fusion of the rate of change of the complex impedance modulus relative to the reference value and the phase stability factor; Dynamically determine the pressure fluctuation amplitude based on the double layer perturbation index; Generate micro-pressure difference modulation control parameters according to the pressure fluctuation amplitude; Calculating the phase stability factor involves: The absolute value of the time derivative of the phase angle is divided by a preset phase change threshold to obtain a normalized phase change rate; Calculate the difference between 1 and the normalized phase change rate to obtain the phase stability factor; Among them, when the absolute value of the time derivative of the phase angle is larger, the phase stability factor is smaller, indicating that the double layer stability is worse; Dynamic determination of pressure fluctuation amplitude includes: Obtaining the time derivative of the oxygen concentration signal as concentration gradient data; Determine the sensor stability state according to the double layer disturbance index and set the corresponding state correction factor; Calculate gradient response factors based on concentration gradient data; The disturbance compensation factor is calculated based on the double layer disturbance index. The larger the double layer disturbance index, the smaller the disturbance compensation factor. Multiply the preset basic pressure difference value with the state correction factor, gradient response factor and disturbance compensation factor to obtain the pressure fluctuation amplitude; The control parameters for generating micro-pressure differential modulation according to the pressure fluctuation amplitude include: The initial pulse interval is calculated based on the absolute value of the concentration gradient data. The larger the concentration gradient data, the smaller the initial pulse interval. Based on this, the pulse time series is generated using the non-uniform interval increasing rule, where the interval of the nth pulse is the initial pulse interval multiplied by (1+αn) β , α and β are preset increasing coefficients; The pressure fluctuation amplitude is used as the initial pulse intensity, and the intensity of subsequent pulses decays exponentially while maintaining the minimum intensity threshold; Each time point in the pulse time series is combined with the corresponding pulse intensity to form the micro-voltage difference modulation control parameter.

2. The method according to claim 1, characterized in that Calculation of the double layer perturbation index includes: Multiplying the absolute value of the rate of change of the complex impedance modulus relative to the reference value by a first weight coefficient to obtain a first weighted result; Multiplying the difference between 1 and the phase stability factor by the second weight coefficient to obtain a second weighted result; Adding the first and second weighted results to obtain the double layer perturbation index; The first weight coefficient is greater than the second weight coefficient.

3. The method according to claim 1, characterized in that Determining the sensor stability state based on the double layer perturbation index includes: When the double layer perturbation index is less than the first threshold, it is determined to be in a stable state; When the double layer disturbance index is greater than or equal to the first threshold and less than the second threshold, it is determined to be a slight disturbance state; When the double layer disturbance index is greater than or equal to the second threshold, it is determined to be a significant disturbance state; Among them, the second threshold is 2 to 4 times the first threshold, and the state correction factors corresponding to different states decrease in sequence.

4. The method according to claim 1, wherein Calculating the gradient response factor involves: The absolute value of the concentration gradient data is divided by the preset concentration change threshold to obtain a normalized gradient value; The normalized gradient value is multiplied by the gradient response coefficient and then added to 1 to obtain the initial response factor; An upper limit value is set for the initial response factor to obtain a gradient response factor; Among them, the faster the oxygen concentration changes, the larger the gradient response factor, which increases the pressure fluctuation amplitude accordingly.

5. The method according to claim 1, wherein Calculating the disturbance compensation factor includes: Multiply the double layer perturbation index by the compensation coefficient to obtain the compensation amount; Calculate the difference between 1 and the compensation amount to obtain the disturbance compensation factor; The compensation coefficient is less than 1, so that the larger the double layer perturbation index is, the smaller the perturbation compensation factor is.

6. The method according to claim 1, characterized in that The subsequent pulse intensities decay exponentially, including: The pressure fluctuation amplitude was taken as the intensity of the first pulse; For the nth pulse, its intensity is equal to the sum of the pressure fluctuation amplitude multiplied by the exponential decay term and the pressure fluctuation amplitude multiplied by the retention factor; the formula is: Intensity(n) = P_amplitude × exp(-0.15n) + 0.3 × P_amplitude; where 0.15 is the decay rate, 0.3 is the minimum intensity retention factor, P_amplitude is the pressure fluctuation amplitude, n is the pulse number, and Intensity(n) is the pulse intensity. The exponential decay term decreases exponentially with the pulse number, and the holding coefficient ensures that the pulse intensity is not lower than a preset proportion of the initial intensity.

7. The method according to claim 1, characterized in that Calculating the initial pulse interval includes: Add 1 to the absolute value of the concentration gradient data to obtain the first normalized gradient value; Dividing the preset reference interval value by the first normalized gradient value to obtain a dynamic adjustment factor; The dynamic adjustment factor is compared with the minimum interval threshold, and the larger value is taken as the initial pulse interval.

Images