Impurity detection control method and system for high-purity oxygen

Abstract

The invention relates to the technical field of impurity gas detection control, and discloses an impurity detection control method and system for high-purity oxygen, and the method comprises the steps: arranging a plurality of groups of TDLAS detection probes and temperature vibration monitoring units, and obtaining a spectrum harmonic signal and temperature and vibration data; the concentration is inverted through wavelength modulation spectrum demodulation and temperature and pressure disturbance compensation; pCA fusion denoising processing is executed through the edge calculation unit, and correction concentration is output; and abnormity is judged through a multi-scale dynamic threshold mechanism and linkage control is carried out. Compared with a common single-point probe or traditional low-pass filtering in the prior art, the technical problem that interference between impurity concentration change and environmental disturbance cannot be effectively handled especially in a high-vibration and high-temperature-difference environment is solved. According to the invention, through multi-channel cooperative weighting, PCA fusion compensation and a dynamic threshold determination mechanism, accurate detection control of the high-purity oxygen conveying pipeline in an industrial environment is realized, and the automation level of pipeline impurity gas detection control is improved.

Description

Technical Field

[0001] This invention relates to the field of impurity gas detection and control technology, and in particular to a method and system for detecting and controlling impurities in high-purity oxygen. Background Technology

[0002] Currently, high-purity oxygen is an indispensable key gas in high-end industrial fields such as metallurgy, chemical industry, and electronics manufacturing. Controlling its impurity concentration directly affects the quality of end products and process safety. Taking the steel smelting process as an example, the oxygen supply to the furnace typically requires high purity, while the content of trace impurities (such as N2, CO2, Ar, and H2O) needs strict control. Therefore, how to achieve real-time, accurate, and stable online detection of impurity content in oxygen pipelines under complex operating conditions has become a key issue in the field of high-purity gas transportation.

[0003] In existing technologies, tunable diode laser absorption spectroscopy (TDLAS) is often used for online monitoring. Although it theoretically possesses high sensitivity and selectivity, it still has many shortcomings in practical industrial applications. For example, most current methods rely on single-point probe deployment, which cannot cover multiple areas of long-distance pipeline networks; data processing mostly uses static thresholding or low-pass filtering algorithms, making it difficult to distinguish between normal fluctuations and actual impurity increases; and in the face of factors commonly found in steel plants such as high temperature differences, strong mechanical vibrations, and airflow disturbances, TDLAS probe signals are prone to baseline drift and amplitude distortion, leading to frequent false alarms and missed alarms.

[0004] In addition, existing solutions generally lack the ability to fuse and denoise multi-channel detection results and identify anomalies, and cannot build an intelligent control closed loop with predictive capabilities and adaptive response mechanisms. Especially under conditions such as changes in gas flow rate, severe operating disturbances, and deterioration of signal quality, traditional algorithms struggle to maintain detection stability and judgment reliability.

[0005] Therefore, there is an urgent need for a new detection method that can achieve high sensitivity, dynamic discrimination, and self-closed-loop control of trace impurities in high-purity oxygen under complex working conditions and multi-disturbance environments, so as to improve the accuracy, real-time performance, and industrial intelligence level of impurity detection and meet the actual needs of key industries for continuous, stable, and controllable operation of the oxygen supply process. Summary of the Invention

[0006] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a method for detecting and controlling impurities in high-purity oxygen. This method aims to solve the technical problem that common single-point probes or traditional low-pass filters in the prior art, especially in environments with high vibration and temperature differences, cannot effectively cope with the interference between changes in impurity concentration and environmental disturbances.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a method for detecting and controlling impurities in high-purity oxygen.

[0008] The method for detecting and controlling impurities in the high-purity oxygen includes:

[0009] Step S10: Deploy multiple sets of TDLAS detection probes and corresponding temperature and vibration monitoring units along the target high-purity oxygen delivery pipeline; wherein, the TDLAS detection probes are used to measure the intensity of characteristic absorption lines of impurities in the high-purity oxygen delivery pipeline to obtain spectral harmonic signals; the temperature and vibration monitoring units are used to acquire real-time temperature data. and vibration acceleration data ;

[0010] Step S20: The spectral harmonic signal acquired by the TDLAS detection probe is subjected to harmonic normalization processing using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. ;

[0011] Step S30: The relative concentration Real-time temperature data and vibration acceleration data The data is input to a preset edge computing unit, where impurity concentration is denoised and corrected based on a multi-channel collaborative weighted PCA fusion compensation mechanism, and the corrected concentration data set is output.

[0012] Step S40: Obtain the historical concentration data of the kth TDLAS detection probe, fuse the historical concentration data of the output corrected concentration data set, and determine whether the impurity concentration is abnormal based on the multi-scale dynamic threshold judgment mechanism and record the abnormal event.

[0013] Step S50: The abnormal event recorded by the needle will automatically trigger the preset closed-loop linkage control logic.

[0014] Preferably, in step S10, the characteristic absorption line intensities of the impurities include the symmetric stretching vibration absorption line intensity of CO2 at a wavelength of 2.004 μm, the asymmetric stretching vibration absorption line intensity of H2O at a wavelength of 1.395 μm, the indirect line intensity of the O2 absorption linewidth change caused by Ar or N2 dilution within a 760 nm bandwidth, the doublet structure absorption line intensity of O2 at 760.6 nm, and the absorption line intensity of CH4 at 1.651 μm.

[0015] Preferably, in step S20, harmonic normalization is performed on the spectral harmonic signal acquired by the TDLAS detection probe using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. The steps specifically include:

[0016] Step S201: Apply a high-frequency sinusoidal modulation signal within a preset first frequency range to the TDLAS detection probe, periodically scan and acquire the transmitted light intensity signal of the gas under test, and extract the first harmonic component 1f and the second harmonic component 2f.

[0017] Step S202: Normalize the first harmonic component 1f and the second harmonic component 2f using the second harmonic ratio principle with phase self-optimization to obtain the absorption intensity ratio signal R;

[0018] Step S203: Collect real-time temperature data corresponding to the k-th TDLAS detection probe. With vibration acceleration data and utilize real-time temperature data With vibration acceleration data An adaptive Kalman filter coupled temperature and vibration disturbance compensation algorithm is used to perform temperature and pressure disturbance compensation, and the temperature correction spectral line intensity and vibration influence correction factor are obtained.

[0019] Step S204: Perform BeerLambert dynamic inversion calculation with multivariate temperature and pressure perturbation compensation based on temperature-corrected spectral line intensity and vibration influence correction factor, and output the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. .

[0020] Preferably, in step S20, a Beer-Lambert dynamic inversion calculation with multivariable temperature and pressure perturbation compensation is performed based on the temperature-corrected spectral line intensity and vibration influence correction factor, outputting the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. The steps specifically include performing the following calculations:

[0021]

[0022] in, For real-time temperature data The intensity of the temperature-corrected spectral line of the i-th gas; For vibration acceleration data The correction factor for the vibrational influence of the i-th impurity spectral line; For reference light intensity; This represents the actual measured intensity of transmitted light. The vibration response factor related to impurity i is used to account for the influence of vibration on the absorbed signal and to eliminate vibration acceleration data. The influence of dimensions; It is the relative length of the optical path; Let represent the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. These are dimensionless data, and their values ​​reflect the concentration level.

[0023] Preferably, in step S30, the relative concentration is... Real-time temperature data and vibration acceleration data The process of inputting data into a preset edge computing unit, performing impurity concentration denoising correction based on a multi-channel collaborative weighted PCA fusion compensation mechanism in the edge computing unit, and outputting a corrected concentration data set specifically includes:

[0024] Step S301: Window Construction and Spatiotemporal Alignment: Based on Relative Concentration Concentration sequence C(t) associated with time t is constructed, real-time gas flow rate and TDLAS probe spacing are obtained, and time translation and interpolation processing are performed on concentration sequences C(t) from different locations to form a spatiotemporal alignment window W covering the same gas mass;

[0025] Step S302: Channel Health Assessment and Co-weighting: Calculate the health weight for each TDLAS detection probe within the spatiotemporal alignment window W. The health weight is determined by a combination of the following dimensional indicators: mean square of spectral line fitting residuals. Temperature drift sensitivity coefficient and vibration sensitivity coefficient And based on health score weights Forming a weighted observation matrix ;

[0026] Step S303: PCA fusion and perturbation separation: This involves adjusting the weighted observation matrix... Perform principal component analysis to obtain the principal component loading matrix. Combined with the scoring matrix, real-time temperature data and vibration acceleration data Constructing the exogenous feature matrix A principal component working condition regression sub-model is established; where the principal components of the principal component working condition regression sub-model are working condition related components, the reconstructed value of the principal component working condition regression sub-model is the systematic shift caused by temperature and vibration, and the residual R of the principal component working condition regression sub-model is regarded as a mixture of the actual impurity change and the measurement noise after debiasing.

[0027] Step S304: Based on the residual R of the principal component working condition regression sub-model, the Tukey function is used for correction. The corrected residual is added to the reconstructed value of the principal component working condition regression sub-model to obtain the debiased concentration correction result. Finally, the corrected concentration data set is output.

[0028] Preferably, step S40, which involves acquiring historical concentration data from the k-th TDLAS detection probe, fusing the historical concentration data from the output corrected concentration data set, and determining whether the impurity concentration is abnormal and recording abnormal events based on a multi-scale dynamic threshold determination mechanism, specifically includes:

[0029] Step S401: Obtain the historical concentration data of the k-th TDLAS detection probe, and use Variational Mode Decomposition (VMD) to decompose the historical concentration data of the k-th TDLAS detection probe into low-frequency trend components. With high-frequency fluctuation components Among them, the low-frequency trend component is used to reflect the long-term trend of impurity concentration, and the high-frequency fluctuation component is used to reflect short-term fluctuation characteristics.

[0030] Step S402: For low-frequency trend components Calculate the mean and standard deviation of the sliding window to form a trend feature vector; for high-frequency fluctuation components... Calculate the energy index and kurtosis index to form a fluctuation characteristic vector;

[0031] Step S403: Introduce trend fold factor and fluctuation fold factor; based on the trend feature vector and fluctuation feature vector combined with the trend fold factor and fluctuation fold factor, generate a multi-scale dynamic threshold set for each group of TDLAS detection probes; wherein, the multi-scale dynamic threshold set includes low-frequency trend thresholds. and high-frequency fluctuation threshold ;

[0032] Step S404: When the relative concentration at the k-th TDLAS detection probe position... Low-frequency trend components that simultaneously meet the following three conditions Greater than or equal to the low-frequency trend threshold And high-frequency fluctuation components Greater than or equal to the high-frequency fluctuation threshold When this occurs, it is determined to be a valid abnormal event.

[0033] Preferably, in step S50, the step of automatically triggering a preset closed-loop linkage control logic based on the abnormal event recorded by the needle specifically includes:

[0034] When an abnormal event is detected as a single point or local anomaly, the opening of the vent valve of the corresponding pipeline section is automatically adjusted or the backup oxygen supply channel is switched.

[0035] When an abnormal event is detected as an abnormal global concentration trend, a flow reduction or shutdown command is sent to the upstream oxygen supply source, and the whole network purging mode is initiated.

[0036] All control execution commands and sensor status data are uploaded to the cloud monitoring platform in real time. The cloud monitoring platform performs pattern analysis and threshold optimization, and sends updated dynamic threshold parameter sets and control strategies to the edge computing unit to achieve self-learning and adaptive parameter updates.

[0037] This invention also provides a high-purity oxygen impurity detection and control system, comprising:

[0038] The probe deployment and environmental monitoring module is used to deploy multiple sets of TDLAS detection probes and corresponding temperature and vibration monitoring units along the target high-purity oxygen delivery pipeline. The TDLAS detection probes are used to measure the intensity of characteristic absorption lines of impurities in the high-purity oxygen delivery pipeline to obtain spectral harmonic signals; the temperature and vibration monitoring units are used to acquire real-time temperature data. and vibration acceleration data ;

[0039] The spectral demodulation and dynamic inversion module is used to perform harmonic normalization processing on the spectral harmonic signal acquired by the TDLAS detection probe using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. ;

[0040] The concentration denoising and fusion compensation module is used to reduce relative concentration. Real-time temperature data and vibration acceleration data The data is input to a preset edge computing unit, where impurity concentration is denoised and corrected based on a multi-channel collaborative weighted PCA fusion compensation mechanism, and the corrected concentration data set is output.

[0041] The historical data fusion and anomaly detection module is used to acquire the historical concentration data of the k-th TDLAS detection probe, fuse the historical concentration data of the output corrected concentration data set, and determine whether the impurity concentration is abnormal and record the abnormal event based on the multi-scale dynamic threshold judgment mechanism.

[0042] The abnormal response and closed-loop control module is used to automatically trigger preset closed-loop linkage control logic for recorded abnormal events.

[0043] The present invention also provides a high-purity oxygen impurity detection and control device, comprising: a memory, a processor, and a high-purity oxygen impurity detection and control program stored in the memory and executable on the processor. When the high-purity oxygen impurity detection and control program is executed by the processor, a high-purity oxygen impurity detection and control method is implemented.

[0044] The present invention also provides a computer program product, including a high-purity oxygen impurity detection and control program, which, when executed by a processor, implements the high-purity oxygen impurity detection and control method.

[0045] The beneficial effects of this invention are as follows: By introducing a dynamic inversion method with multivariable temperature and pressure disturbance compensation and a PCA denoising mechanism, this invention achieves high-precision, low-error real-time detection of trace impurities in high-purity oxygen under high temperature difference and high vibration environment.

[0046] This invention employs a multi-scale dynamic threshold determination and closed-loop linkage control strategy, which can identify abnormal concentration trends in advance and automatically trigger control responses. This avoids false alarms and missed alarms caused by static threshold lag in traditional methods, and significantly enhances the safety and intelligence level of the oxygen delivery system. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating the first embodiment of a method for detecting and controlling impurities in high-purity oxygen according to the present invention.

[0049] Figure 2 This is a schematic diagram of the equipment for the impurity detection and control method of high-purity oxygen according to the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the high-purity oxygen impurity detection and control method of the present invention, which presents the first embodiment of the high-purity oxygen impurity detection and control method of the present invention.

[0052] In the first embodiment, the impurity detection and control method for high-purity oxygen includes:

[0053] Step S10: Deploy multiple sets of TDLAS detection probes and corresponding temperature and vibration monitoring units along the target high-purity oxygen delivery pipeline; wherein, the TDLAS detection probes are used to measure the intensity of characteristic absorption lines of impurities in the high-purity oxygen delivery pipeline to obtain spectral harmonic signals; the temperature and vibration monitoring units are used to acquire real-time temperature data. and vibration acceleration data ;

[0054] It should be noted that the TDLAS detection probe is a transmission spectral detection component using a tunable semiconductor laser, preferably deployed in typical thermal disturbance sections at the beginning, middle and end of the pipeline. Each probe group corresponds to at least one temperature and vibration monitoring unit, which includes sensors such as thermocouples and MEMS accelerometers to sense local environmental fluctuations in the probe area.

[0055] It is understandable that, since the impurity concentration of high-purity oxygen is extremely low (usually below 5 ppm) and there are temperature gradients and mechanical interferences at different locations, this invention can achieve spatial distribution sensing and redundant signal correction by deploying probes and environmental sensors in multiple key areas, effectively enhancing the spatial representativeness of the measurement.

[0056] It should be understood that, compared with the existing technology that only detects at a single fixed endpoint, this deployment scheme not only covers the possible disturbance area of ​​the entire pipeline, but also provides a foundation for subsequent spatiotemporal alignment, channel weighting and denoising modeling, and constitutes the data input prerequisite for subsequent edge computing and fusion analysis.

[0057] For example, in the oxygen supply system of a converter in a steel company, the temperature difference range of the 200-meter-long main oxygen pipeline is as high as 45°C. Pressure fluctuations and equipment vibrations can cause the laser signal intensity to drift. This invention installs a set of TDLAS probes and triaxial vibration and temperature monitoring units in the initial section, compression section and branch section of the pipeline. In subsequent steps, through multi-point data synchronization, the dynamic change trend of CO2 and H2O is effectively identified, and a 5-minute early warning is achieved for concentration increases of 0.5ppm.

[0058] Step S20: The spectral harmonic signal acquired by the TDLAS detection probe is subjected to harmonic normalization processing using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. ;

[0059] It should be noted that the wavelength modulation spectral demodulation mechanism adopts second harmonic demodulation (2f / 1f) to improve the resolution of weak absorption signals, and constructs the absorption intensity ratio signal R by normalizing the baseline offset to weaken the influence of light source fluctuations and path disturbances. At the same time, in order to adapt to the complex non-stationary thermal vibration coupling disturbances in the pipeline environment, this step inputs real-time temperature and vibration acceleration as independent variables into the multivariate compensation model to dynamically correct the optical path intensity and absorption coefficient offset terms in the Beer–Lambert model.

[0060] Understandably, traditional TDLAS demodulation typically achieves high accuracy in environments with constant temperature and stable platforms. However, in industrial settings, high-frequency vibrations and pipe wall temperature drift significantly affect harmonic amplitudes and baseline stability, leading to systematic deviations in the derived impurity concentrations. Therefore, introducing a multivariate perturbation compensation mechanism not only improves the environmental adaptability of the inversion calculation but also lays the foundation for subsequent concentration correction and anomaly identification.

[0061] It should be understood that the "multivariable temperature and pressure disturbance compensation type Beer-Lambert dynamic inversion method" constructed in this step not only introduces a nonlinear disturbance correction term in the formula structure, but also adopts a dynamically estimated absorption coefficient matrix and an unsteady-state fitting residual feedback mechanism in the modeling strategy. It has high responsiveness, high robustness and cross-channel adaptability, and is significantly better than the traditional static linear inversion model.

[0062] Step S30: The relative concentration Real-time temperature data and vibration acceleration data The data is input to a preset edge computing unit, where impurity concentration is denoised and corrected based on a multi-channel collaborative weighted PCA fusion compensation mechanism, and the corrected concentration data set is output.

[0063] It should be noted that the multi-channel collaborative weighted PCA fusion compensation mechanism refers to the following: For concentration signals from multiple TDLAS detection points, each channel carries the temperature and vibration interference characteristics at the corresponding time. First, a joint feature matrix of "signal-interference" is constructed, and a weight model based on residual fitting goodness, temperature and vibration sensitivity, and historical stability is introduced to dynamically weight the contribution of each channel. Then, PCA is used to perform dimensionality reduction and purification on the joint feature matrix to extract the main components and filter out environmentally induced non-common interference. Finally, a low-noise concentration signal set close to the real working conditions is recovered through linear reconstruction.

[0064] Understandably, this step significantly improves the stability and accuracy of concentration signals in high-purity oxygen transportation scenarios, especially addressing issues such as signal baseline drift and inconsistencies in multi-point data caused by temperature gradient fluctuations and pipeline vibrations. Through a PCA compensation model deployed at the edge, near real-time data cleaning and correction can be achieved, effectively alleviating the processing bottleneck of traditional centralized computing in high-frequency sampling environments.

[0065] It should be understood that, compared to traditional denoising methods that primarily rely on time-sliding window averaging or fixed filtering algorithms, this mechanism offers stronger robustness and adaptability while preserving the dynamic nature of the concentration response. Traditional algorithms often fail to distinguish between "measurement errors" and "true concentration disturbances" when faced with multi-dimensional interference sources. In contrast, this invention effectively distinguishes between systematic and incidental disturbances through principal component modeling and dynamically adjusts data reliability by combining channel weights, resulting in a final output that more closely reflects the actual trend of oxygen composition changes.

[0066] For example, in a test verification at a special steel plant, the pipeline temperature fluctuated within ±20℃, and the maximum vibration acceleration reached 0.8g. After processing with the traditional low-pass filtering algorithm, the impurity concentration signal still had high-frequency noise of ±0.2ppm. However, after processing with the PCA fusion compensation mechanism in this invention, the standard deviation significantly converged to ±0.04ppm, the false error rate was reduced by 73.5%, and the detection response to normal impurity mutation events was advanced by an average of 4.8 seconds. This fully demonstrates the high precision, fast response, and stable performance of this invention in industrial disturbance environments.

[0067] Step S40: Obtain the historical concentration data of the kth TDLAS detection probe, fuse the historical concentration data of the output corrected concentration data set, and determine whether the impurity concentration is abnormal based on the multi-scale dynamic threshold judgment mechanism and record the abnormal event.

[0068] It should be noted that the multi-scale dynamic threshold determination mechanism refers to: after aligning the corrected impurity concentration data with the corresponding historical concentration data sequence, constructing concentration statistical models for three time scales: short-term window (e.g., within 5 minutes), medium-term window (e.g., within 1 hour), and long-term window (e.g., within 24 hours). For each scale, dynamically calculate the mean fluctuation range, skewness distribution index, and adaptive tolerance bandwidth. At the same time, combined with the operating cycle and process load level of the TDLAS probe's region, the determination results of each scale are weighted and fused to output a cross-scale dynamic determination threshold set, which is used to identify abnormal concentration shifts and abrupt trends in real time.

[0069] Understandably, this mechanism effectively bridges the contradiction between "rapid mutation" and "chronic drift" in detection response, enabling early identification of subtle concentration anomalies and suppression of short-term abnormal fluctuations, thus ensuring accurate detection of critical impurity events in complex industrial environments. Furthermore, by dynamically adjusting threshold sensitivity and tolerance at various scales, it can adapt to changes in operating conditions at different oxygen delivery stages (such as start-up, steady state, and overload).

[0070] It should be understood that traditional anomaly identification methods based on fixed thresholds or single-window sliding detection suffer from problems such as "threshold rigidity" or frequent "false alarms / missed alarms," ​​making it difficult to adapt to concentration fluctuations caused by changes in pipeline temperature, vibration, and operating pressure. This invention, by integrating historical operating condition features, dynamic window scale adaptive modeling, and cross-scale comprehensive judgment, achieves a more sensitive and robust concentration anomaly identification strategy than traditional methods, effectively improving detection reliability and response accuracy.

[0071] For example, during a 7-day continuous monitoring of a medical liquid oxygen pipeline, when the ambient temperature suddenly dropped at night, causing changes in pipe wall stress and inducing concentration disturbances, the traditional 3σ fixed threshold model had a false alarm rate of 18.4%. However, when using the multi-scale dynamic threshold mechanism of this invention, through cross-scale deviation aggregation calculation, only 2 actual impurity transition events were marked, reducing the false alarm rate to 2.1%. Moreover, the average response to each sudden change was 3.2 minutes earlier, winning a critical time window for subsequent automated linkage control.

[0072] Step S50: The abnormal event recorded by the needle will automatically trigger the preset closed-loop linkage control logic.

[0073] It should be noted that the closed-loop linkage control logic refers to the following: after recording an abnormal impurity concentration event of any TDLAS detection probe, the linkage response template bound to the probe number is automatically invoked to execute a rapid response control process based on the three-layer mapping relationship of "detection point - positioning area - control unit". Specifically, this includes: initiating local gas source bypass switching, adjusting the flow output of the booster device, activating the exhaust purification subsystem, and feeding back the response results to the upper-level control platform in real time, thereby forming a closed-loop linkage control chain consisting of "abnormal detection - strategy execution - status feedback".

[0074] Understandably, by setting up the above closed-loop control mechanism, it is possible to ensure that when the impurity concentration exceeds the safety threshold, the pollution source can be quickly located and the pipeline delivery status can be actively adjusted without human intervention, preventing impurities from spreading to critical downstream gas consumption links. This is especially suitable for application scenarios with extremely high requirements for gas purity, such as electronic-grade oxygen and medical liquid oxygen.

[0075] It should be understood that, compared to the traditional response process that relies on manual inspection and handling, the closed-loop linkage control mechanism not only significantly reduces response delay (from the traditional 5-10 minutes to the second level), but also completes the self-verification of strategy execution through the feedback mechanism, realizing true "intelligent early warning + autonomous handling"; especially in remote unattended management sections, this mechanism can significantly improve the safety level of automatic operation and the timeliness of accident handling.

[0076] For example, when testing a 300-meter-long high-purity oxygen transmission line, after the 6th group of TDLAS probes recorded that the nitrogen impurity concentration rose to 7 ppm (exceeding the set 5 ppm threshold), it completed the positioning within 3 seconds and automatically triggered the bypass valve switching and buffer gas injection. There were no downstream pressure fluctuations or purity deviations during the entire process, and the control platform received complete feedback logs and fault reports in just 6 seconds. This is 10 times faster than the traditional "detection-reporting-decision-execution" process, demonstrating good closed-loop control efficiency and collaborative capabilities.

[0077] Example 2: Furthermore, the present invention provides a high-purity oxygen impurity detection and control system, employing a high-purity oxygen impurity detection and control method from the above embodiments, which can solve the technical problem of high-purity oxygen impurity detection and control. Compared with the prior art, the beneficial effects of the high-purity oxygen impurity detection and control system provided by the present invention are the same as those of the high-purity oxygen impurity detection and control method from the above embodiments, and other technical features in the high-purity oxygen impurity detection and control system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0078] Example 3: This invention provides a device for detecting and controlling impurities in high-purity oxygen. Please refer to... Figure 2A high-purity oxygen impurity detection and control device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to perform the high-purity oxygen impurity detection and control method described in Embodiment 1 above. The high-purity oxygen impurity detection and control device in this embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and vehicle terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. This high-purity oxygen impurity detection and control device is merely an example and should not limit the functionality or scope of use of the embodiments of this invention. The high-purity oxygen impurity detection and control device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. Random access memory 1004 also stores various programs and data required for the operation of a high-purity oxygen impurity detection and control device. Processing device 1001, read-only memory 1002, and random access memory 1004 are interconnected via bus 1005. I / O interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows a high-purity oxygen impurity detection and control device to communicate wirelessly or wiredly with other devices to exchange data. Although a high-purity oxygen impurity detection and control device with various systems is shown in the figures, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented or possessed alternatively.

[0079] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the high-purity oxygen impurity detection and control method described above. The computer program product provided by this invention can solve the technical problem of high-purity oxygen impurity detection and control. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as the beneficial effects of the high-purity oxygen impurity detection and control method provided in the above embodiments, and will not be repeated here.

[0080] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0081] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0082] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for detecting and controlling impurities in high-purity oxygen, characterized in that, The methods include: Step S10: Deploy multiple sets of TDLAS detection probes and corresponding temperature and vibration monitoring units along the target high-purity oxygen delivery pipeline; wherein, the TDLAS detection probes are used to measure the intensity of characteristic absorption lines of impurities in the high-purity oxygen delivery pipeline to obtain spectral harmonic signals; the temperature and vibration monitoring units are used to acquire real-time temperature data. and vibration acceleration data ; Step S20: The spectral harmonic signal acquired by the TDLAS detection probe is subjected to harmonic normalization processing using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. ; Step S30: The relative concentration Real-time temperature data and vibration acceleration data The data is input to a preset edge computing unit, where impurity concentration is denoised and corrected based on a multi-channel collaborative weighted PCA fusion compensation mechanism, and the corrected concentration data set is output. Step S40: Obtain the historical concentration data of the kth TDLAS detection probe, fuse the historical concentration data of the output corrected concentration data set, and determine whether the impurity concentration is abnormal based on the multi-scale dynamic threshold judgment mechanism and record the abnormal event. Step S50: The abnormal event recorded by the needle will automatically trigger the preset closed-loop linkage control logic.

2. The method for detecting and controlling impurities in high-purity oxygen as described in claim 1, characterized in that, In step S10, the characteristic absorption line intensities of the impurities include the symmetric stretching vibration absorption line intensity of CO2 at a wavelength of 2.004 μm, the asymmetric stretching vibration absorption line intensity of H2O at a wavelength of 1.395 μm, the indirect line intensity of the O2 absorption linewidth change caused by Ar or N2 dilution within a 760 nm bandwidth, the doublet structure absorption line intensity of O2 at 760.6 nm, and the absorption line intensity of CH4 at 1.651 μm.

3. The method for detecting and controlling impurities in high-purity oxygen as described in claim 1, characterized in that, In step S20, harmonic normalization is performed on the spectral harmonic signal acquired by the TDLAS detection probe using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. The steps specifically include: Step S201: Apply a high-frequency sinusoidal modulation signal within a preset first frequency range to the TDLAS detection probe, periodically scan and acquire the transmitted light intensity signal of the gas under test, and extract the first harmonic component 1f and the second harmonic component 2f. Step S202: Normalize the first harmonic component 1f and the second harmonic component 2f using the second harmonic ratio principle with phase self-optimization to obtain the absorption intensity ratio signal R; Step S203: Collect real-time temperature data corresponding to the k-th TDLAS detection probe. With vibration acceleration data and utilize real-time temperature data With vibration acceleration data An adaptive Kalman filter coupled temperature and vibration disturbance compensation algorithm is used to perform temperature and pressure disturbance compensation, and the temperature correction spectral line intensity and vibration influence correction factor are obtained. Step S204: Perform BeerLambert dynamic inversion calculation with multivariate temperature and pressure perturbation compensation based on temperature-corrected spectral line intensity and vibration influence correction factor, and output the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. .

4. The method for detecting and controlling impurities in high-purity oxygen as described in claim 3, characterized in that, In step S20, a BeerLambert dynamic inversion calculation with multivariable temperature and pressure perturbation compensation is performed based on the temperature-corrected spectral line intensity and vibration influence correction factor, outputting the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. The steps specifically include performing the following calculations: ； in, For real-time temperature data The intensity of the temperature-corrected spectral line of the i-th gas; For vibration acceleration data The correction factor for the vibrational influence of the i-th impurity spectral line; For reference light intensity; This represents the actual measured intensity of transmitted light. The vibration response factor related to impurity i is used to account for the influence of vibration on the absorbed signal and to eliminate vibration acceleration data. The influence of dimensions; It is the relative length of the optical path; Let represent the relative concentration of the i-th impurity at the k-th TDLAS detection probe position. These are dimensionless data, and their values ​​reflect the concentration level.

5. The method for detecting and controlling impurities in high-purity oxygen as described in claim 1, characterized in that, In step S30, the relative concentration is... Real-time temperature data and vibration acceleration data The process of inputting data into a preset edge computing unit, performing impurity concentration denoising correction based on a multi-channel collaborative weighted PCA fusion compensation mechanism in the edge computing unit, and outputting a corrected concentration data set specifically includes: Step S301: Window Construction and Spatiotemporal Alignment: Based on Relative Concentration Concentration sequence C(t) associated with time t is constructed, real-time gas flow rate and TDLAS probe spacing are obtained, and time translation and interpolation processing are performed on concentration sequences C(t) from different locations to form a spatiotemporal alignment window W covering the same gas mass; Step S302: Channel Health Assessment and Co-weighting: Calculate the health weight for each TDLAS detection probe within the spatiotemporal alignment window W. The health weight is determined by a combination of the following dimensional indicators: mean square of spectral line fitting residuals. Temperature drift sensitivity coefficient and vibration sensitivity coefficient And based on health score weights Forming a weighted observation matrix ; Step S303: PCA fusion and perturbation separation: This involves adjusting the weighted observation matrix... Perform principal component analysis to obtain the principal component loading matrix. Combined with the scoring matrix, real-time temperature data and vibration acceleration data Constructing the exogenous feature matrix A principal component working condition regression sub-model is established; where the principal components of the principal component working condition regression sub-model are working condition related components, the reconstructed value of the principal component working condition regression sub-model is the systematic shift caused by temperature and vibration, and the residual R of the principal component working condition regression sub-model is regarded as a mixture of the actual impurity change and the measurement noise after debiasing. Step S304: Based on the residual R of the principal component working condition regression sub-model, the Tukey function is used for correction. The corrected residual is added to the reconstructed value of the principal component working condition regression sub-model to obtain the debiased concentration correction result. Finally, the corrected concentration data set is output.

6. The method for detecting and controlling impurities in high-purity oxygen as described in claim 1, characterized in that, Step S40 involves acquiring historical concentration data from the k-th TDLAS detection probe, fusing the historical concentration data from the output corrected concentration data set, and determining whether the impurity concentration is abnormal and recording abnormal events based on a multi-scale dynamic threshold determination mechanism. Specifically, this includes: Step S401: Obtain the historical concentration data of the k-th TDLAS detection probe, and use Variational Mode Decomposition (VMD) to decompose the historical concentration data of the k-th TDLAS detection probe into low-frequency trend components. With high-frequency fluctuation components Among them, the low-frequency trend component is used to reflect the long-term trend of impurity concentration, and the high-frequency fluctuation component is used to reflect short-term fluctuation characteristics. Step S402: For low-frequency trend components Calculate the mean and standard deviation of the sliding window to form a trend feature vector; for high-frequency fluctuation components... Calculate the energy index and kurtosis index to form a fluctuation characteristic vector; Step S403: Introduce trend fold factor and fluctuation fold factor; based on the trend feature vector and fluctuation feature vector combined with the trend fold factor and fluctuation fold factor, generate a multi-scale dynamic threshold set for each group of TDLAS detection probes; wherein, the multi-scale dynamic threshold set includes low-frequency trend thresholds. and high-frequency fluctuation threshold ; Step S404: When the relative concentration at the k-th TDLAS detection probe position... Low-frequency trend components that simultaneously meet the following three conditions Greater than or equal to the low-frequency trend threshold And high-frequency fluctuation components Greater than or equal to the high-frequency fluctuation threshold When this occurs, it is determined to be a valid abnormal event.

7. The method for detecting and controlling impurities in high-purity oxygen as described in claim 1, characterized in that, In step S50, the step of automatically triggering the preset closed-loop linkage control logic based on the abnormal event recorded by the needle includes: When an abnormal event is detected as a single point or local anomaly, the opening of the vent valve of the corresponding pipeline section is automatically adjusted or the backup oxygen supply channel is switched. When an abnormal event is detected as an abnormal global concentration trend, a flow reduction or shutdown command is sent to the upstream oxygen supply source, and the whole network purging mode is initiated. All control execution commands and sensor status data are uploaded to the cloud monitoring platform in real time. The cloud monitoring platform performs pattern analysis and threshold optimization, and sends updated dynamic threshold parameter sets and control strategies to the edge computing unit to achieve self-learning and adaptive parameter updates.

8. A high-purity oxygen impurity detection and control system, applied to the high-purity oxygen impurity detection and control method according to any one of claims 1 to 7, characterized in that, The impurity detection and control system for high-purity oxygen includes: The probe deployment and environmental monitoring module is used to deploy multiple sets of TDLAS detection probes and corresponding temperature and vibration monitoring units along the target high-purity oxygen delivery pipeline. The TDLAS detection probes are used to measure the intensity of characteristic absorption lines of impurities in the high-purity oxygen delivery pipeline to obtain spectral harmonic signals; the temperature and vibration monitoring units are used to acquire real-time temperature data. and vibration acceleration data ; The spectral demodulation and dynamic inversion module is used to perform harmonic normalization processing on the spectral harmonic signal acquired by the TDLAS detection probe using a wavelength modulation spectral demodulation mechanism to obtain the absorption intensity ratio signal R; based on the absorption intensity ratio signal R and real-time temperature data... and vibration acceleration data The relative concentration of the i-th impurity at the k-th TDLAS detection probe location was calculated using the Beer-Lambert dynamic inversion method with multivariate temperature and pressure perturbation compensation. ; The concentration denoising and fusion compensation module is used to reduce relative concentration. Real-time temperature data and vibration acceleration data The data is input to a preset edge computing unit, where impurity concentration is denoised and corrected based on a multi-channel collaborative weighted PCA fusion compensation mechanism, and the corrected concentration data set is output. The historical data fusion and anomaly detection module is used to acquire the historical concentration data of the k-th TDLAS detection probe, fuse the historical concentration data of the output corrected concentration data set, and determine whether the impurity concentration is abnormal and record the abnormal event based on the multi-scale dynamic threshold judgment mechanism. The abnormal response and closed-loop control module is used to automatically trigger preset closed-loop linkage control logic for recorded abnormal events.

9. A device for detecting and controlling impurities in high-purity oxygen, characterized in that, The high-purity oxygen impurity detection and control device includes: a memory, a processor, and a high-purity oxygen impurity detection and control program stored in the memory and executable on the processor. When the high-purity oxygen impurity detection and control program is executed by the processor, it implements a high-purity oxygen impurity detection and control method according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a high-purity oxygen impurity detection and control program, which, when executed by a processor, implements the features described in any one of claims 1 to 7.

Images