Underground operator breathing flow monitoring method and system based on differential pressure transducer
By installing differential pressure sensors inside breathing masks or pipelines in the downhole working environment, and combining various filtering algorithms and flow calculation models, the signal distortion problem caused by the installation location of traditional equipment is solved, realizing accurate breathing flow monitoring and anomaly detection, and improving the safety and efficiency of downhole workers.
Patent Information
- Application Number
- CN202511378876.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-25
- Publication Date
- 2025-11-11
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In existing downhole working environments, traditional breathing flow monitoring equipment is easily affected by humidity and temperature interference because it is fixed to the outside of the mask, resulting in signal distortion. It is also difficult to adapt to the bending and deformation of the breathing tubing, affecting data accuracy and calculation accuracy. Especially in emergency scenarios, the error is large, affecting the reliability of protection decisions.
The differential pressure sensor is installed inside the breathing mask or in the tubing. The differential pressure signal is denoised and corrected using Gaussian filtering, wavelet transform, adaptive filtering and Kalman filtering algorithms. The flow calculation model is established by fitting with the least squares method. Time series decomposition and sliding window are used to detect abnormal patterns to achieve accurate respiratory flow monitoring.
It improves the accuracy and real-time performance of breathing flow calculation, reduces environmental noise interference, ensures data stability and reliability of anomaly detection under complex downhole conditions, and enhances safety protection capabilities.
Smart Images
Figure CN120918628A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of respiratory monitoring technology, and in particular to a method and system for monitoring the respiratory flow of downhole workers based on a differential pressure sensor. Background Technology
[0002] In underground coal mine operations, ensuring the respiratory safety of workers is paramount, as poor air circulation or the accumulation of harmful gases often leads to sudden accidents that threaten life and property. Respiratory flow monitoring, as a core component of the protective system, directly relates to whether personnel can obtain sufficient oxygen and expel waste gases in a timely manner. Its accuracy and real-time performance have become key factors in improving operational efficiency and safety levels. Traditional respiratory monitoring methods mostly rely on external air sampling equipment or manual recording. While these methods can roughly assess ambient air quality, they ignore dynamic changes in individual breathing and cannot capture subtle airflow fluctuations generated by an individual in narrow tunnels or high-dust environments.
[0003] A deeper problem lies in the fact that these devices are fixed outside the mask, causing airflow paths to be affected by underground humidity and temperature, leading to signal distortion. This is especially true when workers move or bend over, as external sensors struggle to adapt to the bending and deformation of the breathing tubing, amplifying data deviations. This installation limitation further triggers a chain reaction of unstable signal transmission, as underground electromagnetic interference and tubing vibrations interfere with the accuracy of pressure signal acquisition, making it difficult to reliably capture even subtle airflow changes. For example, in a simulated underground operation, the pressure difference signal recorded by the external sensor fluctuated by up to 20% due to slight tubing distortion, directly misleading research involving underground workers. Respiratory health monitoring of underground workers is a crucial aspect of safe production, directly impacting worker safety and efficiency. Insufficient acquisition accuracy further exacerbates the challenges in computation. Existing systems struggle to convert unstable pressure signals into accurate respiratory flow values, particularly in emergency scenarios with accelerated breathing rhythms, where flow estimation errors can exceed 15%, affecting the reliability of protective decisions. Therefore, achieving stable pressure difference acquisition and accurate conversion into flow indicators under complex underground conditions is a key issue in improving the reliability of personal respiratory monitoring. Summary of the Invention
[0004] The purpose of this invention is to provide a method and system for monitoring the breathing flow of downhole workers based on differential pressure sensors, which overcomes the limitations of traditional flow monitoring that requires complex devices and improves the accuracy of breathing flow calculation.
[0005] To achieve the above objectives, the present invention provides the following solution:
[0006] A method for monitoring the respiratory flow of downhole workers based on a differential pressure sensor, comprising:
[0007] Install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to collect the raw pressure difference signal;
[0008] The original pressure difference signal is denoised and corrected to obtain the corrected pressure difference signal;
[0009] According to a preset flow calculation model, the corrected pressure difference signal is converted into corresponding respiratory flow data, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow.
[0010] Anomaly detection is performed on the respiratory flow data to obtain an anomaly alarm signal.
[0011] Optionally, the differential pressure sensor includes two pressure sampling nozzles, which extend through polytetrafluoroethylene conduits to upstream and downstream monitoring points of airflow within the breathing mask or breathing tubing, respectively. The distance between the two monitoring points is 5-15 cm, and the end of the conduit is provided with a dust filter with a pore size of no more than 50 μm.
[0012] Optionally, the denoising and correction processing of the original pressure difference signal includes:
[0013] The original pressure difference signal is filtered using a Gaussian filter to obtain a preliminary filtered signal;
[0014] If the humidity noise amplitude in the preliminary filtered signal exceeds a preset threshold, wavelet transform is used to remove the humidity noise and obtain a humidity-removed signal.
[0015] Based on the period of temperature noise in the humidity removal signal, an adaptive filter is used to remove temperature noise and obtain the temperature removal signal;
[0016] The pressure difference value is extracted from the temperature removal signal by differential calculation to obtain the denoised pressure difference signal;
[0017] The denoised pressure signal is corrected for deviation using a Kalman filter algorithm to determine the corrected pressure difference signal.
[0018] Optionally, the Kalman filter algorithm can be used to correct the deviation, including
[0019] For the denoised pressure signal, calculate the sequence fluctuation amplitude;
[0020] The fluctuation amplitude of the sequence is compared with a preset threshold. If the fluctuation amplitude exceeds the preset threshold, it is determined that there is a deviation.
[0021] From the biased sequence, the bias estimate is extracted, and the bias estimate is processed by the Kalman filter algorithm to obtain the filtered sequence;
[0022] The corrected pressure difference signal is obtained from the filtered sequence.
[0023] Optionally, obtaining the preset traffic calculation model includes:
[0024] The simulation includes breathing airflow correlation parameters such as downhole temperature, humidity and air pressure. Airflow with different known flow rates is generated through a standard flow device, and corresponding pressure difference data is collected simultaneously.
[0025] Establish a database of correspondences between known flow rates and corresponding pressure differences. Based on the database, use the least squares method to fit and obtain the linear or nonlinear parameters of the flow calculation model, and obtain the preset flow calculation model.
[0026] Optionally, performing anomaly detection on the respiratory flow data and obtaining anomaly alarm signals includes:
[0027] The respiratory flow data is decomposed to obtain the peak and trough amplitude characteristics of the respiratory flow data;
[0028] The peak and valley amplitude features are divided into windows of fixed length to obtain feature subsets within the window;
[0029] Analyze the abnormal patterns of the feature subset within the window. If the amplitude deviation of the feature subset within the window exceeds a preset threshold, it is marked as a potential anomaly, and the marked anomaly subset is obtained.
[0030] An abnormal alarm signal is generated based on the marked abnormal subset, and the alarm level is determined by threshold comparison to obtain the abnormal alarm signal.
[0031] Optionally, decomposing the respiratory flow data includes:
[0032] The respiratory flow data was decomposed into trend and seasonal components using a time series decomposition method to obtain the decomposed series.
[0033] The peak and valley positions are obtained from the decomposed sequence, and the peak and valley positions are identified by the local maximum search algorithm to obtain the peak and valley position sequence.
[0034] The peak amplitude and valley amplitude are extracted based on the peak and valley position sequence to obtain peak and valley amplitude features.
[0035] The present invention also discloses a breathing flow monitoring system for downhole workers based on a differential pressure sensor, comprising: a data acquisition module, a data processing module, a breathing flow monitoring module, and an anomaly detection module;
[0036] The data acquisition module is used to install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to acquire the raw pressure difference signal;
[0037] The data processing module is used to perform noise reduction and correction processing on the original pressure difference signal to obtain the corrected pressure difference signal;
[0038] The respiratory flow monitoring module is used to convert the corrected pressure difference signal into corresponding respiratory flow data according to a preset flow calculation model, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow.
[0039] The anomaly detection module is used to detect anomalies in the respiratory flow data and obtain anomaly alarm signals.
[0040] The beneficial effects of this invention are as follows: Addressing the problem of inaccurate airflow control in respiratory protection scenarios caused by environmental noise, individual respiratory differences, and the need for long-term monitoring, this invention acquires raw pressure signals through sensors, combines filtering processing with Kalman filtering algorithms to correct deviations, and obtains a stable pressure difference sequence. This provides core data support for subsequent accurate calculations.
[0041] By simulating downhole parameters, standard flow rate calibration, and least squares fitting, a precise pressure difference-flow rate mapping relationship is established. This significantly improves the accuracy of breathing flow rate calculation, solves the problem of flow rate conversion deviation caused by fluctuations in downhole environmental parameters, and keeps the error between the calculated results and the actual breathing flow rate within a very small range, providing reliable data for assessing the breathing status of operators.
[0042] By extracting peak and trough characteristics and using a sliding window anomaly detection mode, accurate respiratory airflow monitoring and risk assessment are achieved, improving the real-time response capability and safety of protective masks. Attached Figure Description
[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments 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.
[0044] Figure 1 This is a flowchart illustrating a method for monitoring the breathing flow of downhole workers based on a differential pressure sensor, according to an embodiment of the present invention. Detailed Implementation
[0045] 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.
[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0047] like Figure 1 As shown, this embodiment provides a method for monitoring the breathing flow of downhole workers based on a differential pressure sensor, including:
[0048] Install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to collect the raw pressure difference signal;
[0049] The original pressure difference signal is denoised and corrected to obtain the corrected pressure difference signal;
[0050] According to a preset flow calculation model, the corrected pressure difference signal is converted into corresponding respiratory flow data, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow.
[0051] Anomaly detection is performed on the respiratory flow data to obtain an anomaly alarm signal.
[0052] Furthermore, the differential pressure sensor includes two pressure sampling nozzles, which extend through polytetrafluoroethylene conduits to upstream and downstream monitoring points of airflow within the breathing mask or breathing tubing, respectively. The distance between the two monitoring points is 5-15 cm, and the end of the conduit is equipped with a dust filter with a pore size of no more than 50 μm.
[0053] Furthermore, the denoising and correction processing of the original pressure difference signal includes:
[0054] The original pressure difference signal is filtered using a Gaussian filter to obtain a preliminary filtered signal;
[0055] If the humidity noise amplitude in the preliminary filtered signal exceeds a preset threshold, wavelet transform is used to remove the humidity noise and obtain a humidity-removed signal.
[0056] Based on the period of temperature noise in the humidity removal signal, an adaptive filter is used to remove temperature noise and obtain the temperature removal signal;
[0057] The pressure difference value is extracted from the temperature removal signal by differential calculation to obtain the denoised pressure difference signal;
[0058] The denoised pressure signal is corrected for deviation using a Kalman filter algorithm to determine the corrected pressure difference signal.
[0059] Furthermore, the Kalman filter algorithm is used to correct the deviation, including
[0060] For the denoised pressure signal, calculate the sequence fluctuation amplitude;
[0061] The fluctuation amplitude of the sequence is compared with a preset threshold. If the fluctuation amplitude exceeds the preset threshold, it is determined that there is a deviation.
[0062] From the biased sequence, the bias estimate is extracted, and the bias estimate is processed by the Kalman filter algorithm to obtain the filtered sequence;
[0063] The corrected pressure difference signal is obtained from the filtered sequence.
[0064] Specifically, when acquiring the raw pressure signal within the mask tubing, the sensor captures airflow changes. Assuming the sensor acquires pressure data 1000 times per second, the raw signal contains respiratory pressure fluctuations, humidity noise, and temperature noise. The raw signal may exhibit irregular waveforms, such as pressure values fluctuating between 0.5 and 1.5 kPa, but it contains high-frequency noise. A Gaussian filter is used to smooth the raw signal. Setting the filter window size to 5 sampling points and the standard deviation to 1 effectively reduces high-frequency noise while preserving the main trend of the respiratory signal. The initial filtered signal shows smoother pressure fluctuations and a noise amplitude reduction of approximately 30%, but humidity noise can still cause signal offset. For example, humidity noise may cause the pressure value to be 0.2 kPa higher. If the humidity noise amplitude in the initial filtered signal exceeds a preset threshold of 0.1 kPa, wavelet transform is used to remove humidity noise. A Daubechies wavelet basis is selected, decomposed into three layers, and low-frequency components are extracted to filter out high-frequency humidity noise. The humidity-removed signal is closer to the actual respiratory pressure, with the offset reduced to within 0.05 kPa, significantly improving signal purity. An adaptive filter is used to eliminate temperature noise. Temperature noise often exhibits periodic fluctuations, such as 10 low-frequency disturbances per minute. An adaptive filter based on the least mean square algorithm can dynamically track the temperature noise period and adjust the filtering parameters in real time. The fluctuation period of the temperature-removed signal disappears, the pressure value becomes more stable, and the error is controlled within 0.02 kPa. Pressure differences are extracted through differential calculation. The pressure difference sequence between adjacent sampling points reflects changes in respiratory rate; for example, the difference for normal breathing is 0.1 to 0.3 kPa, while rapid breathing may reach 0.5 kPa. The initial pressure difference sequence contains approximately 1000 data points, preserving dynamic characteristics.
[0065] In a 1000-point sequence, a sliding window of length 50 is used to calculate the difference between the mean within each window and the overall mean, resulting in a bias estimate sequence. The bias estimates are assumed to range from 0.05 to 0.15 kPa. A Kalman filter algorithm is then applied to process these bias estimates. The Kalman filter progressively optimizes the sequence data through prediction and update steps. Inputting the bias estimates into the Kalman filter with an initial state covariance of 0.1, process noise of 0.01, and measurement noise of 0.05, the filtered sequence shows a significant reduction in fluctuation amplitude and smoother data points.
[0066] Furthermore, obtaining the preset traffic calculation model includes:
[0067] In a laboratory environment, breathing airflow-related parameters, including downhole temperature, humidity, and air pressure, are simulated. Airflow with different known flow rates in the range of 0-100 L / min is generated using a standard flow device, and corresponding pressure difference data is collected simultaneously.
[0068] Establish a database of correspondences between known flow rates and corresponding pressure differences. Based on the database, use the least squares method to fit and obtain the linear or nonlinear parameters of the flow calculation model, and obtain the preset flow calculation model.
[0069] Furthermore, anomaly detection is performed on the respiratory flow data to obtain anomaly alarm signals, including:
[0070] The respiratory flow data is decomposed to obtain the peak and trough amplitude characteristics of the respiratory flow data;
[0071] The peak and valley amplitude features are divided into windows of fixed length to obtain feature subsets within the window;
[0072] Analyze the abnormal patterns of the feature subset within the window. If the amplitude deviation of the feature subset within the window exceeds a preset threshold, it is marked as a potential anomaly, and the marked anomaly subset is obtained.
[0073] An abnormal alarm signal is generated based on the marked abnormal subset, and the alarm level is determined by threshold comparison to obtain the abnormal alarm signal.
[0074] Furthermore, the decomposition of the respiratory flow data includes:
[0075] The respiratory flow data was decomposed into trend and seasonal components using a time series decomposition method to obtain the decomposed series.
[0076] The peak and valley positions are obtained from the decomposed sequence, and the peak and valley positions are identified by the local maximum search algorithm to obtain the peak and valley position sequence.
[0077] The peak amplitude and valley amplitude are extracted based on the peak and valley position sequence to obtain peak and valley amplitude features.
[0078] Specifically, when using time series decomposition to separate trend and seasonal components, an additive decomposition model can be used. The trend component reflects long-term changes in respiratory flow, such as a slow increase in overall flow over time; the seasonal component captures periodic fluctuations, such as peaks and troughs in each respiratory cycle. Taking a certain dataset as an example, after decomposition, the trend component shows that the flow steadily increases from 2.0 L / min to 2.3 L / min within 10 minutes, while the seasonal component shows regular fluctuations with a respiratory cycle of 4 seconds. This decomposition helps to identify regular patterns in the data. When identifying peaks and troughs using a local maximum search algorithm, a sliding window can be set to analyze local extrema within the window. Within a 5-second window, the algorithm detected a flow peak of 3.0 L / min and a trough of 1.8 L / min, corresponding to the inspiratory and expiratory phases of the respiratory cycle, respectively. The peak and trough position sequence records the timestamps of these extreme points, such as the peak occurring at 2.5 seconds and the trough at 4.0 seconds. This method can accurately locate key points in the respiratory cycle. When extracting amplitude features from the peak and trough position sequence, the flow difference between the peak and trough can be calculated. For example, in a certain sequence, the peak velocity is 3.0 L / min, the trough velocity is 1.8 L / min, and the amplitude difference is 1.2 L / min. The feature set includes the amplitude value of each breath, such as the amplitudes of three consecutive cycles being 1.2 L / min, 1.3 L / min, and 1.1 L / min, respectively. These features reflect the variation pattern of respiratory intensity.
[0079] In one embodiment, when constructing a sliding window dataset, the window length can be set to 10 seconds, the step size to 2 seconds, and multiple subsets can be created. For example, a window may contain 5 peaks and troughs, and the feature subsets may include the corresponding amplitudes and time intervals. This partitioning method facilitates the capture of changes in local patterns. For example, when analyzing abnormal patterns, if the amplitude deviation within a certain window exceeds a preset threshold of 0.5 L / min, such as a sudden increase in amplitude to 1.8 L / min in a certain cycle, far exceeding the average of 1.2 L / min, it is marked as a potential abnormality. This marking helps in the detection of respiratory abnormalities, such as sudden rapid breathing.
[0080] When generating abnormal alarm signals, alarm levels can be defined based on the degree of deviation. For example, a deviation of 0.5-1.0 L / min is a low-level alarm, while a deviation exceeding 1.0 L / min is a high-level alarm. The final alarm signal sequence records the time and level of each abnormality, such as "Time 15 seconds, Low-level alarm". This mechanism can promptly alert to abnormal respiratory states, facilitating subsequent handling.
[0081] Based on the matching of abnormal alarm signals with historical data, if the matching degree is higher than a threshold, a feedback control command is generated to determine the oxygen replenishment adjustment plan. This includes: acquiring abnormal alarm signals; collecting environmental data in real time through sensors to determine the current signal characteristics; extracting historical data patterns from the data storage structure based on the current signal characteristics; performing pattern matching using the k-nearest neighbor algorithm to obtain a matching degree value; if the matching degree value exceeds a preset threshold, generating a feedback control command according to the command generation rules; determining the oxygen replenishment plan based on the feedback control command and adjustment plan parameters; acquiring the oxygen replenishment plan; adjusting the signal acquisition frequency; updating the historical data samples in the data storage structure; optimizing the alarm triggering conditions based on the updated historical data samples to obtain a new triggering threshold; and monitoring abnormal alarm signals using the new triggering threshold to determine subsequent signal characteristics.
[0082] In one possible implementation, historical data patterns are extracted from a data storage structure. The data storage structure uses a time-series database, storing oxygen concentration, temperature, and humidity data for the past 30 days. Historical data from the last 7 days is extracted to form a pattern library, containing a daily average oxygen concentration of 20% and a fluctuation range of ±2%. The k-nearest neighbor algorithm is used to calculate the Euclidean distance between the current signal features and historical patterns, obtaining a matching degree value.
[0083] This embodiment also provides a downhole worker breathing flow monitoring system based on a differential pressure sensor, including: a data acquisition module, a data processing module, a breathing flow monitoring module, and an anomaly detection module;
[0084] The data acquisition module is used to install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to acquire the raw pressure difference signal;
[0085] The data processing module is used to perform noise reduction and correction processing on the original pressure difference signal to obtain the corrected pressure difference signal;
[0086] Specifically, a sensor array collects raw pressure signals within the mask's tubing. These signals are then filtered using a Gaussian filter to obtain a preliminary filtered signal. If the humidity noise amplitude in the preliminary filtered signal exceeds a preset threshold, wavelet transform is used to remove the humidity noise, resulting in a humidity-removed signal. Based on the periodicity of temperature noise in the humidity-removed signal, an adaptive filter is applied to eliminate temperature noise, yielding a temperature-removed signal.
[0087] For the denoised pressure difference sequence, the sequence fluctuation amplitude is calculated. The fluctuation amplitude is compared with a preset threshold; if the fluctuation amplitude exceeds the preset threshold, a deviation is identified. From the sequences identified as having deviations, deviation estimates are extracted. The deviation estimates are then processed using a Kalman filter algorithm to obtain a filtered sequence. From the filtered sequence, the corrected pressure difference value is determined.
[0088] The respiratory flow monitoring module is used to convert the corrected pressure difference signal into corresponding respiratory flow data according to a preset flow calculation model, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow.
[0089] Specifically, in a laboratory environment, breathing airflow parameters such as downhole temperature, humidity, and air pressure are simulated. Airflow with different known flow rates in the range of 0-100 L / min is generated through a standard flow device. Pressure difference signals output by differential pressure sensors are collected simultaneously to establish a database of the correspondence between pressure difference and known flow rate. Based on the database, the least squares method is used to fit and obtain the linear or nonlinear parameters of the flow calculation model.
[0090] The anomaly detection module is used to detect anomalies in the respiratory flow data and obtain anomaly alarm signals.
[0091] Specifically, raw sequence data is obtained from traffic data, and time series decomposition methods are used to separate trend and seasonal components, resulting in a decomposed sequence. Peak and trough positions are determined from the decomposed sequence, and a local maximum search algorithm is used to identify peaks and troughs, resulting in a peak-trough position sequence. Peak and trough amplitudes are extracted from the peak-trough position sequence, and amplitude differences are calculated based on the position sequence to obtain a peak-trough amplitude feature set. A sliding window dataset is constructed from the peak-trough amplitude feature set, and the sequence is divided into feature subsets within the window using a fixed-length window. Abnormal patterns are analyzed from the feature subsets within the window; if the amplitude deviation in the subset exceeds a preset threshold, it is marked as a potential anomaly, resulting in a marked anomaly subset. Anomaly alarm signals are generated from the marked anomaly subset, and the alarm level is determined using threshold comparison logic, resulting in the final alarm signal sequence.
[0092] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.
Claims
1. A method for monitoring the breathing flow of downhole workers based on a differential pressure sensor, characterized in that, include: Install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to collect the raw pressure difference signal; The original pressure difference signal is denoised and corrected to obtain the corrected pressure difference signal; According to a preset flow calculation model, the corrected pressure difference signal is converted into corresponding respiratory flow data, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow. Anomaly detection is performed on the respiratory flow data to obtain an anomaly alarm signal.
2. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 1, characterized in that, The differential pressure sensor includes two pressure sampling nozzles, which are respectively extended through polytetrafluoroethylene conduits to upstream and downstream monitoring points of airflow in the breathing mask or breathing tubing. The distance between the two monitoring points is 5-15cm, and the end of the conduit is provided with a dust filter with a pore size of no more than 50μm.
3. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 1, characterized in that, The denoising and correction processing of the original pressure difference signal includes: The original pressure difference signal is filtered using a Gaussian filter to obtain a preliminary filtered signal; If the humidity noise amplitude in the preliminary filtered signal exceeds a preset threshold, wavelet transform is used to remove the humidity noise and obtain a humidity-removed signal. Based on the period of temperature noise in the humidity removal signal, an adaptive filter is used to remove temperature noise and obtain the temperature removal signal. The pressure difference value is extracted from the temperature removal signal by differential calculation to obtain the denoised pressure difference signal; The denoised pressure signal is corrected for deviation using a Kalman filter algorithm to determine the corrected pressure difference signal.
4. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 3, characterized in that, The Kalman filter algorithm is used to correct the deviation, including For the denoised pressure signal, calculate the sequence fluctuation amplitude; The fluctuation amplitude of the sequence is compared with a preset threshold. If the fluctuation amplitude exceeds the preset threshold, it is determined that there is a deviation. From the biased sequence, the bias estimate is extracted, and the bias estimate is processed by the Kalman filter algorithm to obtain the filtered sequence; The corrected pressure difference signal is obtained from the filtered sequence.
5. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 1, characterized in that, Obtaining the preset traffic calculation model includes: The simulation includes breathing airflow correlation parameters such as downhole temperature, humidity and gas pressure. Airflow with different known flow rates is generated through a standard flow device, and corresponding pressure difference data is collected simultaneously. Establish a database of correspondences between known flow rates and corresponding pressure differences. Based on the database, use the least squares method to fit and obtain the linear or nonlinear parameters of the flow calculation model, and obtain the preset flow calculation model.
6. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 1, characterized in that, Anomaly detection of the respiratory flow data and acquisition of anomaly alarm signals include: The respiratory flow data is decomposed to obtain the peak and trough amplitude characteristics of the respiratory flow data; The peak and valley amplitude features are divided into windows of fixed length to obtain feature subsets within the window; Analyze the abnormal patterns of the feature subset within the window. If the amplitude deviation of the feature subset within the window exceeds a preset threshold, it is marked as a potential anomaly, and the marked anomaly subset is obtained. An abnormal alarm signal is generated based on the marked abnormal subset, and the alarm level is determined by threshold comparison to obtain the abnormal alarm signal.
7. The method for monitoring the breathing flow of downhole workers based on a differential pressure sensor according to claim 1, characterized in that, Decomposing the respiratory flow data includes: The respiratory flow data was decomposed into trend and seasonal components using a time series decomposition method to obtain the decomposed series. The peak and valley positions are obtained from the decomposed sequence, and the peak and valley positions are identified by the local maximum search algorithm to obtain the peak and valley position sequence. The peak amplitude and valley amplitude are extracted based on the peak and valley position sequence to obtain peak and valley amplitude features.
8. A downhole worker breathing flow monitoring system based on a differential pressure sensor, used to implement the downhole worker breathing flow monitoring method based on a differential pressure sensor as described in any one of claims 1-7, characterized in that, include: Data acquisition module, data processing module, respiratory flow monitoring module, and anomaly detection module; The data acquisition module is used to install the differential pressure sensor inside the breathing mask or breathing tubing of the downhole worker to acquire the raw pressure difference signal; The data processing module is used to perform noise reduction and correction processing on the original pressure difference signal to obtain the corrected pressure difference signal; The respiratory flow monitoring module is used to convert the corrected pressure difference signal into corresponding respiratory flow data according to a preset flow calculation model, wherein the preset flow calculation model is a correlation mapping model between pressure difference and respiratory flow. The anomaly detection module is used to detect anomalies in the respiratory flow data and obtain anomaly alarm signals.