A cloud-based oxygen humidification status management system

Abstract

This invention relates to the field of process control technology, specifically a cloud-based oxygen humidification status management system. The system includes an interface visual monitoring module, a spectrum feature extraction module, a flow mapping inversion module, a deviation calculation module, and a valve execution control module. In this invention, by acquiring video streams of the bubble generation area and extracting grayscale time-domain waveforms, non-contact visual information is transformed into frequency-domain features reflecting the bubble oscillation rhythm. Discrete Fourier transform is used to analyze the physical mapping relationship between bubble generation frequency and gas flow rate, achieving non-invasive and accurate inversion of oxygen volumetric flow rate. This eliminates errors from manual readings and mechanical wear interference. Combined with proportional-integral-differential logic, flow deviation is calculated in real time, driving the electronic actuator to automatically correct valve opening, ensuring that the oxygen supply flow rate strictly matches the target standard, effectively improving the automation level and control stability of the humidification oxygen supply process.

Description

Technical Field

[0001] This invention relates to the field of process control technology, and in particular to a cloud-based oxygen humidification status management system. Background Technology

[0002] The field of process control technology encompasses a collection of technologies that utilize control theory and automated instruments to detect and regulate physical quantities during production or operation. Core aspects include real-time acquisition and closed-loop control of key parameters such as temperature, pressure, flow rate, liquid level, and composition. Systems typically consist of sensors, controllers, and actuators. Data interaction is achieved through the conversion of analog and digital signals. Based on preset control logic, such as PID control strategies, deviation values ​​are calculated, which then drive regulating valves, frequency converters, or pumps to correct process variables and ensure system operation meets specific process requirements. Traditional oxygen humidification status management systems refer to technologies that monitor humidification fluid consumption and gas path patency in oxygen therapy. These systems employ mechanical float flow meters in conjunction with transparent bubble-type humidification bottles. Medical personnel visually observe the scale lines on the humidification bottle to determine the level of sterile water or saline solution inside. The oxygen flow rate is read based on the float's position in the conical tube, and the humidification status is roughly assessed by observing the density of bubbles escaping from the inlet. When the fluid level is depleted, the flow switch is manually turned off, and the bottle is disassembled for refilling.

[0003] Traditional oxygen humidification management relies on medical staff to visually observe the position of the buoy and the density of escaping bubbles to determine the oxygen flow rate. Manual visual estimation is easily affected by the observation angle and subjective experience, leading to reading deviations. Furthermore, mechanical buoy flow meters cannot accurately quantify the nonlinear relationship between the bubble generation rate and the small fluctuations in gas flow. Relying solely on rough visual assessment cannot capture real-time dynamic changes in flow. Manual valve adjustment is slow to respond and lacks accuracy, making it impossible to maintain a constant oxygen supply flow rate. The high-frequency manual inspection and intervention mode not only increases the workload of medical staff but also poses operational risks such as depletion of humidifying fluid or abnormal oxygen supply due to inspection intervals. Summary of the Invention

[0004] To address the limitations of existing technologies, such as susceptibility to reading deviations due to observation angles and subjective experience, the difficulty of accurately quantifying the nonlinear relationship between bubble generation rate and minute fluctuations in gas flow rate using mechanical float flowmeters, the inability to capture real-time dynamic changes in flow rate through rough visual assessment, and the inability to maintain a constant oxygen supply flow rate due to lag in response and inconsistent accuracy of manual valve adjustment, coupled with the high-frequency manual inspection and intervention methods that increase the workload of medical staff and pose operational risks such as depletion of humidification fluid or abnormal oxygen supply due to inspection intervals, this invention provides a cloud-based oxygen humidification status management system. The technical solution is as follows. On the one hand, a cloud-based oxygen humidification status management system is provided, which includes: The interface visual monitoring module collects real-time video stream data output by the visual sensor, delineates a rectangular monitoring area covering the area above the nozzle of the duct in the cloud server, calculates the arithmetic mean of the grayscale data of the pixels in the rectangular monitoring area, and generates a grayscale time-domain waveform signal of the gas-liquid interface. The spectral feature extraction module performs a discrete Fourier transform on the gray-scale time-domain waveform signal of the gas-liquid interface, converting the gray-scale information fluctuating in the time dimension into energy spectrum data distributed in the frequency dimension, and obtaining the fundamental frequency index of the bubble oscillation. The flow mapping inversion module uses the fundamental frequency index of the bubble oscillation as an index to obtain the fluid dynamics mapping table pre-stored in the cloud server, including the nonlinear correspondence between frequency data and standard gas volume flow rate levels, retrieves the oxygen supply at the current moment, and obtains real-time oxygen volume flow rate data. The deviation calculation module compares the real-time oxygen volumetric flow rate data with the preset target flow rate standard, calculates the deviation amplitude and polarity between the real-time oxygen volumetric flow rate data and the target flow rate standard, generates digital instructions for adjusting the valve opening, and establishes a valve opening correction digital signal.

[0005] As a further embodiment of the present invention, the gas-liquid interface grayscale time-domain waveform signal includes discrete time node coordinates, instantaneous average grayscale amplitude, and waveform sequence length information; the bubble oscillation fundamental frequency index includes the main peak frequency value, power spectral density peak intensity, and frequency resolution parameter; the real-time oxygen volumetric flow rate data includes the instantaneous flow velocity value, gas volume measurement unit, and flow rate confidence interval; and the valve opening correction digital signal includes the number of step drive pulses, adjustment direction identifier, and correction amplitude encoding.

[0006] As a further aspect of the present invention, the interface visual monitoring module includes: The grayscale extraction submodule of the monitoring area collects real-time video stream data output by the visual sensor. In the cloud server, a rectangular monitoring area covering the area above the nozzle of the duct is defined for the real-time video stream data. The two-dimensional coordinate positions of the pixels inside the rectangular monitoring area are traversed, and the grayscale data values ​​corresponding to the pixels are extracted one by one. The arithmetic average calculation operation is performed on all grayscale data within the range of a single frame image to obtain the instantaneous average grayscale value of the monitoring area. The transmittance sequence construction submodule reads the frame generation timestamps generated along with the instantaneous average gray value of the monitoring area based on the instantaneous average gray value of the monitoring area, and stores multiple instantaneous average gray values ​​of the monitoring area into a linear buffer queue in sequence according to the order of the frame generation timestamps to establish a gas-liquid two-phase interface transmittance sequence. The grayscale time-domain waveform generation submodule performs signal waveform processing, sets the frame generation timestamp as the time-domain horizontal axis coordinate, sets the numerical amplitude in the gas-liquid two-phase interface transmittance sequence as the vertical axis intensity coordinate, maps discrete sequence data points to a two-dimensional coordinate system, connects adjacent data points through interpolation to form a continuously changing fluctuation curve, calculates the weighted smooth value of the interpolated grayscale trajectory, performs signal formatting and encapsulation on the fluctuation curve, and generates a gas-liquid interface grayscale time-domain waveform signal.

[0007] As a further aspect of the present invention, the spectral feature extraction module includes: The frequency domain transformation submodule calls the gray-scale time-domain waveform signal of the gas-liquid interface, performs time-frequency domain transformation operation, and resolves the gray-scale amplitude fluctuation information originally distributed along the time axis into spectral components distributed along the frequency axis, calculates the square value of the modulus corresponding to the frequency component, and generates the gray-scale fluctuation frequency domain energy spectrum. The frequency band locking submodule sets the bandpass filter boundary based on the gray-scale fluctuation frequency domain energy spectrum and the preset frequency range of the bubble generation physical process. It delineates the target range of bubble oscillation energy on the frequency axis of the energy spectrum data, separates the local spectrum data segments corresponding to the bubble generation rhythm, removes the frequency band interference of background noise, and establishes the bubble generation feature screening frequency band. The fundamental frequency analysis submodule filters frequency bands based on the bubble generation characteristics, performs extreme value search operations on discrete spectral data, compares the power spectral density values ​​corresponding to the frequency points within the frequency band one by one, locates the spectral peak position of the peak energy density, reads the horizontal axis frequency value corresponding to the spectral peak position as the core parameter for quantifying the periodic generation rate of bubbles, and obtains the fundamental frequency index of bubble oscillation.

[0008] As a further aspect of the present invention, the execution of the time-frequency domain conversion operation is specifically defined as: performing discrete frequency domain transformation on the gas-liquid interface grayscale time-domain waveform signal within a preset sampling window length, and retaining only the spectral data on the non-negative frequency axis of the transformation result; The calculation of the square value of the modulus corresponding to the frequency component is specifically limited to squaring the complex spectral amplitude of each discrete frequency point and forming a one-to-one energy value sequence. The preset frequency range setting bandpass filter boundary for the combined bubble generation physical process is specifically limited to setting the lower frequency limit and the upper frequency limit to preset lower frequency limit values ​​and upper frequency limit values, respectively, and only extracting the grayscale fluctuation frequency domain energy spectrum data between the lower frequency limit and the upper frequency limit.

[0009] As a further aspect of the present invention, the traffic mapping inversion module includes: The mapping table node positioning submodule calls the bubble oscillation fundamental frequency index, accesses the pre-constructed fluid dynamics mapping table which includes the nonlinear relationship between frequency and flow rate, performs a numerical comparison operation in the frequency index column of the fluid dynamics mapping table, determines the numerical difference between the bubble oscillation fundamental frequency index and the discrete frequency points in the fluid dynamics mapping table, determines the numerical range of the bubble oscillation fundamental frequency index on the frequency axis, locks the standard frequency node row number position corresponding to the upper and lower boundaries of the numerical range, and establishes the adjacent frequency node index. The flow calibration data extraction submodule reads the standard frequency node values ​​and the standard gas volume flow rate level mapped by the node from the fluid dynamics mapping table based on the adjacent frequency node index, and combines the standard frequency node boundary with the standard gas volume flow rate level to obtain the reference flow calibration data. The flow interpolation calculation submodule calculates the magnitude of the deviation of the bubble oscillation fundamental frequency index from the low-frequency boundary point based on the reference flow calibration data, converts the magnitude of the deviation into the corresponding flow increment step, and accumulates the flow increment step to the flow reference value corresponding to the low-frequency boundary point to obtain real-time oxygen volumetric flow rate data.

[0010] As a further aspect of the present invention, the deviation calculation module includes: The flow deviation quantification submodule reads the preset target flow standard through the real-time oxygen volume flow rate data, performs numerical subtraction, calculates the difference between the real-time oxygen volume flow rate data and the target flow standard, extracts the absolute value of the difference as the deviation amplitude, identifies the positive or negative sign of the difference to determine the state attribute of the flow rate being in positive or negative deviation, and generates an instantaneous flow deviation vector. Based on the instantaneous flow deviation vector, the feedback control submodule calculates the real-time amplitude term, the time integral term of the deviation, and the differential rate of change term of the deviation over time of the instantaneous flow deviation vector. It then calls the preset proportional coefficient, integral time constant, and differential time constant to perform a linear weighted summation calculation on the above three deviation characteristics to obtain the theoretical control output value required to offset the flow error, determine the logical control quantity of the regulating valve core displacement, and obtain the valve stroke adjustment quantity. The correction instruction generation submodule, for the valve stroke adjustment amount, maps the logic control amount into the number of step pulses based on the stepper motor drive characteristics and mechanical transmission ratio parameters of the electronic flow control valve, sets the motor rotation direction identifier bit according to the sign bit of the adjustment amount, and encapsulates the pulse number information and direction identifier bit into a control instruction data packet according to the predetermined underlying digital communication protocol format, and establishes a valve opening correction digital signal.

[0011] As a further aspect of the present invention, the calculation of the difference between the real-time oxygen volumetric flow rate data and the target flow rate standard is specifically limited to: comparing the real-time oxygen volumetric flow rate data with the target flow rate standard one by one, extracting the difference magnitude by using absolute value calculation, and determining whether the flow rate is in a positive or negative deviation state by comparing the calculated difference with a preset threshold. The generation of the instantaneous flow deviation vector is specifically defined as follows: based on the absolute value of the difference, the positive and negative signs of the difference are identified, and corresponding flow deviation vectors are generated respectively; the flow adjustment requirements are determined according to the change in the magnitude of the difference.

[0012] As a further aspect of the present invention, the system also includes a valve execution and control module: The valve control module sends the valve opening correction digital signal to the front-end device, drives the valve core of the electronic flow control valve to produce mechanical displacement, adjusts the oxygen flow rate input to the humidification bottle, verifies the consistency between the real-time flow rate and the target flow rate standard, and generates a closed-loop control state for humidification oxygen supply. The closed-loop control status of the humidification oxygen supply includes the steady-state convergence indicator of the flow rate, the dynamic following error rate, and the operation mode feedback information.

[0013] As a further aspect of the present invention, the valve control module includes: The drive signal execution submodule calls the valve opening correction digital signal, parses the pulse count and direction identifier, controls the motor drive bridge to conduct in sequence to generate electromagnetic torque, drives the stepper motor rotor of the electronic flow control valve to rotate, and converts the rotational torque into the linear feed action of the valve core axially through the mechanical transmission component, changes the physical position of the valve core relative to the valve seat, and obtains the valve core physical displacement increment. The fluid resistance control submodule, based on the physical displacement increment of the valve core, corrects the flow area of ​​the throttling orifice inside the valve, changes the local flow resistance coefficient when oxygen flows through the pipeline, adjusts the volumetric flow rate of the gas entering the humidification bottle under the action of the gas source pressure, performs mechanical intervention and regulation of the physical gas path flow rate, and establishes the inlet flow rate after regulation. The closed-loop performance verification submodule performs numerical matching analysis between the adjusted inlet flow rate and the target flow rate standard, calculates the instantaneous residual value after the adjustment action is completed, and compares the instantaneous residual value with the preset steady-state convergence judgment threshold. If the residual value is less than the steady-state convergence judgment threshold, it is judged as a locked state; otherwise, it is marked as a dynamic following state, generating a humidification oxygen supply closed-loop control state.

[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By acquiring video streams of the bubble generation area and extracting grayscale time-domain waveforms, non-contact visual information is transformed into frequency domain features reflecting the bubble oscillation rhythm. Discrete Fourier transform is used to analyze the physical mapping relationship between bubble generation frequency and gas flow rate, achieving non-invasive and accurate inversion of oxygen volumetric flow rate. This eliminates errors from manual readings and mechanical wear interference. Combined with proportional-integral-differential logic, flow deviation is calculated in real time, driving the electronic actuator to automatically correct valve opening. A high-response-speed closed-loop feedback regulation mechanism is constructed to ensure that the oxygen supply flow rate is strictly consistent with the target standard, effectively improving the automation level and control stability of the humidification oxygen supply process. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0016] Figure 1 A schematic diagram of the system provided by the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the interface visual monitoring module in this invention; Figure 4 This is a flowchart of the spectral feature extraction module in this invention; Figure 5 This is a flowchart of the flow mapping inversion module in this invention; Figure 6 This is a flowchart of the deviation calculation module in this invention; Figure 7 This is a flowchart of the valve control module in this invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] This invention provides a cloud-based oxygen humidification status management system, such as... Figure 1-2 The diagram shown illustrates a cloud-based oxygen humidification status management system, which includes: The interface visual monitoring module collects real-time video stream data output by the visual sensor, delineates a rectangular monitoring area covering the area above the nozzle of the duct in the cloud server, calculates the arithmetic mean of the grayscale data of the pixels in the rectangular monitoring area, arranges the arithmetic mean in the order of the frame generation time, constructs a sequence of the dynamic change of light transmittance of the gas-liquid interface over time, and generates a time-domain waveform signal of grayscale of the gas-liquid interface. The spectrum feature extraction module performs discrete Fourier transform on the gray-scale time-domain waveform signal of the gas-liquid interface, converting the gray-scale information fluctuating in the time dimension into energy spectrum data distributed in the frequency dimension. In the energy spectrum data, a screening interval corresponding to the physical bubble generation rhythm is set, and the frequency point of the peak power spectral density within the screening interval is searched to obtain the fundamental frequency index of bubble oscillation. The flow mapping inversion module uses the fundamental frequency index of bubble oscillation as an index to obtain the fluid dynamics mapping table pre-stored in the cloud server, including the nonlinear correspondence between frequency data and standard gas volume flow rate levels, retrieves the oxygen supply at the current moment, and obtains real-time oxygen volume flow rate data. The deviation calculation module compares the real-time oxygen volume flow rate data with the preset target flow rate standard, calculates the deviation magnitude and polarity between the real-time oxygen volume flow rate data and the target flow rate standard, and generates digital instructions for adjusting the valve opening by combining proportional-integral-derivative control logic, and establishes a valve opening correction digital signal. The valve control module sends the valve opening correction digital signal to the front-end device, drives the valve core of the electronic flow control valve to produce mechanical displacement, adjusts the oxygen flow rate input to the humidification bottle, verifies the consistency between the real-time flow rate and the target flow rate standard, and generates a closed-loop control state for humidification oxygen supply. The gas-liquid interface grayscale time-domain waveform signal includes discrete time node coordinates, instantaneous average grayscale amplitude, and waveform sequence length information. The fundamental frequency index of bubble oscillation includes the main peak frequency value, peak intensity of power spectral density, and frequency resolution parameter. The real-time oxygen volumetric flow rate data includes instantaneous flow rate value, gas volume measurement unit, and flow rate confidence interval. The valve opening correction digital signal includes the number of step drive pulses, adjustment direction identifier, and correction amplitude encoding. The humidification oxygen supply closed-loop control status includes the steady-state convergence indicator of flow rate, dynamic following error rate, and operation mode feedback information.

[0023] Specifically, such as Figure 2 , 3 As shown, the interface visual monitoring module includes: The grayscale extraction submodule of the monitoring area collects real-time video stream data output by the visual sensor. In the cloud server, a rectangular monitoring area covering the area above the nozzle of the duct is defined for the real-time video stream data. The two-dimensional coordinate positions of the pixels inside the rectangular monitoring area are traversed, and the grayscale data values ​​corresponding to the pixels are extracted one by one. The arithmetic average calculation operation is performed on all grayscale data within the range of a single frame image to obtain the instantaneous average grayscale value of the monitoring area. Set the industrial camera's acquisition resolution to pixels and frame rate The original image data is used to ensure that the requirements for high-frequency bubble oscillation (maximum) are met. To meet the Nyquist sampling requirements, a Cartesian coordinate system with the center of the duct nozzle as the origin is established in the image processing unit of the cloud server, and the horizontal coordinates are defined. And the vertical coordinate The rectangular region is designated as the region of interest (ROI), which covers the area above the duct nozzle. The physical spatial range of pixels is determined by initiating a double-loop program to traverse the entire rectangular monitoring area. The two-dimensional coordinates of each pixel are used to read the grayscale conversion function. The corresponding 8-bit depth grayscale value The extracted grayscale values ​​are stored in a temporary accumulator register, and a summation operation is performed. If the total grayscale value of the pixels in the current frame is... Then use the formula (in The value is calculated based on the total number of pixels. The numerical value represents the overall light transmittance characteristics of the gas-liquid interface at the current moment, and the instantaneous average gray value of the monitoring area is obtained.

[0024] The transmittance sequence construction submodule reads the frame generation timestamps generated along with the instantaneous average gray value of the monitoring area based on the instantaneous average gray value of the monitoring area, and stores the instantaneous average gray values ​​of multiple monitoring areas into a linear buffer queue in sequence according to the order of the frame generation timestamps to establish the transmittance sequence of the gas-liquid two-phase interface. Initialize a capacity of The system uses a first-in-first-out (FIFO) double-ended circular queue, and the system clock records the moment each frame of image capture is completed with microsecond-level precision. The sampling rate, frame interval is For example, the time when the first frame is generated The second frame was generated at the time. The calculated instantaneous average gray value With timestamp Composition data pairs Pushed to the end of the queue, as the video stream continues to be input, the queue stores data sequentially. , For consecutive data points, the earliest time point data is automatically removed when the queue is full, and a time period of approximately seconds is maintained in memory. A dataset of time-grayscale coordinates was used to establish a gas-liquid two-phase interface transmittance sequence.

[0025] The grayscale time-domain waveform generation submodule performs signal waveform processing, sets the frame generation timestamp as the time-domain horizontal axis coordinate, sets the numerical amplitude in the gas-liquid two-phase interface transmittance sequence as the vertical axis intensity coordinate, maps discrete sequence data points to a two-dimensional coordinate system, connects adjacent data points through interpolation to form a continuously changing fluctuation curve, calculates the weighted smooth value of the interpolated grayscale trajectory, performs signal formatting and encapsulation on the fluctuation curve, and generates a gas-liquid interface grayscale time-domain waveform signal. The weighted smoothing value of the interpolated grayscale trajectory is calculated using the following formula: ; in, This represents the weighted smoothing value of the interpolated grayscale trajectory. This represents the instantaneous average grayscale value of the monitoring area at a real-time time point. This represents the instantaneous average grayscale value of the monitoring area at the next adjacent time point. Represents the first sampling point within a local sampling window centered on the current time node. The grayscale data point values ​​at each location, This represents the arithmetic mean of the grayscale data points within the local sampling window. Represents the total number of data points; Formula calculation logic: The first part of the formula uses... That is, the root mean square (RMS) operation. The logic is that the bubble movement belongs to an energy oscillation process. Compared with the arithmetic mean, the RMS can better retain the energy characteristics of the signal and avoid signal distortion caused by the rapid cancellation of high and low gray levels; the second part Calculate the degree of dispersion of the local gray level relative to the overall mean value using the absolute value (i.e., the mean absolute deviation MAD), and normalize it by dividing the total gray level intensity. Logically, a "turbulence coefficient" is constructed. The larger the coefficient, the more unstable the interface; multiply by That is, the change amplitude between the current frame and the next frame, which means that when the interface is unstable (the turbulence coefficient is large), a larger correction amount needs to be superimposed on the interpolation result to dynamically adapt to the texture change of the bubble surface; The meaning of the parameters and the formula calculation process: Represents the instantaneous average gray level value of the monitoring area at the current time node (e.g., ) in the transmittance sequence of the gas-liquid two-phase interface, and the value is obtained from the mean calculation of the pixel gray levels in the monitoring area; Represents the instantaneous average gray level value of the monitoring area at the next time node adjacent in time sequence (e.g., ) in the transmittance sequence of the gas-liquid two-phase interface, reflecting the overall light transmittance of the next frame; Represents the local average gray level value in the th sub-region after dividing the monitoring area ROI into sub-regions, which is used to characterize the local spatial non-uniformity of the gas-liquid interface (such as the bubble surface texture or micro-turbulence); Represents the arithmetic mean of the local gray level values of these sub-regions, which numerically approximates Represents the total number of sub-region divisions, which is a dimensionless integer constant; Call the defined pixel ROI area, extract the "instantaneous average gray level value of the monitoring area" as the assignment of the current node parameter , the system reads the data corresponding to the next frame timestamp in the first-in-first-out queue to obtain the gray level value as the assignment of the parameter . In order to obtain the local feature parameter , the image processing unit divides the ROI area into four equal-area rectangular sub-blocks in a "field" shape logic (i.e., set ), calculates the average gray level in each sub-block respectively. After real-time pixel statistics, the gray level values of the four sub-blocks are measured as , , , (See Table 1 for data) The system calculates the local arithmetic mean based on these four sampled values. For the fixed constants in the formula, the system is set. The grayscale data are all 8-bit unsigned integers (range 0-255), which are used as floating-point numbers in the calculation to ensure precision. Table 1. Sampling table of grayscale interpolation operation parameters; As shown in Table 1, the key input data and their sources required to perform this interpolation operation are listed. All grayscale values ​​are retained to two decimal places for floating-point operations. Formula calculation example calculation: Based on the real-time data collected in Table 1, the parameters are substituted into the formula for step-by-step calculation: Step 1: Calculate the Energy Reference Term (RMS) and Sum of squares: ; Take the average value: ; Taking the square root yields the baseline value: ; Step 2: Calculate the local fluctuation correction coefficient. Calculate the absolute deviation of each sub-region relative to the mean and sum them: ; Calculate the normalized denominator: ; The volatility coefficient is obtained as follows: ; Step 3: Calculate the span-weighted sum and the final result to calculate the inter-frame variation amplitude. ; Calculate the correction increment: ; Merge result: ; Calculation results (Retain two decimal places) and the result of traditional arithmetic mean linear interpolation ( Compared to ), there exists The numerical offset results show that, due to the local gray-level inhomogeneity within the monitoring area (i.e., texture fluctuations or slight deformations on the bubble surface), the system automatically identifies that the signal at that moment contains high "fluctuation energy." Therefore, it assigns a slightly higher weight than the geometric center during the interpolation process, making the generated waveform closer to the physical reality of the bubble's dynamic changes. The advantage of the formula is that, by introducing root mean square operation and local variance weighting mechanism, it solves the peak clipping and energy loss problems caused by traditional linear interpolation when processing high-frequency bubble oscillation signals. This allows the generated gray-level time-domain waveform signal to retain the bubble's micro-vibration characteristics, providing high-fidelity raw data support for frequency domain analysis.

[0026] Specifically, such as Figure 2 , 4 As shown, the spectral feature extraction module includes: The frequency domain transformation submodule calls the gray-scale time-domain waveform signal of the gas-liquid interface, performs time-frequency domain transformation operation, and resolves the gray-scale amplitude fluctuation information originally distributed along the time axis into spectral components distributed along the frequency axis. It calculates the square value of the modulus corresponding to the frequency component and generates the gray-scale fluctuation frequency domain energy spectrum. Set the window size for the Fast Fourier Transform (FFT) to... The Hanning window function is applied to the time-domain signal to suppress spectral leakage, and then a discrete Fourier transform operation is performed to convert the time-domain signal... Convert to frequency domain complex sequence For each frequency component Extract its actual part With the imaginary part Using the formula Calculate the power energy value at that frequency point, for example, at that frequency point. The real part is calculated to be The imaginary part is Then the squared value of the modulus at that point is According to the Nyquist sampling theorem, the effective analysis bandwidth of the system is to By traversing the frequency points within this range, a gray-scale fluctuation frequency domain energy spectrum is generated.

[0027] The frequency band locking submodule is based on the gray-scale fluctuation frequency domain energy spectrum. It sets the bandpass filter boundary by combining the preset frequency range of the bubble generation physical process. It delineates the target range of bubble oscillation energy on the frequency axis of the energy spectrum data, separates the local spectrum data segments corresponding to the bubble generation rhythm, removes the frequency band interference of background noise, and establishes the bubble generation feature screening frequency band. Referring to the effective oscillation frequency range determined by the physical experiment of bubble formation, the low-frequency cutoff threshold is set as follows: To filter out low-frequency drift caused by liquid sloshing, a high-frequency cutoff threshold is set as follows: To eliminate high-frequency thermal noise from electronic circuits, a rectangular window function is constructed on the frequency axis of the energy spectrum. When frequency hour ,otherwise The original energy spectrum data is multiplied point-by-point with the window function, retaining... to The effective spectral components and their corresponding energy intensity data within the interval will and The amplitudes of the above spectral components are forced to zero to establish a bubble generation feature screening frequency band.

[0028] The fundamental frequency analysis submodule filters frequency bands based on bubble generation characteristics, performs extreme value search operations on discrete spectral data, compares the power spectral density values ​​corresponding to the frequency points within the frequency band one by one, locates the peak position of the peak energy density, reads the horizontal axis frequency value corresponding to the peak position, and uses it as the core parameter for quantifying the periodic generation rate of bubbles to obtain the fundamental frequency index of bubble oscillation. Initialize peak search variables and corresponding frequencies A hill-climbing algorithm is used to traverse and filter power spectral density data within the frequency band, comparing the power spectral density values ​​corresponding to each frequency point within the band. For example, the system scan found that: The energy is ,exist Energy jump to ,exist Energy level drops to The system identified It is a local maximum point, and its energy is determined. Significantly higher than the preset effective signal noise floor threshold This confirms that the peak value represents the dominant frequency characteristic of the periodic detachment of bubbles from the nozzle, and the corresponding horizontal axis value is then determined. The fundamental frequency index of bubble oscillation is obtained.

[0029] Specifically, such as Figure 2 , 5 As shown, the traffic mapping inversion module includes: The mapping table node positioning submodule calls the bubble oscillation fundamental frequency index, accesses the pre-built fluid dynamics mapping table which includes the nonlinear relationship between frequency and flow rate, performs a numerical comparison operation in the frequency index column of the fluid dynamics mapping table, determines the numerical difference between the bubble oscillation fundamental frequency index and the discrete frequency points in the fluid dynamics mapping table, determines the numerical range of the bubble oscillation fundamental frequency index on the frequency axis, locks the standard frequency node row number position corresponding to the upper and lower boundaries of the numerical range, and establishes the adjacent frequency node index. Load the fluid dynamics characteristic database stored in the control system ROM. This database contains discretized standard data as shown in Table 2. If the current input bubble oscillation fundamental frequency index is... The system iterates through the "Oscillation Frequency" column in Table 2 and performs a numerical comparison operation: [The system then determines...] Determine if the real-time frequency falls within the index row number ( ) and index row number ( Within the numerical range formed by these two numerical ranges, lock the standard frequency node row number positions corresponding to the upper and lower boundaries of these two numerical ranges, and establish a neighboring frequency node index; Table 2: Fluid dynamics frequency-flow mapping table; As shown in Table 2, the system records the nonlinear relationship between bubble oscillation frequency and gas volume flow rate and local gradient characteristics under standard operating conditions.

[0030] The flow calibration data extraction submodule reads the standard frequency node values ​​and the standard gas volume flow rate level mapped by the node from the fluid dynamics mapping table based on the neighboring frequency node index. It then combines the standard frequency node boundary with the standard gas volume flow rate level to obtain the baseline flow calibration data. Read the standard frequency node values ​​and standard gas volumetric flow rate levels that serve as interval boundaries from the fluid dynamics mapping table (Table 2), and read the data for the lower boundary node (index row number 3): , Read the data from the upper boundary node (index row number 4): , By combining the above standard frequency node boundaries with standard gas volume flow rates, reference flow calibration data can be obtained.

[0031] The flow interpolation calculation submodule calculates the magnitude of the deviation of the fundamental frequency index of bubble oscillation relative to the low-frequency boundary point based on the reference flow calibration data, converts the magnitude of the deviation into the corresponding flow increment step, and accumulates the flow increment step to the flow reference value corresponding to the low-frequency boundary point to obtain real-time oxygen volume flow rate data. Calculate the magnitude of the deviation of the fundamental frequency exponent of bubble oscillation relative to the low-frequency boundary point. Interpolation formula is used Perform the calculations and substitute in the specific values: Increase the step size of the traffic increment Accumulated to the flow baseline value corresponding to the low-frequency boundary point The above provides real-time oxygen volumetric flow rate data. .

[0032] Specifically, such as Figure 2 , 6 As shown, the deviation calculation module includes: The flow deviation quantification submodule reads the preset target flow standard through real-time oxygen volume flow rate data, performs numerical subtraction to calculate the difference between the real-time oxygen volume flow rate data and the target flow standard, extracts the absolute value of the difference as the deviation magnitude, identifies the positive or negative sign of the difference to determine the state attribute of the flow rate being in positive or negative deviation, and generates an instantaneous flow deviation vector. Read the preset target traffic standard Perform numerical subtraction to calculate the difference between the real-time oxygen volumetric flow rate data and the target flow rate standard. Substitute the numerical values ​​into the calculation: Extract the absolute value of the difference As a measure of deviation, the sign of the difference is identified as negative, indicating that the flow is in a "loss" state. The signed value is then... Generate the instantaneous flow deviation vector.

[0033] The feedback control submodule calculates the real-time amplitude term, time integral term, and differential rate of change term of the instantaneous flow deviation vector based on the instantaneous flow deviation vector. It calls the preset proportional coefficient, integral time constant, and differential time constant to perform linear weighted summation calculation on the above three deviation characteristics to obtain the theoretical control output value required to offset the flow error, determine the logical control quantity of the regulating valve core displacement, and obtain the valve stroke adjustment quantity. An incremental PID algorithm is used, with preset parameters: proportional coefficient. Integral coefficient Differential coefficients Reading deviation: , Current deviation Calculate the PID output increment : Substitute the values: ; ; ; because A negative value indicates insufficient flow, requiring an increase in valve opening. Define the control output. (Scaling factor), calculated If the logic control quantity is determined to be positive, it indicates that a positive opening adjustment needs to be performed, and the valve stroke adjustment amount is obtained.

[0034] The correction instruction generation submodule, for the valve stroke adjustment amount, maps the logic control amount into the number of step pulses based on the stepper motor drive characteristics and mechanical transmission ratio parameters of the electronic flow control valve, sets the motor rotation direction indicator bit according to the sign bit of the adjustment amount, and encapsulates the pulse number information and direction indicator bit into a control instruction data packet according to the predetermined underlying digital communication protocol format, and establishes the valve opening correction digital signal. Set each Each logical unit corresponds to the axial displacement of the valve core. And the stepper motor rotates once ( (Promote the movement of the valve core) The motor step angle is First, calculate the physical displacement requirement: Secondly, the displacement is mapped to the number of step pulses: Each pulse, based on the sign of the adjustment (positive), sets the motor rotation direction indicator to 0x01 (positive on), and according to the predetermined communication protocol (format: frame header-direction-pulse high bit-pulse low bit), transmits... Converted to hexadecimal 0x0103, it is combined and encapsulated into a control command data packet 0xA50x010x010x03 to establish a digital signal for valve opening correction.

[0035] Specifically, such as Figure 2 , 7 As shown, the valve control module includes: The drive signal execution submodule calls the valve opening correction digital signal, parses the pulse count and direction identifier, controls the motor drive bridge to conduct in sequence to generate electromagnetic torque, drives the stepper motor rotor of the electronic flow control valve to rotate, and converts the rotational torque into the linear feed action of the valve core axially through the mechanical transmission component, changes the physical position of the valve core relative to the valve seat, and obtains the valve core physical displacement increment. Analysis includes pulse counting With the direction identifier bit 0x01, the control motor drive bridge circuit is sequentially turned on to generate electromagnetic torque, so as to... The pulse frequency drives the stepper motor rotor of the electronic flow control valve to rotate, and the rotational torque is converted into a linear feed motion of the valve core axially through a mechanical transmission component, driving the valve core from its current position (e.g., To retreat, actually producing The physical displacement causes the valve core to reach its final position. By changing the physical position of the valve core relative to the valve seat, the physical displacement increment of the valve core can be obtained.

[0036] The fluid resistance control submodule modifies the flow area of ​​the valve's internal throttling orifice based on the valve core's physical displacement increment, changes the local flow resistance coefficient when oxygen flows through the pipeline, adjusts the volumetric flow rate of the gas entering the humidification bottle under the action of the gas source pressure, performs mechanical intervention and regulation of the physical gas path flow rate, and establishes the inlet flow rate after regulation. The flow area of ​​the internal throttling orifice of the valve is adjusted as the valve core displacement increases to... The throttling annular area is changed from before adjustment Expand to Maintaining gas source pressure Under these conditions, increasing the flow area reduces the local flow resistance coefficient. After adjustment, the volumetric flow rate of gas through the throttling orifice increases to [value missing]. (approximately) At this point, the average flow velocity of the gas through the throttling orifice is approximately Mechanical intervention is performed to regulate the flow rate of the physical gas path and establish the inlet velocity after regulation.

[0037] The closed-loop performance verification submodule performs numerical matching analysis between the adjusted inlet flow rate and the target flow rate standard, calculates the instantaneous residual value after the adjustment action is completed, and compares the instantaneous residual value with the preset steady-state convergence judgment threshold. If the residual value is less than the steady-state convergence judgment threshold, it is judged as a locked state; otherwise, it is marked as a dynamic following state, generating a humidification oxygen supply closed-loop control state. Adjust the flow rate corresponding to the inlet velocity. With target flow standard Perform numerical matching analysis to calculate the instantaneous residual value after the adjustment action is completed. The instantaneous residual value is compared with the preset steady-state convergence threshold. Perform comparison and judgment: Due to The system determines that the current control deviation has fallen into the allowable dead zone range, identifies it as a locked state, and freezes the accumulation of integral terms to avoid overshoot; if the residual value is greater than the threshold, it is marked as a dynamic following state and continues the next round of adjustment, generating a humidified oxygen supply closed-loop control state.

[0038] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A cloud-based oxygen humidification status management system, characterized in that, The system includes: The interface visual monitoring module collects real-time video stream data output by the visual sensor, delineates a rectangular monitoring area covering the area above the nozzle of the duct in the cloud server, calculates the arithmetic mean of the grayscale data of the pixels in the rectangular monitoring area, and generates a grayscale time-domain waveform signal of the gas-liquid interface. The spectral feature extraction module performs a discrete Fourier transform on the gray-scale time-domain waveform signal of the gas-liquid interface, converting the gray-scale information fluctuating in the time dimension into energy spectrum data distributed in the frequency dimension, and obtaining the fundamental frequency index of the bubble oscillation. The flow mapping inversion module uses the fundamental frequency index of the bubble oscillation as an index to obtain the fluid dynamics mapping table pre-stored in the cloud server, including the nonlinear correspondence between frequency data and standard gas volume flow rate levels, retrieves the oxygen supply at the current moment, and obtains real-time oxygen volume flow rate data. The deviation calculation module compares the real-time oxygen volumetric flow rate data with the preset target flow rate standard, calculates the deviation amplitude and polarity between the real-time oxygen volumetric flow rate data and the target flow rate standard, generates digital instructions for adjusting the valve opening, and establishes a valve opening correction digital signal.

2. The cloud-based oxygen humidification status management system according to claim 1, characterized in that: The gas-liquid interface grayscale time-domain waveform signal includes discrete time node coordinates, instantaneous average grayscale amplitude, and waveform sequence length information. The bubble oscillation fundamental frequency index includes the main peak frequency value, power spectral density peak intensity, and frequency resolution parameter. The real-time oxygen volumetric flow rate data includes the instantaneous flow velocity value, gas volume measurement unit, and flow rate confidence interval. The valve opening correction digital signal includes the number of step drive pulses, adjustment direction identifier, and correction amplitude encoding.

3. The cloud-based oxygen humidification status management system according to claim 1, characterized in that, The interface visual monitoring module includes: The grayscale extraction submodule of the monitoring area collects real-time video stream data output by the visual sensor. In the cloud server, a rectangular monitoring area covering the area above the nozzle of the duct is defined for the real-time video stream data. The two-dimensional coordinate positions of the pixels inside the rectangular monitoring area are traversed, and the grayscale data values ​​corresponding to the pixels are extracted one by one. The arithmetic average calculation operation is performed on all grayscale data within the range of a single frame image to obtain the instantaneous average grayscale value of the monitoring area. The transmittance sequence construction submodule reads the frame generation timestamps generated along with the instantaneous average gray value of the monitoring area based on the instantaneous average gray value of the monitoring area, and stores multiple instantaneous average gray values ​​of the monitoring area into a linear buffer queue in sequence according to the order of the frame generation timestamps to establish a gas-liquid two-phase interface transmittance sequence. The grayscale time-domain waveform generation submodule performs signal waveform processing, sets the frame generation timestamp as the time-domain horizontal axis coordinate, sets the numerical amplitude in the gas-liquid two-phase interface transmittance sequence as the vertical axis intensity coordinate, maps discrete sequence data points to a two-dimensional coordinate system, connects adjacent data points through interpolation to form a continuously changing fluctuation curve, calculates the weighted smooth value of the interpolated grayscale trajectory, performs signal formatting and encapsulation on the fluctuation curve, and generates a gas-liquid interface grayscale time-domain waveform signal.

4. The cloud-based oxygen humidification status management system according to claim 3, characterized in that, The spectral feature extraction module includes: The frequency domain transformation submodule calls the gray-scale time-domain waveform signal of the gas-liquid interface, performs time-frequency domain transformation operation, and resolves the gray-scale amplitude fluctuation information originally distributed along the time axis into spectral components distributed along the frequency axis, calculates the square value of the modulus corresponding to the frequency component, and generates the gray-scale fluctuation frequency domain energy spectrum. The frequency band locking submodule sets the bandpass filter boundary based on the gray-scale fluctuation frequency domain energy spectrum and the preset frequency range of the bubble generation physical process. It delineates the target range of bubble oscillation energy on the frequency axis of the energy spectrum data, separates the local spectrum data segments corresponding to the bubble generation rhythm, removes the frequency band interference of background noise, and establishes the bubble generation feature screening frequency band. The fundamental frequency analysis submodule filters frequency bands based on the bubble generation characteristics, performs extreme value search operations on discrete spectral data, compares the power spectral density values ​​corresponding to the frequency points within the frequency band one by one, locates the spectral peak position of the peak energy density, reads the horizontal axis frequency value corresponding to the spectral peak position as the core parameter for quantifying the periodic generation rate of bubbles, and obtains the fundamental frequency index of bubble oscillation.

5. The cloud-based oxygen humidification status management system according to claim 4, characterized in that, The specific definition of the time-frequency domain conversion operation is as follows: the gas-liquid interface grayscale time-domain waveform signal is subjected to discrete frequency domain transformation within a preset sampling window length, and only the spectrum data on the non-negative frequency axis is retained in the transformation result; The calculation of the square value of the modulus corresponding to the frequency component is specifically limited to squaring the complex spectral amplitude of each discrete frequency point and forming a one-to-one energy value sequence. The preset frequency range setting bandpass filter boundary for the combined bubble generation physical process is specifically limited to setting the lower frequency limit and the upper frequency limit to preset lower frequency limit values ​​and upper frequency limit values, respectively, and only extracting the grayscale fluctuation frequency domain energy spectrum data between the lower frequency limit and the upper frequency limit.

6. The cloud-based oxygen humidification status management system according to claim 4, characterized in that, The traffic mapping inversion module includes: The mapping table node positioning submodule calls the bubble oscillation fundamental frequency index, accesses the pre-constructed fluid dynamics mapping table which includes the nonlinear relationship between frequency and flow rate, performs a numerical comparison operation in the frequency index column of the fluid dynamics mapping table, determines the numerical difference between the bubble oscillation fundamental frequency index and the discrete frequency points in the fluid dynamics mapping table, determines the numerical range of the bubble oscillation fundamental frequency index on the frequency axis, locks the standard frequency node row number position corresponding to the upper and lower boundaries of the numerical range, and establishes the adjacent frequency node index. The flow calibration data extraction submodule reads the standard frequency node values ​​and the standard gas volume flow rate level mapped by the node from the fluid dynamics mapping table based on the adjacent frequency node index, and combines the standard frequency node boundary with the standard gas volume flow rate level to obtain the reference flow calibration data. The flow interpolation calculation submodule calculates the magnitude of the deviation of the bubble oscillation fundamental frequency index from the low-frequency boundary point based on the reference flow calibration data, converts the magnitude of the deviation into the corresponding flow increment step, and accumulates the flow increment step to the flow reference value corresponding to the low-frequency boundary point to obtain real-time oxygen volumetric flow rate data.

7. The cloud-based oxygen humidification status management system according to claim 6, characterized in that, The deviation calculation module includes: The flow deviation quantification submodule reads the preset target flow standard through the real-time oxygen volume flow rate data, performs numerical subtraction, calculates the difference between the real-time oxygen volume flow rate data and the target flow standard, extracts the absolute value of the difference as the deviation amplitude, identifies the positive or negative sign of the difference to determine the state attribute of the flow rate being in positive or negative deviation, and generates an instantaneous flow deviation vector. Based on the instantaneous flow deviation vector, the feedback control submodule calculates the real-time amplitude term, the time integral term of the deviation, and the differential rate of change term of the deviation over time of the instantaneous flow deviation vector. It then calls the preset proportional coefficient, integral time constant, and differential time constant to perform a linear weighted summation calculation on the above three deviation characteristics to obtain the theoretical control output value required to offset the flow error, determine the logical control quantity of the regulating valve core displacement, and obtain the valve stroke adjustment quantity. The correction instruction generation submodule, for the valve stroke adjustment amount, maps the logic control amount into the number of step pulses based on the stepper motor drive characteristics and mechanical transmission ratio parameters of the electronic flow control valve, sets the motor rotation direction identifier bit according to the sign bit of the adjustment amount, and encapsulates the pulse number information and direction identifier bit into a control instruction data packet according to the predetermined underlying digital communication protocol format, and establishes a valve opening correction digital signal.

8. The cloud-based oxygen humidification status management system according to claim 7, characterized in that, The calculation of the difference between the real-time oxygen volumetric flow rate data and the target flow rate standard is specifically limited to: comparing the real-time oxygen volumetric flow rate data with the target flow rate standard one by one, extracting the difference range by absolute value calculation, and determining whether the flow rate is in a positive or negative deviation state by comparing the calculated difference with a preset threshold. The generation of the instantaneous flow deviation vector is specifically defined as follows: based on the absolute value of the difference, the positive and negative signs of the difference are identified, and corresponding flow deviation vectors are generated respectively; the flow adjustment requirements are determined according to the change in the magnitude of the difference.

9. The cloud-based oxygen humidification status management system according to claim 1, characterized in that, The system also includes a valve actuation and control module: The valve control module sends the valve opening correction digital signal to the front-end device, drives the valve core of the electronic flow control valve to produce mechanical displacement, adjusts the oxygen flow rate input to the humidification bottle, verifies the consistency between the real-time flow rate and the target flow rate standard, and generates a closed-loop control state for humidification oxygen supply. The closed-loop control status of the humidification oxygen supply includes the steady-state convergence indicator of the flow rate, the dynamic following error rate, and the operation mode feedback information.

10. The cloud-based oxygen humidification status management system according to claim 9, characterized in that: The valve control module includes: The drive signal execution submodule calls the valve opening correction digital signal, parses the pulse count and direction identifier, controls the motor drive bridge to conduct in sequence to generate electromagnetic torque, drives the stepper motor rotor of the electronic flow control valve to rotate, and converts the rotational torque into the linear feed action of the valve core axially through the mechanical transmission component, changes the physical position of the valve core relative to the valve seat, and obtains the valve core physical displacement increment. The fluid resistance control submodule, based on the physical displacement increment of the valve core, corrects the flow area of ​​the throttling orifice inside the valve, changes the local flow resistance coefficient when oxygen flows through the pipeline, adjusts the volumetric flow rate of the gas entering the humidification bottle under the action of the gas source pressure, performs mechanical intervention and regulation of the physical gas path flow rate, and establishes the inlet flow rate after regulation. The closed-loop performance verification submodule performs numerical matching analysis between the adjusted inlet flow rate and the target flow rate standard, calculates the instantaneous residual value after the adjustment action is completed, and compares the instantaneous residual value with the preset steady-state convergence judgment threshold. If the residual value is less than the steady-state convergence judgment threshold, it is judged as a locked state; otherwise, it is marked as a dynamic following state, generating a humidification oxygen supply closed-loop control state.

Images