High-pressure environment large-space multi-component gas partial pressure equalization control method

Abstract

The application discloses a high-pressure environment large-space multi-component gas partial pressure equalization control method, comprising: obtaining real-time monitoring data of oxygen partial pressure and gas flow rate of each sub-region in a high-pressure cabin, and performing time alignment and coordinate unification to obtain a distributed monitoring data set; performing sub-region gas concentration field reconstruction and gas flow state feature extraction to obtain gas concentration distribution features and gas flow field features; constructing a multi-component gas convection-diffusion dynamic model according to the features; constructing a model predictive controller based on the dynamic model and rolling solving an optimal control sequence; estimating disturbances in real time based on an adaptive disturbance observer, and feeding forward compensating the optimal control instruction according to the disturbance estimate value; and performing air outlet adjustment, oxygen supplement control and fan speed regulation and dynamically adjusting oxygen supplement flow distribution to obtain an equalized multi-component gas partial pressure field. The application realizes accurate equalization control of multi-component gas partial pressure in a large space of a high-pressure cabin, and significantly improves the safety and comfort of the gas environment in the cabin.

Description

Technical Field

[0001] This application relates to the field of high-pressure environment gas control technology, and in particular to a method for equalizing the partial pressure of multi-element gases in a large space under high pressure. Background Technology

[0002] With the rapid development of applications such as deep-sea diving, hyperbaric medicine, and high-pressure simulation training, the hyperbaric chamber, as a core piece of equipment providing a high-pressure environment, directly impacts the safety of personnel and operational efficiency through precise control of its internal gas environment. Under dynamic high-pressure environments of 0.6 MPa and above, the oxygen partial pressure distribution within the chamber is highly sensitive to the flow field. The diffusion coefficient, density, and compressibility of gas molecules all change significantly. Traditional strategies based on the assumption of normal pressure, such as fixed airflow and fixed oxygen supply, are ill-suited to changing operating conditions. This can easily lead to an imbalance in the gas partial pressure field, resulting in either excessively high local oxygen partial pressure causing oxygen poisoning or excessively low local oxygen partial pressure causing hypoxia and asphyxiation.

[0003] Existing hyperbaric chamber gas control technologies mostly employ PID control based on single sensor feedback or simple on / off oxygen supplementation strategies. These control loops fail to consider the nonlinear dynamic characteristics of gas convection-diffusion coupling under high pressure, and lack the ability to perceive the spatial distribution of the gas concentration field within the chamber. While some studies have introduced multi-point monitoring, each monitoring point is controlled independently, failing to establish a distributed closed-loop control architecture and thus unable to achieve global equilibrium optimization of the gas partial pressure field within the chamber. Summary of the Invention

[0004] The main purpose of this application is to provide a method for equalizing the partial pressure of multiple gases in a large space under high pressure, which aims to solve the technical problems of the oxygen partial pressure distribution in the cabin being highly sensitive to the flow field under high pressure dynamic environment, the traditional fixed air volume and fixed oxygen replenishment strategy being difficult to adapt to changes in working conditions, resulting in an unbalanced gas partial pressure field and excessively high local oxygen partial pressure causing high oxygen poisoning in personnel.

[0005] To achieve the above objectives, this application proposes a method for controlling the partial pressure equilibrium of multiple gases in a large space under high pressure. The method includes: Real-time monitoring data of oxygen partial pressure and gas flow velocity in each sub-region of the hyperbaric chamber are acquired, and the real-time monitoring data of oxygen partial pressure and gas flow velocity are time-aligned and coordinate-unified to obtain a distributed monitoring dataset. Based on the distributed monitoring dataset, sub-region gas concentration field reconstruction and gas flow dynamic feature extraction are performed to obtain the gas concentration distribution characteristics and gas flow field characteristics of each sub-region. Based on the gas concentration distribution characteristics and gas flow field characteristics, a dynamic model of multi-element gas convection-diffusion in a high-pressure environment is constructed. A model predictive controller is constructed based on the multi-element gas convection-diffusion dynamic model, and the optimal control sequence is solved in a rolling manner within each control cycle to obtain the optimal control command for the current control cycle. Based on the real-time estimation of the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field using an adaptive disturbance observer, disturbance estimates are obtained. Then, feedforward compensation is performed on the optimal control command based on the disturbance estimates to obtain the compensated optimal control command. The air outlet adjustment, oxygen supplementation control, and fan speed regulation of the environmental control system are executed according to the compensated optimal control command. An oxygen supplementation rate regulation algorithm based on concentration gradient feedback is adopted to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation in each region, so as to obtain a balanced multi-element gas partial pressure field.

[0006] Furthermore, to achieve the above objectives, this application also proposes a high-pressure environment large-space multi-gas partial pressure equalization control device, which includes: The acquisition module is used to acquire real-time monitoring data of oxygen partial pressure and gas flow rate of each sub-region in the hyperbaric chamber, and to align the real-time monitoring data of oxygen partial pressure and gas flow rate with time and coordinate to obtain a distributed monitoring dataset. The extraction module is used to reconstruct the gas concentration field of sub-regions and extract the dynamic features of gas flow based on the distributed monitoring dataset, so as to obtain the gas concentration distribution features and gas flow field features of each sub-region; The construction module is used to construct a dynamic model of convection-diffusion of multiple gases in a high-pressure environment based on the gas concentration distribution characteristics and gas flow field characteristics. The construction module is also used to construct a model predictive controller based on the multi-element gas convection-diffusion dynamic model and to solve the optimal control sequence in each control cycle to obtain the optimal control command for the current control cycle. The compensation module is used to estimate the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field in real time based on the adaptive disturbance observer, obtain disturbance estimates, and compensate the optimal control command based on the disturbance estimates to obtain the compensated optimal control command. The adjustment module is used to execute the air outlet adjustment, oxygen supplementation control and fan speed regulation of the environmental control system according to the compensated optimal control command, and adopts the oxygen supplementation rate adjustment algorithm based on concentration gradient feedback to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation of each region to obtain the balanced multi-element gas partial pressure field.

[0007] In addition, to achieve the above objectives, this application also proposes a high-pressure environment large-space multi-gas partial pressure equalization control device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the high-pressure environment large-space multi-gas partial pressure equalization control method described above.

[0008] In addition, to achieve the above objectives, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the high-pressure environment large-space multi-element gas partial pressure equalization control method as described above.

[0009] The proposed technical solutions (one or more) achieve full-space coverage monitoring of the gas state inside the cabin through a distributed sensor network. By reconstructing the concentration field and flow field based on radial basis function interpolation and intrinsic orthogonal decomposition, a precise mapping from discrete monitoring data to continuous distribution characteristics is achieved. By using a multi-element gas convection-diffusion dynamic model that considers the effects of non-ideal gases under high pressure, the accuracy of gas partial pressure evolution prediction under high pressure is significantly improved. Through rolling optimization and multi-dimensional constraint processing of the model predictive controller, the foresight and safety of the control strategy are achieved. Through feedforward compensation of the adaptive disturbance observer, the influence of time-varying disturbances such as personnel activities and equipment heat dissipation is effectively suppressed. Through dynamic adjustment of the oxygen replenishment rate based on concentration gradient feedback, the oxygen partial pressure deviation of each sub-region is finely corrected. Finally, a complete closed-loop control system of "monitoring-reconstruction-modeling-prediction-compensation-execution" is formed, realizing precise and balanced control of multi-element gas partial pressure in a large space under high pressure, and significantly improving the safety, comfort, and control accuracy of the gas environment inside the cabin. Attached Figure Description

[0010] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating an embodiment of the high-pressure environment large-space multi-element gas partial pressure equilibrium control method of this application. Figure 2 This is a schematic diagram of the module structure of the high-pressure environment large-space multi-element gas partial pressure equalization control device according to an embodiment of this application.

[0013] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0015] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0016] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a high-pressure environment large-space multi-gas partial pressure equalization control device. The following description uses a high-pressure environment large-space multi-gas partial pressure equalization control device as an example to illustrate this embodiment and the subsequent embodiments.

[0017] Based on this, embodiments of this application provide a method for equalizing the partial pressure of multiple gases in a large space under high pressure, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the high-pressure environment large-space multi-element gas partial pressure equalization control method of this application.

[0018] In this embodiment, the method for controlling the partial pressure equilibrium of multi-element gases in a large space under high pressure includes steps S10 to S60: Step S10: Obtain real-time monitoring data of oxygen partial pressure and gas flow rate in each sub-region of the hyperbaric chamber, and align the real-time monitoring data of oxygen partial pressure and gas flow rate in time and coordinates to obtain a distributed monitoring dataset.

[0019] It should be noted that a high-pressure chamber refers to a sealed chamber environment with an operating pressure of 0.6 MPa or higher, including but not limited to hyperbaric oxygen chambers, saturation diving living quarters, and high-pressure simulation training chambers. Under high pressure, the physical properties of gases, such as density, diffusion coefficient, and viscosity, differ significantly from those under normal pressure. Directly applying traditional normal-pressure gas sensors to high-pressure environments will result in severe measurement errors; therefore, high-pressure environment adaptability correction is necessary. A sub-region refers to several control units that divide the large space within the chamber according to functional zoning or geometric features. Each sub-region has relatively independent gas concentration and flow field characteristics, achieving global equilibrium through distributed control.

[0020] Understandably, real-time oxygen partial pressure monitoring data refers to time-series data reflecting the oxygen partial pressure level of each sub-region, collected by oxygen partial pressure sensors deployed at the center of each sub-region. Its sampling frequency must satisfy the Nyquist sampling theorem and is typically no less than twice the dominant frequency of gas concentration field changes. Real-time gas velocity monitoring data refers to time-series data reflecting the gas flow state of each sub-region, collected by gas velocity sensors deployed at the boundaries of each sub-region, used to characterize the transport effect of gas between sub-regions. Time alignment refers to the process of unifying monitoring data from different sensors, which may have clock deviations, to the same absolute time reference; it is a prerequisite for distributed data fusion. Coordinate unification refers to the process of transforming the local installation coordinates of each sensor to the global coordinate system within the cabin, ensuring the consistency of spatial location information. The distributed monitoring dataset refers to a standardized data structure containing the spatial location, monitoring type, timestamp, and monitoring value of each monitoring point after time alignment and coordinate unification; it serves as the input for subsequent concentration field reconstruction and flow field feature extraction.

[0021] Understandably, at a high pressure of 0.6 MPa, the gas density is approximately six times that at atmospheric pressure, the mean free path of gas molecules is significantly reduced, diffusion is suppressed, and convection dominates gas transport. Simultaneously, the compressibility effect of gases under high pressure is not negligible, the gas equation of state deviates from the ideal gas assumption, and a nonlinear mapping relationship exists between oxygen partial pressure and mole fraction. These high-pressure-specific effects mean that gas monitoring and modeling methods used under atmospheric pressure cannot be directly applied; therefore, dedicated monitoring and processing procedures for high-pressure environments must be established.

[0022] In one feasible implementation, step S10 may include: deploying a distributed sensor network within the hyperbaric chamber, wherein the distributed sensor network includes oxygen partial pressure sensors arranged at the center of each sub-region and gas velocity sensors arranged at the boundaries of each sub-region; acquiring the raw oxygen partial pressure signals of each oxygen partial pressure sensor and the raw flow velocity signals of each gas velocity sensor; performing high-pressure environmental pressure compensation and temperature drift correction on the raw oxygen partial pressure signals to obtain corrected oxygen partial pressure data; performing high-pressure gas density correction and boundary effect correction on the raw flow velocity signals to obtain corrected gas flow velocity data; aligning the corrected oxygen partial pressure data and the corrected gas flow velocity data in time using a timestamp alignment algorithm based on global clock synchronization to obtain time-aligned data; and performing coordinate unification and grid mapping on the time-aligned data according to the spatial coordinate positions of each sub-region to obtain a distributed monitoring dataset.

[0023] It should be noted that a distributed sensor network is a multi-point monitoring system arranged in a specific topology within the cabin space. Its spatial layout must meet the observability requirements of the gas concentration and flow fields within the cabin. The oxygen partial pressure sensor is an electrochemical or optical sensor used to measure the partial pressure of oxygen in a gas mixture. In high-pressure environments, a pressure-resistant sensor must be selected, and its measurement principle must consider the impact of high pressure on the sensor's response characteristics. The gas velocity sensor is used to measure the velocity of gas flow. In this embodiment, a miniature hot-film anemometer is used, which operates based on the principle of heat conduction. The flow velocity is calculated by measuring the heat dissipation rate of the heating element in the airflow. Density correction is required in high-pressure environments.

[0024] Understandably, the division of sub-regions needs to comprehensively consider factors such as the internal geometry, personnel activity areas, equipment layout, and gas flow paths. For regular rectangular compartments, a uniform grid can be used; for irregular compartments or compartments with internal partitions, non-uniform division is required based on the actual spatial structure. The size of each sub-region should ensure that the gas concentration field and flow field within that region have relatively uniform characteristic scales, and the maximum size of the sub-region should not exceed 1 / 3 of the minimum characteristic length of the compartment. An oxygen partial pressure sensor is placed at the center of the sub-region to represent the average oxygen partial pressure level of that region; a gas velocity sensor is placed at the boundary of the sub-region to monitor the intensity of gas exchange between sub-regions.

[0025] It is understandable that the raw oxygen partial pressure signal refers to the electrical or digital signal directly output by the oxygen partial pressure sensor without any processing, and its value is directly related to the physical response of the sensor's sensitive element. The raw flow rate signal refers to the electrical or digital signal directly output by the gas flow rate sensor without any processing, and its value is directly related to the heat dissipation rate of the hot film element.

[0026] High-pressure environmental pressure compensation refers to the correction process that eliminates the influence of the total ambient pressure on the measurement results of the oxygen partial pressure sensor. In electrochemical oxygen partial pressure sensors, the sensor's output current is not only related to the oxygen partial pressure but also affected by the total ambient pressure. Under a high pressure of 0.6 MPa, without pressure compensation, the measurement error can reach over 20%. Temperature drift correction refers to the correction process that eliminates the influence of changes in the sensor's operating temperature on the measurement results. The permeability and diffusion coefficient of the sensitive membrane of the electrochemical sensor both change with temperature, leading to output drift.

[0027] Pressure compensation in high-pressure environments employs a lookup table method or a fitting method based on sensor calibration data. Before the sensor leaves the factory, it is calibrated at standard temperatures at different pressure points, such as 0.1 MPa, 0.3 MPa, 0.6 MPa, 0.8 MPa, and 1.0 MPa. The correspondence between the sensor output and oxygen partial pressure at each pressure point is recorded, and a pressure compensation lookup table or polynomial fitting function is established. During actual measurement, the compensation coefficient is looked up based on the current total environmental pressure to correct the original signal. in, The corrected oxygen partial pressure data. This is the original oxygen partial pressure signal. For total pressure The corresponding compensation coefficient.

[0028] Temperature drift correction employs a real-time correction method based on synchronous temperature sensor measurements. A temperature sensor is placed near the oxygen partial pressure sensor to monitor its operating temperature in real time. Corrections are made based on the temperature drift model: in, This is the final corrected oxygen partial pressure data. This is the temperature drift coefficient. For calibration reference temperature.

[0029] High-pressure gas density correction refers to the correction process that eliminates the influence of changes in ambient gas density on the measurement results of hot-film anemometers. The heat dissipation rate of a hot-film anemometer is related to the gas's density, viscosity, thermal conductivity, and other physical properties. Under high pressure, the gas density increases significantly, resulting in a different heat dissipation rate at the same flow velocity compared to atmospheric pressure, leading to a systematic deviation in the directly output flow velocity value. Boundary effect correction refers to the correction process that eliminates the influence of walls, obstacles, etc., near the sensor installation location on the flow measurement results. Near the chamber boundary, the flow is constrained by the wall, and the velocity distribution differs from that of free flow.

[0030] High-pressure gas density correction is based on the working principle of a hot-film anemometer. The heat dissipation power of the hot-film element... With flow rate and gas density The relationship can be represented as: in, It is a natural convection heat dissipation item. To achieve the forced convection heat dissipation coefficient, The speed index is set to 0.45 to 0.5.

[0031] Under high pressure, Increase, Same Below The increase in pressure causes the flow velocity calculated by the sensor based on the atmospheric pressure calibration curve to be too high. The correction method is as follows: in, The corrected flow rate. The actual gas density under the current high-pressure environment is calculated by the gas state equation based on the total pressure, temperature, and gas composition.

[0032] Boundary effect correction employs the wall function method based on computational fluid dynamics simulation. The velocity distribution near the sensor installation location is obtained through CFD simulation, establishing a mapping relationship between local velocity and free flow velocity. Correction coefficients are then queried based on the sensor's distance from the wall to correct the measured values.

[0033] It is worth noting that global clock synchronization refers to the process of unifying the local clocks of all sensors to the same absolute time reference through methods such as the IEEE 1588 precision time protocol or GPS time synchronization. Timestamp alignment algorithms refer to methods for aligning data sequences from different sensors, which may have different sampling frequencies, to a unified timeline.

[0034] Each sensor adds a local timestamp at the sampling time, and the local timestamp is converted into a global absolute timestamp through global clock synchronization. For sensors with different sampling frequencies, linear interpolation or spline interpolation methods are used to resample the data onto a unified time grid. Let the sampling period of the oxygen partial pressure sensor be... The sampling period of the gas flow rate sensor is The step size of the unified time grid is ,Right now and The greatest common divisor. For moments without original sample values ​​on a uniform time grid, linear interpolation of adjacent sample values ​​is used: in, To unify the target time on the time grid, and For adjacent original sampling times, and This is the corresponding sampled value.

[0035] Coordinate unification refers to the process of transforming the local installation coordinates of each sensor to the global coordinate system within the cabin, typically achieved through a coordinate transformation matrix. Mesh mapping refers to the process of spatially associating sensor monitoring data with discrete cells of the computational grid, providing a data foundation for subsequent concentration field reconstruction.

[0036] Establish a global Cartesian coordinate system within the cabin, with the cabin's geometric center as the origin and the cabin's length, width, and height as the x, y, and z axes, respectively. The installation positions of each sensor are represented in the global coordinate system as ( , , The time-aligned monitoring data is correlated with the spatial location of the sensors to form a four-dimensional dataset of space-time-physical quantities. For the computational grid required for concentration field reconstruction, a mapping relationship between grid nodes and sensor locations is established, typically using nearest neighbor mapping or distance-weighted mapping. in, Let be the spatial distance between the i-th sensor and the j-th grid node. The monitoring data mapping value of grid node j is: in, For the number of sensors, Distance weights To avoid small positive numbers that are divided by zero, we take 10. -6 ~10 -3 , The monitoring data value of the i-th sensor In this embodiment, the impact of the 0.6MPa high-pressure environment on the measurement accuracy of the oxygen partial pressure sensor is eliminated by high-pressure environment pressure compensation and temperature drift correction; the measurement accuracy of the hot-film anemometer is ensured by high-pressure gas density correction and boundary effect correction; and the standardized fusion of distributed sensor network data is achieved by global clock synchronization timestamp alignment and unified coordinate grid mapping, providing high-quality input data for subsequent concentration field reconstruction and flow field feature extraction, significantly improving the reliability and accuracy of gas environment monitoring under high-pressure environment.

[0037] In one feasible implementation, after the step of unifying the time-aligned data and mapping it to a grid based on the spatial coordinates of each sub-region to obtain a distributed monitoring dataset, the method further includes: acquiring the gas concentration change rate of each sub-region based on the distributed sensor network; adjusting the sampling frequency of the distributed sensor network according to the gas concentration change rate to obtain adaptively sampled monitoring data, wherein sub-regions with a gas concentration change rate higher than a preset change rate threshold use a high-frequency sampling mode, and sub-regions with a gas concentration change rate lower than the preset change rate threshold use a low-frequency sampling mode; when an abnormal data or communication interruption is detected in the oxygen partial pressure sensor or gas flow rate sensor in the distributed sensor network, the corresponding sensor is determined to be faulty, and the location of the faulty sensor is marked; based on the monitoring data of adjacent normal sensors, a virtual sensor reconstruction algorithm based on Kriging interpolation is used to generate virtual monitoring data of the faulty sensor according to the spatial distance and monitoring data correlation between the location of the faulty sensor and the adjacent normal sensors; replacing the abnormal data of the faulty sensor with the virtual monitoring data, updating the distributed monitoring dataset to obtain the reconstructed distributed monitoring dataset.

[0038] It should be noted that the rate of change of gas concentration refers to the amount of change in gas concentration per unit time, reflecting the dynamic evolution trend of the gas concentration field. Under high pressure, due to the decrease in gas diffusion coefficient, the spatial differences in the rate of change of concentration are more significant. Regions with high rates of change typically correspond to strong disturbance sources or the areas affected by control actions. The rate of change of gas concentration is obtained by numerical differentiation calculation of oxygen partial pressure monitoring data.

[0039] Adaptive sampling frequency adjustment is a method that automatically adjusts the sampling frequency based on the dynamic characteristics of the measured signal. Its purpose is to reduce data volume and computational burden while maintaining monitoring accuracy. In hyperbaric chamber gas control scenarios, the rate of change of gas concentration varies significantly across different sub-regions. Using a uniform fixed sampling frequency can lead to data redundancy in some areas and insufficient data in others. Therefore, a rate of change threshold needs to be set. ,when When the i-th sub-region adopts a high-frequency sampling mode, the sampling period is adjusted to... ;when When the i-th sub-region adopts a low-frequency sampling mode, the sampling period is adjusted to... Through adaptive sampling, more intensive monitoring data is obtained in high-dynamic areas, while unnecessary data collection is reduced in low-dynamic areas, thus achieving efficient utilization of monitoring resources.

[0040] Understandably, in the confined environment of a hyperbaric chamber, sensors may malfunction due to high-pressure impacts, temperature changes, electromagnetic interference, etc. If not detected and addressed promptly, control decisions will be based on erroneous data, endangering personnel safety. A dual fault detection mechanism based on statistical thresholds and communication timeouts is employed. Statistical threshold detection: The mean and standard deviation of recent data from each sensor are calculated. When the current measurement value deviates from the mean by more than three times the standard deviation, it is considered an abnormal data point. Communication timeout detection: If a sensor fails to return data within a preset time, communication is considered interrupted. Once a fault is detected, the faulty sensor's number, location, and fault type are recorded, triggering the fault handling process.

[0041] It's worth noting that Kriging interpolation is a statistical interpolation method based on variogram theory. It considers not only the spatial distance between the interpolation point and known points but also the spatial correlation between known points, generating optimal unbiased estimates. In sensor failure scenarios, Kriging interpolation utilizes data from adjacent normal sensors to generate virtual monitoring data at the location of the failed sensor, maintaining the continuous operation of the control system. Data replacement refers to using virtual monitoring data to overwrite the abnormal or missing data from the failed sensor, ensuring the integrity of the distributed monitoring dataset. The reconstructed distributed monitoring dataset refers to the updated dataset after the failed sensor data replacement; its data integrity and spatial coverage remain consistent with those before the failure, ensuring that subsequent concentration field reconstruction and controller design are unaffected by sensor failure.

[0042] The adaptive sampling frequency adjustment enables the optimal allocation of monitoring resources in dynamically changing areas, and the Kriging interpolation virtual sensor reconstruction enables the accurate recovery of faulty sensor data, significantly improving the robustness and reliability of the distributed monitoring network and providing data support for the continuous and stable operation of gas partial pressure equalization control under high pressure environment.

[0043] Step S20: Based on the distributed monitoring dataset, reconstruct the gas concentration field of the sub-region and extract the dynamic features of gas flow to obtain the gas concentration distribution features and gas flow field features of each sub-region.

[0044] It should be noted that gas concentration field reconstruction refers to the process of generating a continuous spatial concentration distribution field based on discrete monitoring point data using mathematical interpolation methods. Its accuracy directly affects the accuracy of subsequent control decisions. Gas flow dynamic feature extraction refers to the process of identifying the dominant structure, time-varying characteristics, and turbulence intensity of the flow field from velocity monitoring data, used to characterize the spatiotemporal evolution of gas flow. Gas concentration distribution characteristics refer to the set of parameters describing the spatial distribution characteristics of oxygen partial pressure in each sub-region, including the oxygen partial pressure gradient vector, oxygen partial pressure curvature tensor, etc., reflecting the spatial non-uniformity of the concentration field. Gas flow field characteristics refer to the set of parameters describing the gas flow state in each sub-region, including the dominant flow mode, time-varying coefficient, Reynolds number, turbulence intensity index, etc., reflecting the dynamic characteristics of the flow field.

[0045] Understandably, due to the reduced gas diffusion coefficient under high pressure, the spatial gradient of the concentration field is steeper, and the concentration changes between monitoring points are more drastic, making it difficult for traditional linear interpolation methods to accurately reconstruct the concentration field. Simultaneously, the Reynolds number of gas flow increases under high pressure, enhancing turbulence and resulting in a multi-scale, unsteady, and complex flow field, necessitating order reduction reconstruction to extract the dominant dynamic modes. This embodiment employs radial basis function interpolation for concentration field reconstruction and intrinsic orthogonal decomposition for order reduction reconstruction of the flow field, effectively addressing the complexity of the gas field under high pressure.

[0046] In one feasible implementation, step S20 may include: generating a continuous gas concentration distribution field for each sub-region based on discrete monitoring point data in the distributed monitoring dataset using a gas concentration field reconstruction algorithm based on radial basis function interpolation, to obtain an initial concentration field; correcting the initial concentration field using the high-pressure environment gas state equation to obtain a corrected concentration field; determining the oxygen partial pressure gradient vector and oxygen partial pressure curvature tensor for each sub-region based on the corrected concentration field, to obtain gas concentration distribution characteristics; extracting the dominant modes and time-varying coefficients of the gas flow field based on gas velocity data in the distributed monitoring dataset using a gas flow field order reduction reconstruction method based on intrinsic orthogonal decomposition, to obtain flow field modal characteristics; and determining the gas flow Reynolds number and turbulence intensity index for each sub-region based on the flow field modal characteristics, to obtain gas flow field characteristics.

[0047] It should be noted that radial basis function interpolation is a scattered data interpolation method based on radially symmetric kernel functions. It reconstructs the concentration field from discrete monitoring points to continuous space by constructing a linear combination of basis functions with the Euclidean distance between the interpolation point and known data points as the independent variable. Under high pressure, the spatial variation of the concentration field is more drastic due to the reduced gas diffusion coefficient. Traditional linear or polynomial interpolation struggles to accurately capture the concentration gradient, while radial basis function interpolation, through the superposition of locally supporting kernel functions, can better adapt to the nonlinear spatial distribution of the concentration field under high pressure.

[0048] Let the internal space domain be The set of discrete monitoring points is The corresponding oxygen partial pressure monitoring value is The goal of radial basis function interpolation is to construct... This ensures accurate interpolation at all monitoring points: in, For radial basis functions, The coefficients are undetermined, determined by the interpolation conditions. Sure, Let be the total number of sensors. Under high-pressure conditions, a Wendland C2 type radial basis function with compact support characteristics is selected: in, For the radius of the compact support, This indicates the positive part operation, i.e., when... The original value is used if the condition is met, otherwise 0 is used. The compact support property makes the interpolation matrix have a sparse structure, which greatly reduces the computational complexity of large-scale matrix inversion and is suitable for real-time reconstruction scenarios of multi-point monitoring in large cabin spaces.

[0049] Correction of the gas state equation under high pressure refers to the process of correcting the concentration field calculated based on the ideal gas assumption due to the non-ideal behavior of gases under high pressure. At pressures of 0.6 MPa and above, the intermolecular forces and molecular volumes are not negligible, causing significant deviations in the ideal gas state equation PV=nRT. Therefore, a non-ideal gas state equation must be used for correction.

[0050] The Peng-Robinson equation of state is used to describe the non-ideal behavior of multi-component gases under high pressure: in, The total gas pressure For temperature, For molar volume, Here, is the universal gas constant, and a and b are component-dependent equation of state parameters. This is a temperature correction factor. For multi-component gas mixtures, mixing parameters are calculated using mixing rules. in, The overall energy parameter of the mixture. This refers to the overall volume parameter of the mixture. and Let be the mole fractions of the i-th and j-th gases, respectively. Let be the interaction parameter between component i and component j. , For binary interaction coefficients, Let i be the energy parameters of component i and component j. Let i be the volume parameter of component i.

[0051] Based on the Peng-Robinson equation of state, a nonlinear mapping relationship between oxygen partial pressure and mole fraction is established. The oxygen partial pressure value in the initial concentration field based on the ideal gas assumption is corrected to the real oxygen partial pressure value under high pressure through the nonlinear mapping relationship, thus obtaining the corrected concentration field.

[0052] The oxygen partial pressure gradient vector describes the spatial rate of change and direction of oxygen partial pressure, pointing in the direction of the fastest increase in oxygen partial pressure, with its magnitude being the rate of change in that direction. The oxygen partial pressure curvature tensor describes the second-order characteristics of the spatial change of oxygen partial pressure; its eigenvalues ​​reflect the degree of curvature of the concentration field in different directions. Under high pressure, due to the weakened diffusion effect, the concentration gradient is steeper, and the curvature change is more dramatic. The gradient vector and curvature tensor can accurately characterize the spatial non-uniformity of the concentration field. The gradient vector and curvature tensor are calculated at the center of each sub-region, serving as the gas concentration distribution characteristics of that sub-region.

[0053] Intrinsic orthogonal decomposition (IOD) is a data-driven flow field order reduction method. It extracts the dominant modes with the highest energy proportions by performing eigenvalue decomposition on the covariance matrix of flow field snapshot data, projecting the high-dimensional flow field data into a low-dimensional modal space, thus achieving efficient characterization of the flow field. In the large-space gas flow of a high-pressure chamber, the flow field exhibits complex characteristics of multi-scale and unsteady operation. Directly modeling and controlling the full-dimensional flow field involves enormous computational costs. Extracting dominant modes through IOD can significantly reduce computational complexity while preserving the main dynamic characteristics of the flow field.

[0054] Let the snapshot matrix composed of gas flow rate monitoring data be... ,in The number of flow rate sensors. The time step is used. Singular value decomposition is performed on the snapshot matrix, and the spatial modes corresponding to the top K largest singular values ​​are selected as the dominant modes. The cumulative energy percentage exceeds a preset threshold, such as 95%. The flow field at any given time can be represented as a linear combination of the dominant modes: in, These are time-varying coefficients, calculated by projection. Dominant mode. and time-varying coefficients Together they constitute the modal characteristics of the flow field.

[0055] The Reynolds number for gas flow is a dimensionless number characterizing the relative magnitudes of inertial and viscous forces in a flow. Its magnitude determines the flow state, such as laminar, transitional, or turbulent flow. Under high pressure, gas density increases while viscosity changes relatively little, resulting in a significant increase in the Reynolds number and a more pronounced turbulence effect. The turbulence intensity index is a parameter characterizing the severity of flow field fluctuations. It is defined as the ratio of the root mean square of the velocity fluctuations to the average velocity, and its magnitude reflects the instability and mixing efficiency of the flow field.

[0056] Reynolds number of the j-th subregion The calculation is as follows: in, The density of gas under high pressure. Let be the average flow velocity in subregion j. Let j be the feature length of subregion j. This represents the gas dynamic viscosity.

[0057] Turbulence intensity index The calculation is as follows: in, Let be the instantaneous flow velocity of the j-th sub-region at the t-th sampling time. The length of the sampling time series.

[0058] The Reynolds number and turbulence intensity index together constitute the characteristics of the gas flow field, which are used to characterize the flow state and mixing properties of each sub-region under high pressure.

[0059] In this embodiment, accurate reconstruction of the concentration field under high pressure was achieved through radial basis function interpolation and Peng-Robinson equation of state correction; the dominant flow field modes were extracted through intrinsic orthogonal decomposition, achieving low-dimensional and efficient characterization of the high-dimensional flow field; and complete gas concentration distribution characteristics and gas flow field characteristics were constructed through gradient vector, curvature tensor, Reynolds number, and turbulence intensity index, providing high-quality input features for the subsequent construction of dynamic models, significantly improving the accuracy and efficiency of gas field modeling under high pressure.

[0060] It should be noted that under high pressure, the intermolecular distance of gas molecules decreases and the intermolecular forces increase, leading to a significant gas compressibility effect. The traditional RBF interpolation kernel function width parameter based on the assumption of normal pressure cannot accurately describe the spatial correlation of the gas concentration field under high pressure. Therefore, pressure-related corrections are needed for the RBF kernel function width parameter. Furthermore, the degree of gas concentration non-uniformity varies in different sub-regions within the chamber; using a uniform grid resolution would lead to… In one feasible implementation, based on discrete monitoring point data in the distributed monitoring dataset, a gas concentration field reconstruction algorithm based on radial basis function interpolation is used to generate a continuous gas concentration distribution field for each sub-region to obtain an initial concentration field. This may include: applying a pressure-related correction to the kernel function width parameter in the radial basis function interpolation based on the compressibility effect of gas molecules under high pressure to obtain a corrected radial basis function kernel function, wherein the pressure-related correction is inversely proportional to the absolute pressure inside the hyperbaric chamber; and regenerating the continuous gas concentration distribution field for each sub-region based on the corrected radial basis function kernel function to obtain a high-pressure corrected concentration field. The initial concentration field is obtained by: determining the concentration field nonuniformity index of each sub-region based on the oxygen partial pressure curvature tensor, wherein the concentration field nonuniformity index is the Frobenius norm of the oxygen partial pressure curvature tensor; assigning adaptive mesh refinement weights to each sub-region based on the concentration field nonuniformity index, wherein sub-regions with a nonuniformity index higher than a preset nonuniformity threshold are assigned higher mesh refinement weights, and sub-regions with a nonuniformity index lower than the preset nonuniformity threshold are assigned lower mesh refinement weights; and dynamically adjusting the mesh resolution of each sub-region based on the adaptive mesh refinement weights to obtain an adaptively meshed concentration field, which serves as the initial concentration field.

[0061] It should be noted that under high pressure, the compressibility of gas molecules alters the spatial scale of the concentration field. Traditional radial basis function kernel width parameters, designed for atmospheric pressure conditions, may lead to over-smoothing or excessive oscillation in high-pressure environments. Pressure-related corrections aim to adaptively adjust the kernel width based on the physical properties of the gas under high pressure, ensuring the interpolation more accurately reflects the concentration field distribution. As the absolute pressure inside the hyperbaric chamber increases, the corrected kernel width decreases, the kernel's support range shrinks, and the interpolation results become more localized, adapting to the steeper concentration gradients under high pressure. Subsequently, radial basis functions are regenerated based on the corrected kernel width parameters.

[0062] Understandably, the concentration field nonuniformity index is a scalar indicator that quantifies the severity of spatial changes in the concentration field. A higher value indicates a more dramatic change in the concentration gradient within that region, requiring a higher grid resolution for accurate capture. Adaptive mesh refinement is a method that dynamically adjusts the grid density based on the spatial variation characteristics of the field variable. It uses a denser grid in regions of dramatic field changes and a sparser grid in regions of moderate changes, achieving higher accuracy with the same computational resources. The preset nonuniformity threshold is the critical value that distinguishes between highly and low-uniform regions, determined based on the required concentration field reconstruction accuracy and computational resource constraints.

[0063] Step S30: Construct a dynamic model of convection-diffusion of multiple gases in the chamber under high pressure based on the gas concentration distribution characteristics and gas flow field characteristics.

[0064] It should be noted that the multi-component gas convection-diffusion dynamic model is a mathematical model describing the spatial redistribution of various gas components such as oxygen, nitrogen, and carbon dioxide in a hyperbaric chamber under the influence of a time-varying flow field through convective transport and molecular diffusion. It forms the core foundation for the design of model predictive controllers. The convection term characterizes the transport effect of gas velocity on the gas partial pressure distribution. Under high pressure, due to increased gas density, the convective transport capacity is enhanced, and the dominant role of the convection term becomes more significant. The diffusion term characterizes the diffusion effect of multi-component gas molecules driven by the concentration gradient. Under high pressure, due to the decrease in the mean free path of molecules and the reduction in the diffusion coefficient, the diffusion effect is relatively weakened, but still not negligible, especially in regions with lower flow velocities. The source-sink terms characterize the influence of oxygen input and gas consumption on the gas partial pressure field, with oxygen input as the source term, personnel respiration consumption as the sink term, and natural convection caused by equipment heat dissipation equivalent to distributed source-sink terms.

[0065] Understandably, the compressibility of gases under high pressure means that the relationship between gas partial pressure and mole fraction no longer satisfies a simple linear proportionality, necessitating the introduction of a non-ideal gas law that considers intermolecular interactions. Simultaneously, the local flow field distortion caused by personnel activity within the cabin and the thermal buoyancy effect caused by equipment heat dissipation, as unmodeled disturbances, exhibit randomness and time-varying intensity and spatial distribution. These need to be characterized as lumped disturbances in the dynamic model, providing a physical basis for the subsequent design of an adaptive disturbance observer.

[0066] In one feasible implementation, step S30 may include: determining the initial gas partial pressure field of each sub-region based on the gas concentration distribution characteristics, and determining the gas velocity field of each sub-region based on the gas flow field characteristics; establishing a multi-element gas mass conservation equation based on the initial gas partial pressure field and the gas velocity field, wherein the multi-element gas includes at least oxygen, nitrogen, and carbon dioxide; constructing a dynamic partial differential equation system of convection-diffusion coupling based on the multi-element gas mass conservation equation, wherein the dynamic partial differential equation system consists of convection terms, diffusion terms, and source-sink terms, the convection terms use the compressible Navier-Stokes equation under high pressure to characterize gas momentum transport, the diffusion terms use Fick's diffusion law considering the effects of nonideal gases under high pressure to characterize gas molecule diffusion, and the source-sink terms characterize the influence of oxygen input and gas consumption on the gas partial pressure field; discretizing the dynamic partial differential equation system in state space based on the nonlinear mapping relationship between gas partial pressure and mole fraction under high pressure to obtain a discrete state space model; and performing parameter identification and model verification on the discrete state space model to obtain a multi-element gas convection-diffusion dynamic model.

[0067] It should be noted that the initial gas partial pressure field refers to the spatial distribution of gas partial pressures at various locations within the chamber, determined based on the concentration field reconstruction results; it serves as the initial condition for the dynamic model. The gas velocity field refers to the spatial distribution of gas velocities at various locations within the chamber, determined based on flow field characteristics; it is the input to the convection term in the dynamic model. The initial gas partial pressure field is directly determined by the corrected concentration field. The initial oxygen partial pressure of each sub-region is the oxygen partial pressure value at the center of that sub-region. The nitrogen and carbon dioxide partial pressures are determined according to Dalton's law of partial pressures based on the total pressure and oxygen partial pressure, and require correction for non-ideal gases. The gas velocity field is obtained from the flow field modal characteristics reconstruction; the average velocity of each sub-region is the value of the time-varying coefficient of the flow field modal characteristics combined with the linear combination of the dominant mode at that sub-region.

[0068] Understandably, the mass conservation equation for a multi-component gas is the fundamental equation describing the change in mass of each gas component within the chamber over time. Its physical meaning is: the rate of mass change of a gas component within any control volume is equal to the net mass flow rate of that component entering the control volume through convection and diffusion, plus the mass generated by the source and sink terms. Under high pressure, due to the significant gas compressibility effect, the mass conservation equation must be in a compressible form.

[0069] For gas components The mass conservation equation is: in, Components density, The total density of the gas mixture. For the gas velocity field, Components The effective diffusion coefficient, Components mass fraction, Components The source and sink items.

[0070] Under high pressure, Significantly increased, Significantly reduced, convection term The dominant role of diffusion items has increased. Relatively weakened.

[0071] The compressible Navier-Stokes equations are a set of partial differential equations describing the conservation of momentum in compressible fluids. Under high pressure, due to significant changes in gas density, a compressible form must be used to accurately describe momentum transport. Fick's diffusion law, which considers the effects of nonideal gases under high pressure, is a modification of the classical Fick's law. Under high pressure, it needs to account for the influence of intermolecular interactions on the diffusion coefficient. The source and sink terms include the oxygen supply source term, the oxygen consumption term from personnel respiration, and the equivalent source and sink term caused by equipment heat dissipation. The convection term uses the compressible Navier-Stokes equations, and the diffusion term uses the modified Fick's law. The source and sink terms include the mass flow rate input of the oxygen supply system, the mass flow rate of oxygen consumed by personnel respiration, and the equivalent mass flow rate caused by equipment heat dissipation.

[0072] Understandably, state-space discretization refers to the process of transforming a system of partial differential equations in continuous time and continuous space into a system of difference equations in discrete time and discrete space, which facilitates numerical solutions and controller design. Under high-pressure environments, due to the nonlinear mapping relationship between gas partial pressure and mole fraction, state-space discretization must be performed within a nonlinear mapping framework.

[0073] The spatial domain is discretized using the finite volume method, dividing the cabin space into N parts. f Each control volume is divided into sub-regions. For each control volume, a partial mass conservation equation is obtained in a semi-discrete form. Then, a nonlinear mapping relationship between gas partial pressure and mole fraction is introduced. The state variables are selected as the gas partial pressure of each control volume, and the control inputs are selected as the air outlet opening, oxygen supply inlet flow rate, and fan speed, resulting in a discrete state-space model.

[0074] Specifically, the Peng-Robinson equation of state is used to describe the non-ideal gas behavior of a multi-element gas under high pressure, establishing a nonlinear mapping relationship between gas partial pressure and mole fraction. The Peng-Robinson equation of state considers the attractive forces between gas molecules and the molecular volume effect under high pressure. A time-varying oxygen sink term for personnel respiratory consumption is established based on the number of personnel in the cabin, their activity intensity, and metabolic rate. The time-varying oxygen sink term dynamically adjusts the oxygen consumption rate according to the personnel activity intensity. A time-varying oxygen source term for oxygen supplementation input is established based on the supply pressure and temperature of the oxygen supplementation system. The time-varying oxygen source term dynamically adjusts the oxygen supplementation mass flow rate according to the gas compressibility of the oxygen supplementation system. Substituting the time-varying oxygen sink term and the time-varying oxygen source term into the source and sink term, an enhanced source and sink term considering personnel activity and oxygen supplementation dynamics is obtained. Based on the enhanced source and sink term and the Peng-Robinson equation of state, the dynamic partial differential equation system is updated, and the updated dynamic partial differential equation system is discretized in state space to obtain a discrete state space model.

[0075] It is worth noting that parameter identification refers to the process of determining unknown parameters in the state-space model, such as the effective diffusion coefficient and convective heat transfer coefficient, based on measured data. Model validation refers to the process of evaluating the model's accuracy by comparing the model's predicted values ​​with the measured values. The least squares method or maximum likelihood method is used to identify the parameters of the discrete state-space model, with the objective function being the mean square error between the model's predicted values ​​and the measured oxygen partial pressure data. Model validation employs a cross-validation method, dividing the data into training and validation sets. Parameters are identified on the training set, and the prediction accuracy is evaluated on the validation set. When the prediction error on the validation set is less than a preset threshold, the model is considered to have passed validation.

[0076] In this embodiment, by establishing a multi-element gas mass conservation equation and a compressible Navier-Stokes equation that consider the high-pressure compressibility effect, the dynamic behavior of gas convection-diffusion coupling under high pressure is accurately described. By introducing a nonlinear mapping relationship between gas partial pressure and mole fraction through the Peng-Robinson equation of state, the deviation of the ideal gas assumption under high pressure is overcome. Through state-space discretization and parameter identification verification using the finite volume method, a discrete dynamic model suitable for model predictive control design is obtained, providing a reliable predictive model basis for subsequent controller design.

[0077] Step S40: Construct a model predictive controller based on the multi-element gas convection-diffusion dynamic model and solve for the optimal control sequence in each control cycle to obtain the optimal control command for the current control cycle.

[0078] It should be noted that Model Predictive Controller (MMCC) is an advanced control strategy based on a dynamic model of the controlled object. It solves a finite-time open-loop optimal control problem in each control cycle to obtain the optimal control action at the current moment, and then re-solves the problem based on new measurements in the next cycle to achieve closed-loop control. The prediction time domain refers to the step size range within which the MMCC predicts future states; its length must balance prediction accuracy and computational complexity. The control time domain refers to the step size range within which the MMCC optimizes control variables, typically not exceeding the prediction time domain. The optimization objective is the performance index function minimized by the MMCC in each control cycle. In this implementation, the optimization objectives are the oxygen partial pressure uniformity index and the total oxygen concentration stability index for each sub-region; the former characterizes spatial uniformity, and the latter characterizes temporal stability. Control variable constraints limit the physical capabilities of the environmental control system's actuators, including the upper and lower limits of vent opening, oxygen supply flow rate, and fan speed. State variable constraints limit the safety status of the cabin gas environment, including the safe oxygen partial pressure range for each sub-region, the upper limit of gas flow velocity, and the amplitude of total pressure fluctuation.

[0079] Understandably, the core advantage of model predictive control lies in its ability to explicitly handle the coupling constraints of multi-input multi-output systems, achieving a balance between global optimality and constraint satisfaction through rolling optimization within each control cycle. In the gas control scenario of a hyperbaric chamber, the oxygen partial pressures of different sub-regions are coupled, and there are complex nonlinear interactions between the oxygen supply inlet flow rate, vent opening, and fan speed. Traditional single-loop PID control struggles to coordinate these multi-variable coupling relationships, while model predictive control, by constructing a unified optimization problem, can simultaneously optimize all control variables, achieving global equilibrium of the gas partial pressure field within the chamber.

[0080] In one feasible implementation, step S40 may include: using the multi-element gas convection-diffusion dynamic model as a prediction model, and setting a prediction time domain and a control time domain; constructing an optimization objective, wherein the optimization objective includes at least an oxygen partial pressure uniformity index and a total oxygen concentration stability index for each sub-region, the oxygen partial pressure uniformity index for each sub-region being the weighted sum of squares of the deviations between the oxygen partial pressure of each sub-region and the target oxygen partial pressure, and the total oxygen concentration stability index being the square integral of the rate of change of the total oxygen concentration in the chamber; constructing control variable constraints, wherein the control variable constraints include at least a physical limit constraint on the air vent opening, a safe upper limit constraint on the oxygen supply inlet flow rate, and a fan speed range constraint; constructing state variable constraints, wherein the state variable constraints include at least a safe range constraint on the oxygen partial pressure of each sub-region, a gas flow rate upper limit constraint, and a total pressure fluctuation amplitude constraint; and within each control cycle, based on the current state measurement value, solving the minimization problem of the optimization objective under the control variable constraints and state variable constraints to obtain the optimal control command for the current control cycle.

[0081] It should be noted that the prediction time domain N p This refers to the number of steps the model predictive controller takes to predict future states. Its selection needs to balance prediction accuracy and computational complexity, and is typically taken as 5 to 10 times the system's dominant time constant. Control time domain N c This refers to the number of steps in optimizing the control input. In N c Subsequent control quantities are assumed to remain constant or decay according to a specific pattern. The sampling period T is set based on the dominant time constant τ of the hyperbaric chamber gas concentration field. s Typically, the time constant τ is taken as 1 to 10 seconds, and the dominant time constant τ is determined by the convection-diffusion characteristics, predicting the time domain. Control time domain .

[0082] Understandably, the optimization objective is to minimize the performance index of the model predictive controller in each control cycle, and its design directly determines the behavior of the control system. The oxygen partial pressure uniformity index characterizes the spatial uniformity of the oxygen partial pressure distribution in each sub-region of the chamber relative to the target oxygen partial pressure; a smaller value indicates a more uniform spatial distribution. The total oxygen concentration stability index characterizes the stability of the total oxygen concentration in the chamber over time; a smaller value indicates smaller concentration fluctuations. The optimization objective function is: in, The weight for the oxygen partial pressure uniformity of the i-th sub-region is given by [the weight]. The target oxygen partial pressure is set at 0.21~0.23 bar, corresponding to the partial pressure of 21%~23% oxygen concentration at 0.6 MPa under normal pressure. Let r be the predicted oxygen partial pressure of the i-th sub-region at the k-th prediction step size, and r be the total oxygen concentration stabilization weight. The total oxygen partial pressure inside the cabin at the k-th prediction step. To control the number of variables, The weight of the rate of change of the j-th control variable. Let be the control increment for the j-th control variable at the k-th prediction step. The first term is the oxygen partial pressure uniformity index, the second term is the total oxygen concentration stability index, and the third term is the control increment penalty term, used to suppress drastic changes in control action.

[0083] Understandably, control variable constraints limit the physical capabilities of the environmental control system's actuators and must be strictly met to ensure safe system operation. The physical limit constraint on vent opening refers to the mechanical limitation on the vent blade rotation angle, typically 0% to 100%. The safe upper limit constraint on oxygen supply flow rate refers to the maximum oxygen supply flow rate set to prevent excessively high local oxygen partial pressure; its value is related to the chamber volume, current oxygen partial pressure level, and number of personnel. The fan speed range constraint refers to the allowable speed range of the variable frequency fan, typically 20% to 100% of the rated speed. State variable constraints limit the safe state of the gas environment within the chamber and are crucial constraints for ensuring personnel safety. The safe oxygen partial pressure range constraint for each sub-zone refers to the upper and lower limits of oxygen partial pressure set to prevent oxygen poisoning or hypoxia. Under a 0.6 MPa high-pressure environment, the safe upper limit of oxygen partial pressure typically does not exceed 0.5 bar, corresponding to approximately 83% of the oxygen concentration at 0.6 MPa; exceeding this value poses a risk of oxygen poisoning. The lower limit is not lower than 0.16 bar, corresponding to approximately 27% of the oxygen concentration. The upper limit constraint on gas flow velocity refers to the upper limit of flow velocity set to prevent discomfort or dust caused by strong airflow, and it usually does not exceed 2 m / s. The total pressure fluctuation amplitude constraint refers to the range of total pressure variation set to maintain stable pressure inside the chamber, and it usually does not exceed ±5% of the set pressure.

[0084] It is worth noting that rolling solution refers to resolving the finite-time optimization problem based on the latest measurement value at the current moment within each control cycle, and implementing the control input for only the first control step, repeating this process in the next cycle. This rolling optimization mechanism enables the controller to continuously adapt to changes in system state and the effects of external disturbances.

[0085] The constrained optimization problem described above is solved using a sequential quadratic programming algorithm or an interior-point method. At each sampling time k, the measured value of the current state is used. As initial conditions, in the prediction time domain The internal optimization control sequence is solved to obtain the optimal control sequence for the current control cycle. Only the first control variable is output as the optimal control command for the current control cycle to the environmental control system actuator. At the next sampling time k+1, the latest cabin oxygen partial pressure measurement value is collected again, and the above optimization solution process is repeated to obtain the optimal control command for the next control cycle. This rolling process is used to achieve closed-loop control.

[0086] In this embodiment, by constructing an optimization objective that includes an oxygen partial pressure uniformity index and a total oxygen concentration stability index, the unified optimization of the spatial equilibrium and temporal stability of the gas partial pressure field inside the chamber is achieved; by setting control variable constraints and state variable constraints, the executability of control actions and the safety of the chamber environment are ensured; through a rolling optimization mechanism, the controller can continuously adapt to changes in high-pressure dynamic operating conditions, significantly improving the adaptability and robustness of the gas partial pressure equilibrium control.

[0087] It should be noted that traditional MPC controllers typically use fixed optimization target weights, which cannot adapt to dynamic changes in cabin conditions. For example, when personnel are actively engaged, local oxygen partial pressure fluctuations increase, and in this case, maintaining oxygen partial pressure uniformity should be prioritized. When personnel are at rest, oxygen consumption is stable, and maintaining a stable total oxygen concentration should be prioritized to save energy. Furthermore, under extreme conditions, such as multiple people engaging in strenuous activity simultaneously, control variable constraints and state variable constraints may conflict. For instance, if a large amount of oxygen supplementation is needed but the oxygen supply flow rate has reached its limit, a constraint softening strategy is required. This involves appropriately relaxing non-critical constraints while ensuring safety constraints, thus ensuring the optimization problem is solvable.

[0088] In one feasible implementation, within each control cycle, based on the current state measurement value, a rolling solution is performed to minimize the optimization objective under the constraints of the control variables and the state variables to obtain the optimal control command for the current control cycle. This includes: acquiring the activity intensity level and real-time number of personnel in the cabin, wherein the activity intensity level is classified according to the degree of local gas flow field distortion caused by personnel activity; and dynamically adjusting the weight coefficients of the oxygen partial pressure uniformity index and the total oxygen concentration stability index in each sub-region of the optimization objective based on the activity intensity level and the real-time number of personnel, wherein the oxygen partial pressure uniformity is increased when the activity intensity is high. The weighting coefficients of the indicators are adjusted, with the weighting coefficient of the total oxygen concentration stability indicator increased when the intensity of personnel activity is low. When a conflict is detected between the control variable constraints and the state variable constraints under the current operating conditions, a constraint softening strategy based on priority ranking is adopted. In this strategy, the oxygen partial pressure safety range constraints in each sub-region are non-softenable hard constraints, while the upper limit constraints of gas flow rate and the total pressure fluctuation amplitude constraints are softenable constraints. The softenable constraints are relaxed according to the priority ranking to obtain the softened constraints. Based on the dynamically adjusted weighting coefficients and the softened constraints, the minimization problem of the optimization objective is solved again in a rolling manner to obtain the optimal control command for the current control cycle.

[0089] It should be noted that the personnel activity intensity level is a discrete level based on the degree of flow field distortion, typically divided into four levels: resting, light activity, moderate activity, and heavy activity. The control requirements for the partial pressure field of the gas inside the cabin differ under different activity intensities: at high activity intensities, personnel breathing energy consumption increases and flow field disturbances are severe, requiring greater emphasis on spatial equilibrium; at low activity intensities, the flow field is relatively stable, requiring greater emphasis on temporal stability. The personnel activity intensity level is determined by analyzing the pulsating energy in the flow velocity sensor data.

[0090] Understandably, constraint conflicts refer to situations where control variable constraints and state variable constraints cannot be simultaneously satisfied under certain extreme operating conditions. For example, when a rapid reduction in the oxygen partial pressure of a certain area is required, the upper limit of the fan speed may restrict the adjustment speed. The constraint softening strategy involves appropriately relaxing secondary constraints while ensuring key safety constraints, thus making the optimization problem solvable. The constraint priority order is: oxygen partial pressure safety range constraint (highest priority, cannot be softened) > oxygen inlet flow rate safety upper limit constraint > gas velocity upper limit constraint > total pressure fluctuation amplitude constraint > vent opening physical limit constraint > fan speed range constraint (lowest priority, can be softened). When a constraint conflict is detected, it is gradually softened starting from the lowest priority constraint until the optimization problem is solvable. The softening method involves adding relaxation variables to the constraint boundaries.

[0091] It should be noted that the gas control system in the hyperbaric chamber exhibits significant dual-timescale characteristics: the airflow field responds quickly, regulated by the vent opening and fan speed; the concentration field responds slowly, regulated by the oxygen supply flow rate, with time constants typically on the order of minutes. Traditional single-timescale MPC controllers need to handle both fast and slow dynamics simultaneously, leading to difficulties in setting the prediction time domain. A prediction time domain that is too short cannot cover the slow dynamics of the concentration field, while a time domain that is too long results in a dramatic increase in computational load and limited control accuracy for the fast dynamics. This implementation method divides the control cycle into fast and slow timescales, designs separate controllers for each, and coordinates their synchronization, achieving a dual improvement in computational efficiency and control performance.

[0092] In one feasible implementation, determining the optimal control command for the current control cycle may further include: dividing the control cycle of the model predictive controller into a fast time scale and a slow time scale, wherein the fast time scale corresponds to rapid adjustment of the airflow field of the vent opening and fan speed, and the slow time scale corresponds to slow adjustment of the concentration field of the oxygen supply inlet flow rate, and the time step of the fast time scale is smaller than the time step of the slow time scale; within the fast time scale, a proportional-integral controller is used to rapidly track and adjust the fan speed and vent opening based on the gas flow field characteristics to obtain the airflow field adjustment command for the fast time scale; Within the slow timescale, based on the gas concentration distribution characteristics, the model predictive controller performs rolling optimization of the oxygen supply inlet flow rate to obtain a slow timescale concentration field adjustment command. A timescale decoupling algorithm coordinates and synchronizes the fast timescale airflow field adjustment command and the slow timescale concentration field adjustment command, wherein the timescale decoupling algorithm uses the principle that the airflow field adjustment command does not disrupt the stability of the concentration field as the coordination criterion, resulting in a fused multi-timescale control command. This fused multi-timescale control command is used as the optimal control command to execute the air outlet adjustment, oxygen supply control, and fan speed regulation of the environmental control system.

[0093] It should be noted that multi-timescale decoupled control is a method that decomposes the control system into multiple timescale levels for coordinated control based on the time constant differences of different physical processes. In hyperbaric chamber gas control, the response speed of the airflow field is much faster than that of the concentration field. Using a uniform timescale would lead to excessive computational burden or insufficient control accuracy. The control variables for the fast timescale are the vent opening and fan speed, while the control variable for the slow timescale is the oxygen supply inlet flow rate.

[0094] Understandably, using a proportional-integral (PI) controller to adjust fan speed and vent opening at fast timescales can quickly offset instantaneous disturbances in the cabin flow field. Its simple structure and low computational cost make it suitable for high-frequency adjustment requirements at fast timescales without introducing excessive computational load. The input to the PI controller is the deviation between the actual oxygen partial pressure and the target oxygen partial pressure in the current sub-region, and the output is the corresponding fan speed correction or vent opening correction. The proportional element responds quickly to deviations, and the integral element eliminates steady-state deviations, ensuring the airflow field can be quickly maintained in the expected state, providing the basic flow field conditions for stable concentration field adjustment. Fast timescale adjustment only rapidly adjusts the airflow distribution without changing the oxygen supply flow rate given at slow timescales. This fully utilizes the advantages of both controllers and reduces the optimization dimensionality of the model predictive controller through timescale decomposition, thus reducing the computational load of a single optimization.

[0095] Understandably, the slow-timescale model predictive controller executes at each slow sampling time, using the current concentration field state as initial conditions to predict the future N... p-slow By analyzing the concentration field evolution over a slow time step and optimizing the oxygen supply flow sequence, a concentration field regulation command on a slow time scale is obtained.

[0096] The coordination criteria of the time-scale decoupling algorithm ensure that fast-timescale airflow field regulation does not cause drastic fluctuations in the concentration field. Constraints are added to the fast-timescale controller to limit the rate of change of the vent opening and fan speed, so that the concentration field fluctuations caused by their changes are within the compensation capability of the slow-timescale controller.

[0097] Rapid stabilization of the airflow field was achieved through proportional-integral control with a fast time scale, global optimization of the concentration field was achieved through model predictive control with a slow time scale, and coordination and synchronization between the two time scales were achieved through a time-scale decoupling algorithm, which significantly improved the computational efficiency and hierarchical control capability of the control system.

[0098] Step S50: Based on the adaptive disturbance observer, estimate the impact of personnel activity disturbance and equipment heat dissipation disturbance on the gas flow field in real time to obtain the disturbance estimate value, and perform feedforward compensation on the optimal control command according to the disturbance estimate value to obtain the compensated optimal control command.

[0099] It should be noted that an adaptive disturbance observer is a state observer capable of estimating the amplitude and spatial distribution of unmodeled disturbances in a system in real time. It estimates disturbances as augmented states by expanding the state space and adaptively adjusts the observation gain based on the estimation error, achieving rapid tracking of time-varying disturbances. Lumped disturbance refers to the mathematical processing of multiple physical disturbance sources within the cabin, such as personnel activity and equipment heat dissipation, into a single disturbance vector, facilitating unified estimation and compensation. Feedforward compensation refers to a control strategy that directly superimposes the disturbance estimate onto the control command, preemptively offsetting the disturbance's impact before it affects the system output, resulting in a faster response speed compared to feedback control.

[0100] It is understandable that the local gas flow field distortion caused by personnel activity within the cabin is random and sudden, and its spatial distribution is closely related to personnel location and activity intensity. The thermal buoyancy effect caused by equipment heat dissipation is continuous and directional, and its intensity is related to equipment power, heat dissipation area, and cabin temperature difference. Under high pressure, these two disturbances have a more significant impact on the flow field due to increased gas density, making traditional fixed-gain observers difficult to adapt to the time-varying characteristics of disturbance intensity. This embodiment employs an adaptive gain adjustment module, dynamically adjusting the observation gain according to the flow field turbulence intensity index. The gain is increased during severe disturbances to improve tracking speed, and decreased during stable disturbances to suppress estimation noise, achieving a balance between disturbance estimation accuracy and robustness.

[0101] In one feasible implementation, step S50 may include: constructing an adaptive disturbance observer, wherein the adaptive disturbance observer includes a disturbance state estimation module and an adaptive gain adjustment module; taking the local gas flow field distortion caused by personnel activity in the cabin and the thermal buoyancy effect caused by equipment heat dissipation as lumped disturbance inputs and inputting them into the adaptive disturbance observer; using the disturbance state estimation module based on the state-space representation of the multi-element gas convection-diffusion dynamic model, and employing an extended state observer structure, estimating the amplitude and spatial distribution of the lumped disturbance in real time to obtain an initial disturbance estimate; and using the adaptive gain adjustment module... The observation gain of the extended state observer is dynamically adjusted based on the turbulence intensity index of the gas flow field characteristics to obtain an adaptive observation gain. The initial disturbance estimate is then corrected based on the adaptive observation gain to obtain the disturbance estimate. A disturbance compensation correction amount is determined based on the disturbance estimate and the optimal control command, wherein the disturbance compensation correction amount includes a disturbance compensation correction amount for the oxygen supply flow rate and a disturbance compensation correction amount for the fan speed. The disturbance compensation correction amount is then superimposed onto the corresponding control component of the optimal control command, and the superimposed control component is subjected to constraint projection processing to obtain the compensated optimal control command.

[0102] It should be noted that the adaptive disturbance observer is a state observer capable of dynamically adjusting the observation gain based on the disturbance characteristics. It expands the state space to treat the disturbance as an augmented state, enabling real-time estimation of the disturbance amplitude and spatial distribution. The disturbance state estimation module is responsible for calculating the estimated value of the disturbance based on the system model and measurement output, while the adaptive gain adjustment module is responsible for dynamically adjusting the observation gain according to the estimation error and flow field characteristics to balance estimation speed and noise suppression capability.

[0103] Understandably, lumped disturbance refers to the mathematical processing of equating multiple physical disturbance sources to a single disturbance vector. The local gas flow field distortion caused by personnel activity mainly manifests as random pulsations in the velocity field and vortex injection, the intensity of which is related to the number of personnel, the intensity of the activity, and the activity area. The thermal buoyancy effect caused by equipment heat dissipation mainly manifests as natural convection driven by the temperature gradient, the intensity of which is related to the equipment power, heat dissipation area, and the gas density inside the cabin. Under high pressure, the thermal buoyancy effect is enhanced due to increased gas density, and the momentum exchange caused by personnel activity is also more significant.

[0104] Understandably, the extended state observer is a method that incorporates disturbances as augmented states into the observer design. It achieves synchronous estimation of disturbances by adding disturbance states to the original system state. In the gas control scenario of a hyperbaric chamber, the spatial distribution characteristics of disturbances are crucial because the disturbance intensity varies significantly across different regions, necessitating a distributed observer structure.

[0105] Extend the discrete state-space model into an augmented model that includes perturbation states: in, This is the oxygen partial pressure state vector of each sub-region inside the cabin predicted in step k+1. This is the lumped perturbation augmented state vector predicted at step k+1. and These are the state transition matrix and input matrix of the original discrete state-space model, respectively. The perturbation input matrix is... It is the identity matrix. , The process noise is divided into the original state and the disturbed state, respectively. Let k be the control input vector at time k. Let k be the system measurement output vector at time k. The output matrix is ​​the original state. For measuring noise.

[0106] Design an extended state observer: in, This is the estimated vector of oxygen partial pressure state in each sub-region of the cabin predicted in the k-th step. Let be the lumped disturbance estimation vector predicted in the k-th step. This is the observation gain matrix for the oxygen partial pressure state. The observation gain matrix of the lumped perturbation. To observe the residuals, which reflect the deviation between the actual measured output and the state estimation output, the initial disturbance estimate is obtained by solving the above recursive equations. .

[0107] It is worth noting that adaptive observation gain refers to the observer gain that dynamically changes according to the flow field characteristics. Its design goal is to increase the gain to improve the tracking speed when the flow field is severely disturbed, and to decrease the gain to suppress estimation noise when the flow field is stable. The turbulence intensity index is the key basis for adjusting the gain, because the turbulence intensity directly reflects the instability of the flow field and the level of disturbance energy.

[0108] In practice, the adaptive observation gain is designed using a nonlinear mapping based on the turbulence intensity index: in, For adaptive observation gain, As the reference gain, For turbulence sensitivity coefficient, This represents the turbulence intensity index at the current moment. When... When it increases, The corresponding increase in speed allows the observer to track disturbances faster; when When decreasing, The noise reduction is correspondingly reduced, thus enhancing the observer's noise suppression capability.

[0109] By substituting the adaptive observation gain into the extended state observer equation and recalculating the disturbance estimate, the final disturbance estimate can be obtained.

[0110] It is worth noting that the disturbance compensation correction amount refers to the control command adjustment amount calculated based on the disturbance estimate. Its purpose is to offset the impact of the disturbance before it affects the system output. The disturbance compensation correction amount for oxygen supply flow rate is used to offset changes in oxygen partial pressure caused by personnel activities and equipment heat dissipation, while the disturbance compensation correction amount for fan speed is used to offset changes in flow velocity caused by flow field distortion.

[0111] The calculation of the disturbance compensation correction is based on the sensitivity analysis of the disturbance to the system state. The compensation correction is decomposed into the oxygen supply flow correction and the fan speed correction.

[0112] It is worth noting that constraint projection processing refers to the operation of projecting the superimposed control components onto the feasible region of the control variables, ensuring that the compensated control command still satisfies physical and safety constraints. Since the disturbance compensation correction may cause the control quantity to exceed the constraint range, projection processing is necessary to ensure system safety. The projection operator trims elements exceeding the constraint range to boundary values. The projected control command is the compensated optimal control command.

[0113] In this embodiment, real-time estimation of unmodeled disturbances within the cabin is achieved through an extended state observer, a dynamic balance between estimation accuracy and noise suppression is achieved through adaptive gain adjustment, and rapid feedforward compensation for disturbances is achieved through disturbance compensation correction and constraint projection processing, which significantly improves the system's ability to suppress random disturbances within the cabin and its control robustness.

[0114] It should be noted that the disturbance value estimated by the adaptive disturbance observer may contain estimation errors. Directly adding the disturbance compensation correction to the optimal control command may lead to instability of the closed-loop system. Therefore, stability constraints need to be imposed on the disturbance compensation correction to ensure that the compensated closed-loop control system meets the asymptotic stability condition. When the disturbance compensation correction exceeds the upper limit of the magnitude allowed by the stability constraint, a constraint projection algorithm based on quadratic programming is used to project the correction into the stability region.

[0115] In one feasible implementation, the step of superimposing the disturbance compensation correction amount onto the corresponding control component of the optimal control command, and performing constraint projection processing on the superimposed control component to obtain the compensated optimal control command, includes: constructing Lyapunov stability constraints on the disturbance compensation correction amount based on the disturbance estimate, wherein the Lyapunov stability constraints ensure that the compensated closed-loop control system satisfies the asymptotic stability condition, and the Lyapunov stability constraints are established based on the state-space representation of the multi-element gas convection-diffusion dynamic model and the negative qualitative requirement of the Lyapunov function derivative; based on the L... The Yapunov stability constraint determines the upper limit of the amplitude of the disturbance compensation correction, wherein the upper limit of the amplitude is negatively correlated with the turbulence intensity index of the gas flow field characteristics. When the disturbance compensation correction exceeds the upper limit of the amplitude, a constraint projection algorithm based on quadratic programming is adopted to minimize the deviation of the correction before and after projection. With the upper limit of the amplitude and the control variable constraint conditions as constraints, the disturbance compensation correction is projected into the range of the upper limit of the amplitude to obtain the projected disturbance compensation correction. The projected disturbance compensation correction is superimposed on the corresponding control component of the optimal control command to obtain the compensated optimal control command.

[0116] It should be noted that Lyapunov stability constraints are constraints based on Lyapunov stability theory. They ensure the asymptotic stability of the closed-loop system by constructing a Lyapunov function and requiring its time derivative along the system trajectory to be negative. In hyperbaric chamber gas control, disturbance compensation can lead to system instability, which must be limited by Lyapunov stability constraints. In this embodiment, a quadratic Lyapunov function is selected to construct the stability constraints. Based on this constraint, the upper limit of the disturbance compensation correction magnitude can be derived. When the turbulence intensity index is larger and the disturbance changes faster, the upper limit of the magnitude is smaller, avoiding excessive compensation correction that could cause system oscillations.

[0117] Understandably, as turbulence intensity increases, the upper limit of the amplitude decreases to prevent overcompensation under strong disturbances from leading to instability. When the disturbance compensation correction exceeds the upper limit of the amplitude, a constraint projection algorithm based on quadratic programming is used to project the correction onto a feasible region that satisfies stability and physical constraints. This ensures that the closed-loop control system always meets the asymptotic stability requirements and avoids system divergence or oscillation during the disturbance compensation process.

[0118] Step S60: Execute the air outlet adjustment, oxygen supplementation control and fan speed regulation of the environmental control system according to the compensated optimal control command, and adopt the oxygen supplementation rate regulation algorithm based on concentration gradient feedback to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation of each region to obtain the balanced multi-element gas partial pressure field.

[0119] It should be noted that the compensated optimal control command is a control decision resulting from both model prediction optimization and disturbance feedforward compensation, encompassing the target opening degree of each air outlet, the target flow rate of each oxygen supply inlet, and the target speed of each fan. The air outlet actuator refers to the electromechanical device that drives the air outlet blades to rotate and change the opening degree; its response speed and positioning accuracy directly affect the airflow regulation effect. A variable frequency fan is a fan device that adjusts its speed by changing the power supply frequency; its speed and airflow have an approximately linear relationship. The oxygen supply actuator refers to the valve or mass flow controller that controls the flow rate in the oxygen supply pipeline; its flow rate regulation accuracy directly affects the oxygen partial pressure control accuracy.

[0120] Understandably, the oxygen supplementation rate regulation algorithm based on concentration gradient feedback is one of the core innovations of this application. Traditional fixed-rate oxygen supplementation or simple threshold-triggered oxygen supplementation does not consider the concentration gradient relationship between the oxygen supplementation port and the target area, and the spatial distribution of oxygen supplementation flow lacks optimization, easily leading to local oxygen partial pressure peaks near the oxygen supplementation port. In this embodiment, the concentration gradient feedback gain is positively correlated with the absolute value of the real-time oxygen partial pressure deviation, ensuring that the hypoxic area receives more oxygen supplementation flow; at the same time, it is negatively correlated with the oxygen partial pressure gradient of adjacent sub-regions, avoiding excessive delivery of oxygen supplementation flow to areas already in a high oxygen partial pressure state, fundamentally eliminating the risk of excessively high local oxygen partial pressure. By dynamically correcting the oxygen supplementation flow distribution scheme, active prediction and adaptive equilibrium control of the multi-dimensional gas partial pressure field in the cabin are achieved.

[0121] It should be noted that the risk of oxygen toxicity is significant under high pressure: when the oxygen partial pressure exceeds 23.5 kPa and is continuously inhaled for more than 24 hours, pulmonary oxygen toxicity may occur; when the oxygen partial pressure exceeds 50 kPa and is continuously inhaled for more than 6 hours, cerebral oxygen toxicity may occur. Traditional control methods typically only set a single safety threshold and lack a tiered response mechanism, failing to effectively ensure the safety of personnel inside the cabin in the event of control system failure or extreme conditions. This implementation method establishes a multi-level safety interlock mechanism to achieve a three-level response of early warning, danger, and emergency, and dynamically arbitrates priorities with the MPC controller to ensure that the oxygen partial pressure safety constraint always has the highest priority under any circumstances.

[0122] In one feasible implementation, before executing the air outlet adjustment, oxygen supplementation control, and fan speed regulation of the environmental control system according to the compensated optimal control command, the method further includes: establishing a multi-level safety interlock mechanism for oxygen poisoning protection in high-pressure environments, wherein the multi-level safety interlock mechanism includes an oxygen partial pressure warning threshold, an oxygen partial pressure danger threshold, and an oxygen partial pressure emergency threshold, wherein the oxygen partial pressure warning threshold is lower than the oxygen partial pressure danger threshold, and the oxygen partial pressure danger threshold is lower than the oxygen partial pressure emergency threshold; real-time monitoring of the oxygen partial pressure value of each sub-region, and when the oxygen partial pressure value of any sub-region exceeds the oxygen partial pressure warning threshold, triggering a warning response and increasing the ventilation volume of the corresponding sub-region, while simultaneously sending a warning flag to the model prediction controller; ... triggering a warning response and triggering a fan speed regulation. When the oxygen partial pressure danger threshold is exceeded, a danger response is triggered and an emergency dilution damper is activated. Simultaneously, a danger flag is sent to the model predictive controller, freezing the oxygen supplementation control command of the model predictive controller and maintaining the current state of the fan speed and vent opening. When the oxygen partial pressure value in any sub-region exceeds the oxygen partial pressure emergency threshold, an emergency response is triggered and the oxygen supplementation source is cut off. Simultaneously, all dilution dampers and exhaust dampers are fully opened, and an audible and visual alarm is activated. During the triggering of the multi-level safety interlock mechanism, the constraints of the model predictive controller undergo dynamic priority arbitration. The safety interlock command has a higher priority than the optimal control command. Dynamic priority arbitration ensures that the oxygen partial pressure safety range constraint remains the highest priority constraint during the interlock triggering period.

[0123] It should be noted that the multi-level safety interlock mechanism is a key line of defense for ensuring the safety of personnel inside the cabin. It takes graded response measures when the oxygen partial pressure is abnormal by setting thresholds of different severity levels. In a high-pressure environment of 0.6 MPa, the setting of the oxygen partial pressure safety threshold must take into account the impact of high pressure on the risk of oxygen poisoning. Typically, the warning threshold is set at 0.35 bar, the danger threshold at 0.42 bar, and the emergency threshold at 0.50 bar.

[0124] Understandably, the early warning response includes: increasing the fan speed in the corresponding sub-region by 20%, increasing the vent opening to 80%, and enhancing gas mixing to reduce local oxygen partial pressure. Simultaneously, an early warning flag is sent to the model predictive controller, which then increases the uniformity weight of that sub-region in subsequent optimizations, prioritizing the reduction of oxygen partial pressure in that region.

[0125] Understandably, the hazard response includes: activating the emergency dilution damper to inject nitrogen or air into the chamber to dilute the oxygen concentration; freezing the oxygen supplementation control command, prohibiting any oxygen supplementation operation; maintaining the current fan speed and vent opening to prevent control actions from exacerbating oxygen partial pressure fluctuations. Simultaneously, a hazard warning is sent to the model predictive controller, which then enters a safety hold mode.

[0126] Understandably, an emergency response includes: immediately cutting off all oxygen supply sources and stopping any oxygen supply; fully opening all dilution and exhaust dampers to maximize gas dilution and exhaust; and activating the cabin's audible and visual alarms to notify personnel to take emergency measures. This level of response is of the highest priority and is not affected by any controller commands.

[0127] Dynamic priority arbitration is a decision-making mechanism that determines the execution order based on priority when multiple control commands conflict. In high-pressure manned environments, safety interlock commands involve personnel safety, and their priority must be higher than normal optimization control commands. Dynamic priority arbitration ensures that the oxygen partial pressure safety range constraint is the highest priority constraint under any circumstances. Dynamic priority arbitration is implemented through a constraint priority matrix, where matrix elements decrease sequentially in the following order: oxygen partial pressure safety constraint, personnel thermal comfort constraint, control increment constraint, control amplitude constraint, and partial pressure uniformity target weight. Safety constraints always occupy the highest priority position and will not be overridden by other constraints, thus ensuring that safety constraints remain effective after interlocking is triggered.

[0128] Through a multi-level safety interlock mechanism, graded response and active protection against oxygen partial pressure anomalies are achieved; through dynamic priority arbitration, the highest priority of safety interlock commands in the control system is ensured, fundamentally protecting the lives of personnel inside the cabin and significantly improving the safety and reliability of the gas control system under high pressure.

[0129] This embodiment provides a method for equalizing the partial pressure of multiple gases in a large space under high pressure. It achieves full-space coverage monitoring of the gas state within the cabin through a distributed sensor network. By reconstructing the concentration and flow fields based on radial basis function interpolation and intrinsic orthogonal decomposition, it achieves accurate mapping from discrete monitoring data to continuous distribution characteristics. By using a multi-gas convection-diffusion dynamic model that considers the effects of non-ideal gases under high pressure, it significantly improves the accuracy of predicting gas partial pressure evolution under high pressure. Through rolling optimization and multi-dimensional constraint processing of the model predictive controller, it achieves both forward-looking and safe control strategies. Through feedforward compensation of an adaptive disturbance observer, it effectively suppresses the impact of time-varying disturbances such as personnel activity and equipment heat dissipation. Through dynamic adjustment of the oxygen replenishment rate based on concentration gradient feedback, it achieves fine correction of oxygen partial pressure deviations in each sub-region. Finally, it forms a complete closed-loop control system of "monitoring-reconstruction-modeling-prediction-compensation-execution," realizing precise equalization control of the partial pressure of multiple gases in a large space under high pressure, significantly improving the safety, comfort, and control accuracy of the cabin gas environment.

[0130] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description and will not be repeated hereafter. Based on this, step S60 includes steps S601 to S606: Step S601: Based on the compensated optimal control command, the target opening degree of each air outlet, the target flow rate of each oxygen supply outlet, and the target speed of each fan are obtained.

[0131] It should be noted that analysis refers to the process of decomposing the compensated optimal control command vector into target values ​​for each actuator according to the control input type. The compensated optimal control command is a vector containing the target values ​​of all control degrees of freedom, which needs to be extracted separately for air outlets, oxygen supply outlets, and fans.

[0132] In actual implementation, let the optimal control command after compensation be... Then the target opening degree of each wind direction is The target flow rate for each oxygen supply port is: The target speed of each fan is .

[0133] Step S602: Drive the air outlet actuator to adjust the opening based on the target opening of each air outlet, and drive the variable frequency fan to adjust the speed based on the target speed of each fan, so as to obtain the adjusted gas flow environment.

[0134] It should be noted that the vent actuator refers to the electromechanical device that drives the vent blades to rotate and change the opening degree. It is usually driven by a servo motor, and its response time is typically on the order of hundreds of milliseconds. A variable frequency fan is a fan device that adjusts its speed by changing the power supply frequency. Its speed and air volume have an approximately linear relationship, and the speed adjustment range is usually 20% to 100% of the rated speed.

[0135] In actual implementation, the target implementation agency determines the level of openness based on the objective. A servo motor drive signal is generated, and the blades are rotated to the target angle through position closed-loop control. The variable frequency fan operates according to the target speed n. j The inverter generates frequency commands, and the fan speed is adjusted to the target value through frequency closed-loop control. The adjustment of the air outlet opening and the fan speed together change the gas flow state inside the chamber, forming a regulated gas flow environment.

[0136] Step S603: Under the adjusted gas flow environment, obtain the real-time oxygen partial pressure deviation of each sub-region, wherein the real-time oxygen partial pressure deviation is the difference between the current oxygen partial pressure and the target oxygen partial pressure of each sub-region.

[0137] It should be noted that the real-time oxygen partial pressure deviation is the core input signal for concentration gradient feedback control. Its sign and magnitude directly reflect the degree and direction of deviation of the oxygen partial pressure in each sub-region. A positive deviation indicates that the oxygen partial pressure is higher than the target value, and a negative deviation indicates that the oxygen partial pressure is lower than the target value.

[0138] In actual implementation, the real-time oxygen partial pressure deviation is calculated as follows: in, Let be the real-time oxygen partial pressure deviation of the i-th sub-region at time k. Let be the measured oxygen partial pressure in the i-th sub-region. The target oxygen partial pressure.

[0139] Step S604: Determine the concentration gradient feedback gain of each sub-region based on the real-time oxygen partial pressure deviation, wherein the concentration gradient feedback gain is positively correlated with the absolute value of the real-time oxygen partial pressure deviation and negatively correlated with the oxygen partial pressure gradient of adjacent sub-regions.

[0140] It should be noted that the concentration gradient feedback gain embodies the control principle of "more oxygen supplementation in hypoxic areas, less oxygen supplementation in hyperoxic areas, and slow oxygen supplementation in high gradient areas." It is positively correlated with the absolute value of the real-time oxygen partial pressure deviation, ensuring that hypoxic areas receive more oxygen supplementation flow; it is negatively correlated with the oxygen partial pressure gradient of adjacent sub-regions, avoiding excessive oxygen supplementation flow to areas that are already in a high oxygen partial pressure state.

[0141] In actual implementation, the concentration gradient feedback gain is calculated as follows: in, The concentration gradient feedback gain for the i-th sub-region is... As the reference gain, The gradient sensitivity coefficient, Let be the oxygen partial pressure gradient between the i-th sub-region and the adjacent j-th sub-region. When When it increases, Increase the oxygen supply flow rate; when When it increases, The oxygen supply flow rate is reduced, thus inhibiting excessive delivery to areas with high oxygen partial pressure.

[0142] It should be noted that the oxygen supply port in the hyperbaric chamber is usually some distance from the center of the sub-region. It takes time for oxygen to travel from the supply port to the center of the sub-region. This concentration transfer time lag τ is positively correlated with the spatial distance between the supply port and the sub-region center, and negatively correlated with the gas flow rate. This time lag effect causes phase lag in the concentration gradient feedback control, which may lead to control oscillations or overshoot. The Smith predictor is a classic time lag compensation method. By introducing a time lag model into the feedback loop, it predicts the delayed system response in advance, achieving phase lead compensation.

[0143] In one feasible implementation, step S604 may include: determining the concentration transfer delay time from each oxygen supply port to the corresponding sub-region based on the gas diffusion coefficient and gas density under high pressure, wherein the concentration transfer delay time is positively correlated with the spatial distance between the oxygen supply port and the center of the sub-region and negatively correlated with the gas flow rate; performing phase lead compensation on the real-time oxygen partial pressure deviation of each sub-region based on the concentration transfer delay time to obtain the compensated oxygen partial pressure deviation, wherein the phase lead compensation uses a Smith predictor structure to eliminate the influence of concentration transfer lag; determining the concentration gradient feedback gain based on the compensated oxygen partial pressure deviation and the oxygen partial pressure gradient in the gas concentration distribution characteristics, wherein the concentration gradient feedback gain is adaptively adjusted according to the change of the gas diffusion coefficient under high pressure, increasing the feedback gain when the gas diffusion coefficient decreases and decreasing the feedback gain when the gas diffusion coefficient increases.

[0144] It should be noted that the concentration transfer delay time refers to the time required for the supplemental oxygen gas to be delivered from the oxygen supply port to the target sub-region. In high-pressure, large-space environments, due to the large spatial scale and limited gas flow velocity, this delay time cannot be ignored and may lead to phase lag and overshoot in the control system. The Smith predictor is a classic time delay compensation method. It constructs a model of the time-delay process to predict and compensate for the control action in advance, eliminating the impact of time delay on closed-loop performance.

[0145] The concentration transfer delay time is calculated as follows: in, Let be the concentration transfer delay time from the i-th oxygen supply port to the j-th sub-region. For spatial distance, The average gas flow rate is denoted as .

[0146] The Smith predictor structure is as follows: in, This is the compensated oxygen partial pressure deviation. This refers to the real-time oxygen partial pressure deviation. The oxygen partial pressure is predicted by the time-delay model. The oxygen partial pressure is predicted by the time-delay model. Using the Smith predictor, the controller is based on the compensated deviation. Adjustments were made to eliminate phase lag caused by time delay.

[0147] The concentration gradient feedback gain is determined based on the oxygen partial pressure deviation after compensation and the oxygen partial pressure gradient in the gas concentration distribution characteristics. Since the gas diffusion coefficient is lower under high pressure than under normal pressure, the oxygen diffusion rate is slower. Therefore, the gain is adaptively adjusted with the diffusion coefficient. When the diffusion coefficient decreases, the feedback gain is increased to accelerate the oxygen supplementation response in the hypoxic region. When the diffusion coefficient increases, the feedback gain is decreased to avoid overshoot caused by excessive oxygen supplementation and to ensure the control stability after time delay compensation.

[0148] By calculating the concentration transfer delay time and using Smith predictor phase advance compensation, the adverse effects of gas transfer delay on control performance in high-pressure, large-space environments are eliminated. Through adaptive adjustment of dynamic concentration gradient feedback gain, the influence of gas diffusion coefficient changes on control response characteristics under high-pressure environments is adapted, significantly improving the timeliness and accuracy of oxygen supplementation rate regulation.

[0149] Step S605: Dynamically correct the target flow rate of each oxygen supply port according to the concentration gradient feedback gain to obtain the corrected oxygen supply flow rate allocation scheme.

[0150] It should be noted that dynamic correction refers to the process of adjusting the initial target flow rate of each oxygen supply port based on the concentration gradient feedback gain. The initial target flow rate comes from the optimization results of the model predictive controller. Dynamic correction makes local fine adjustments based on this to further optimize the spatial allocation.

[0151] In actual implementation, the corrected oxygen supplementation flow rate is: in, The corrected flow rate for the i-th oxygen supply port. Let be the initial target flow rate for the i-th oxygen supply port.

[0152] when <0, meaning oxygen deficiency. >0, increase oxygen supplementation flow rate; when >0, meaning when there is excess oxygen. <0, oxygen supplementation flow rate is reduced.

[0153] Step S606: Drive the oxygen supplementation actuator to adjust the oxygen supplementation rate based on the modified oxygen supplementation flow distribution scheme to obtain a balanced multi-element gas partial pressure field.

[0154] It should be noted that the oxygen supplementation actuator refers to the valve or mass flow controller that controls the flow rate of the oxygen supplementation pipeline, and its flow rate regulation accuracy directly affects the oxygen partial pressure control accuracy. Oxygen supplementation rate regulation refers to adjusting the gas supply flow rate at each oxygen supplementation port in real time according to the revised flow distribution scheme, so that the oxygen partial pressure field inside the chamber tends to be more balanced.

[0155] In actual operation, the oxygen supplementation actuator generates valve opening commands or mass flow controller setpoints based on the corrected flow rate, and adjusts the actual oxygen supplementation flow rate to the target value through flow closed-loop control. The coordinated adjustment of each oxygen supplementation port causes the oxygen partial pressure in each sub-region of the chamber to gradually converge to the target value, forming a spatially uniform and time-stable equilibrium multi-dimensional gas partial pressure field.

[0156] In this embodiment, by analyzing the optimal control command after compensation and driving the actuator, the conversion from control decision to physical action is realized; by acquiring the real-time oxygen partial pressure deviation and calculating the concentration gradient feedback gain, the accurate perception and response to the local oxygen partial pressure state is realized; by dynamically correcting the oxygen supplementation flow distribution scheme and driving the oxygen supplementation actuator, the goal of fundamentally eliminating the risk of excessively high local oxygen partial pressure is achieved, providing a safe, stable, and uniform gaseous living environment for the personnel inside the cabin.

[0157] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the pressure equilibrium control method of multi-dimensional gas in a large space under high pressure environment. Any simple modifications based on this technical concept are within the protection scope of this application.

[0158] This application also provides a high-pressure environment, large-space multi-element gas partial pressure equalization control device; please refer to [reference needed]. Figure 2 The high-pressure environment, large-space multi-gas partial pressure equalization control device includes: The acquisition module 10 is used to acquire real-time monitoring data of oxygen partial pressure and gas flow rate of each sub-region in the hyperbaric chamber, and to perform time alignment and coordinate unification of the real-time monitoring data of oxygen partial pressure and gas flow rate to obtain a distributed monitoring dataset.

[0159] Extraction module 20 is used to reconstruct the gas concentration field of sub-regions and extract the dynamic features of gas flow based on the distributed monitoring dataset, so as to obtain the gas concentration distribution features and gas flow field features of each sub-region; The construction module 30 is used to construct a dynamic model of convection-diffusion of multiple gases in a high-pressure environment based on the gas concentration distribution characteristics and gas flow field characteristics.

[0160] The construction module 30 is also used to construct a model predictive controller based on the multi-element gas convection-diffusion dynamic model and to solve the optimal control sequence in each control cycle to obtain the optimal control command for the current control cycle.

[0161] The compensation module 40 is used to estimate the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field in real time based on the adaptive disturbance observer, obtain disturbance estimates, and compensate the optimal control command based on the disturbance estimates to obtain the compensated optimal control command.

[0162] The adjustment module 50 is used to execute the air outlet adjustment, oxygen supplementation control and fan speed regulation of the environmental control system according to the compensated optimal control command, and adopts the oxygen supplementation rate adjustment algorithm based on concentration gradient feedback to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation of each region to obtain a balanced multi-element gas partial pressure field.

[0163] The high-pressure environment large-space multi-gas partial pressure equalization control device provided in this application adopts the high-pressure environment large-space multi-gas partial pressure equalization control method in the above embodiments. It can solve the technical problems of the oxygen partial pressure distribution in the cabin under high-pressure dynamic environment being highly sensitive to the flow field, and the traditional fixed air volume and fixed oxygen supplementation strategy being difficult to adapt to changes in working conditions, resulting in uneven gas partial pressure field and high local oxygen partial pressure leading to high oxygen toxicity among personnel. Compared with the prior art, the beneficial effects of the high-pressure environment large-space multi-gas partial pressure equalization control device provided in this application are the same as the beneficial effects of the high-pressure environment large-space multi-gas partial pressure equalization control method provided in the above embodiments, and other technical features in the high-pressure environment large-space multi-gas partial pressure equalization control device are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0164] This application also proposes a high-pressure environment large-space multi-gas partial pressure equalization control device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the high-pressure environment large-space multi-gas partial pressure equalization control method described above.

[0165] This application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the high-pressure environment large-space multi-element gas partial pressure equalization control method as described above.

[0166] The above are only some embodiments of this application and do not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for equalizing the partial pressure of multiple gases in a large space under high pressure, characterized in that, The method includes: Real-time monitoring data of oxygen partial pressure and gas flow velocity in each sub-region of the hyperbaric chamber are acquired, and the real-time monitoring data of oxygen partial pressure and gas flow velocity are time-aligned and coordinate-unified to obtain a distributed monitoring dataset. Based on the distributed monitoring dataset, sub-region gas concentration field reconstruction and gas flow dynamic feature extraction are performed to obtain the gas concentration distribution characteristics and gas flow field characteristics of each sub-region. Based on the gas concentration distribution characteristics and gas flow field characteristics, a dynamic model of multi-element gas convection-diffusion in a high-pressure environment is constructed. A model predictive controller is constructed based on the multi-element gas convection-diffusion dynamic model, and the optimal control sequence is solved in a rolling manner within each control cycle to obtain the optimal control command for the current control cycle. Based on the real-time estimation of the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field using an adaptive disturbance observer, disturbance estimates are obtained. Then, feedforward compensation is performed on the optimal control command based on the disturbance estimates to obtain the compensated optimal control command. The air outlet adjustment, oxygen supplementation control, and fan speed regulation of the environmental control system are executed according to the compensated optimal control command. An oxygen supplementation rate regulation algorithm based on concentration gradient feedback is adopted to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation in each region, so as to obtain a balanced multi-element gas partial pressure field.

2. The method as described in claim 1, characterized in that, The process involves acquiring real-time monitoring data of oxygen partial pressure and gas velocity in each sub-region of the hyperbaric chamber, and then aligning and unifying the time and coordinates of these data to obtain a distributed monitoring dataset, including: A distributed sensor network is deployed within the hyperbaric chamber, wherein the distributed sensor network includes an oxygen partial pressure sensor arranged at the center of each sub-region and a gas flow rate sensor arranged at the boundary of each sub-region. Acquire the raw oxygen partial pressure signal from each oxygen partial pressure sensor and the raw flow rate signal from each gas flow rate sensor; The original oxygen partial pressure signal is compensated for high pressure environment and temperature drift is corrected to obtain the corrected oxygen partial pressure data. The original flow velocity signal is corrected for high-pressure gas density and boundary effect to obtain corrected gas flow velocity data. The corrected oxygen partial pressure data and the corrected gas flow rate data are time-aligned using a timestamp alignment algorithm based on global clock synchronization to obtain time-aligned data. Based on the spatial coordinates of each sub-region, the time-aligned data is unified in coordinates and mapped to a grid to obtain a distributed monitoring dataset.

3. The method as described in claim 1, characterized in that, The process of reconstructing the gas concentration field and extracting the dynamic features of gas flow in sub-regions based on the distributed monitoring dataset yields the gas concentration distribution characteristics and gas flow field characteristics of each sub-region, including: Based on the discrete monitoring point data in the distributed monitoring dataset, a gas concentration field reconstruction algorithm based on radial basis function interpolation is used to generate the continuous gas concentration distribution field of each sub-region, thus obtaining the initial concentration field. The initial concentration field is corrected using the high-pressure environment gas state equation to obtain the corrected concentration field. Based on the corrected concentration field, the oxygen partial pressure gradient vector and oxygen partial pressure curvature tensor of each sub-region are determined to obtain the gas concentration distribution characteristics. Based on the gas velocity data in the distributed monitoring dataset, the dominant modes and time-varying coefficients of the gas flow field are extracted using the gas flow field order reduction reconstruction method based on intrinsic orthogonal decomposition, and the flow field modal characteristics are obtained. Based on the described flow field modal characteristics, the gas flow Reynolds number and turbulence intensity index of each sub-region are determined, and the gas flow field characteristics are obtained.

4. The method as described in claim 1, characterized in that, The construction of a dynamic model for the convection-diffusion of multiple gases within a high-pressure chamber based on the gas concentration distribution characteristics and gas flow field characteristics includes: The initial gas partial pressure field of each sub-region is determined based on the gas concentration distribution characteristics, and the gas velocity field of each sub-region is determined based on the gas flow field characteristics. Based on the initial gas partial pressure field and the gas velocity field, a mass conservation equation for a multi-component gas is established, wherein the multi-component gas includes at least oxygen, nitrogen, and carbon dioxide. Based on the mass conservation equation of the multi-element gas, a set of dynamic partial differential equations coupled with convection and diffusion is constructed. The set of dynamic partial differential equations consists of convection terms, diffusion terms, and source-sink terms. The convection term uses the compressible Navier-Stokes equation under high pressure to characterize gas momentum transport. The diffusion term uses Fick's diffusion law considering the effects of non-ideal gases under high pressure to characterize gas molecule diffusion. The source-sink terms characterize the effects of oxygen input and gas consumption on the gas partial pressure field. Based on the nonlinear mapping relationship between gas partial pressure and mole fraction under high pressure, the dynamic partial differential equations are discretized in state space to obtain a discrete state space model. The discrete state-space model is subjected to parameter identification and model verification to obtain a multi-element gas convection-diffusion dynamic model.

5. The method as described in claim 1, characterized in that, The process of constructing a model predictive controller based on the multi-element gas convection-diffusion dynamic model and solving for the optimal control sequence in each control cycle to obtain the optimal control command for the current control cycle includes: The multi-element gas convection-diffusion dynamic model is used as the prediction model, and the prediction time domain and control time domain are set. An optimization objective is constructed, wherein the optimization objective includes at least an oxygen partial pressure uniformity index and a total oxygen concentration stability index for each sub-region. The oxygen partial pressure uniformity index for each sub-region is the weighted sum of squares of the deviations of the oxygen partial pressure in each sub-region from the target oxygen partial pressure, and the total oxygen concentration stability index is the square integral of the rate of change of the total oxygen concentration in the cabin. Construct control variable constraints, wherein the control variable constraints include at least the physical limit constraint of the air outlet opening, the safe upper limit constraint of the oxygen supply outlet flow rate, and the fan speed range constraint; Construct state variable constraints, wherein the state variable constraints include at least the oxygen partial pressure safe range constraints, gas flow rate upper limit constraints, and total pressure fluctuation amplitude constraints for each sub-region; Within each control cycle, based on the current state measurement value, the optimization objective is minimized under the constraints of the control variables and the state variables in a rolling solution process to obtain the optimal control command for the current control cycle.

6. The method as described in claim 1, characterized in that, The method involves real-time estimation of the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field based on an adaptive disturbance observer, obtaining disturbance estimates, and then performing feedforward compensation on the optimal control command based on the disturbance estimates to obtain the compensated optimal control command, including: Construct an adaptive disturbance observer, wherein the adaptive disturbance observer includes a disturbance state estimation module and an adaptive gain adjustment module; The local gas flow field distortion caused by personnel activity inside the cabin and the thermal buoyancy effect caused by equipment heat dissipation are used as lumped disturbance inputs and input to the adaptive disturbance observer. The disturbance state estimation module uses the state space representation of the multi-element gas convection-diffusion dynamic model and an extended state observer structure to estimate the amplitude and spatial distribution of the lumped disturbance in real time, thereby obtaining the initial disturbance estimate. The adaptive observation gain is obtained by dynamically adjusting the observation gain of the extended state observer based on the turbulence intensity index of the gas flow field characteristics through the adaptive gain adjustment module. The initial disturbance estimate is corrected based on the adaptive observation gain to obtain the disturbance estimate. The disturbance compensation correction amount is determined based on the disturbance estimate and the optimal control command, wherein the disturbance compensation correction amount includes the disturbance compensation correction amount for the oxygen supply flow rate and the disturbance compensation correction amount for the fan speed. The disturbance compensation correction is superimposed on the corresponding control component of the optimal control command, and the superimposed control component is subjected to constraint projection processing to obtain the compensated optimal control command.

7. The method as described in claim 1, characterized in that, The process of executing the air outlet adjustment, oxygen supplementation control, and fan speed regulation of the environmental control system according to the compensated optimal control command, and using an oxygen supplementation rate regulation algorithm based on concentration gradient feedback to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation in each region, to obtain a balanced multi-element gas partial pressure field, includes: The target opening degree of each air outlet, the target flow rate of each oxygen supply outlet, and the target speed of each fan are obtained by analyzing the compensated optimal control command. Based on the target opening of each air outlet, the air outlet actuator is driven to adjust the opening, and based on the target speed of each fan, the variable frequency fan is driven to adjust the speed, so as to obtain the adjusted gas flow environment. In the adjusted gas flow environment, the real-time oxygen partial pressure deviation of each sub-region is obtained, wherein the real-time oxygen partial pressure deviation is the difference between the current oxygen partial pressure and the target oxygen partial pressure of each sub-region. The concentration gradient feedback gain of each sub-region is determined based on the real-time oxygen partial pressure deviation, wherein the concentration gradient feedback gain is positively correlated with the absolute value of the real-time oxygen partial pressure deviation and negatively correlated with the oxygen partial pressure gradient of adjacent sub-regions. Based on the concentration gradient feedback gain, the target flow rate of each oxygen supply port is dynamically corrected to obtain the corrected oxygen supply flow rate allocation scheme. Based on the modified oxygen supply flow distribution scheme, the oxygen supply actuator is driven to adjust the oxygen supply rate, thereby obtaining a balanced multi-element gas partial pressure field.

8. A pressure equalization control device for multi-element gas components in a large space under high pressure, characterized in that... The device includes: The acquisition module is used to acquire real-time monitoring data of oxygen partial pressure and gas flow rate of each sub-region in the hyperbaric chamber, and to align the real-time monitoring data of oxygen partial pressure and gas flow rate with time and coordinate to obtain a distributed monitoring dataset. The extraction module is used to reconstruct the gas concentration field of sub-regions and extract the dynamic features of gas flow based on the distributed monitoring dataset, so as to obtain the gas concentration distribution features and gas flow field features of each sub-region; The construction module is used to construct a dynamic model of convection-diffusion of multiple gases in a high-pressure environment based on the gas concentration distribution characteristics and gas flow field characteristics. The construction module is also used to construct a model predictive controller based on the multi-element gas convection-diffusion dynamic model and to solve the optimal control sequence in each control cycle to obtain the optimal control command for the current control cycle. The compensation module is used to estimate the impact of personnel activity disturbances and equipment heat dissipation disturbances on the gas flow field in real time based on the adaptive disturbance observer, obtain disturbance estimates, and perform feedforward compensation on the optimal control command based on the disturbance estimates to obtain the compensated optimal control command. The adjustment module is used to execute the air outlet adjustment, oxygen supplementation control and fan speed regulation of the environmental control system according to the compensated optimal control command, and adopts the oxygen supplementation rate adjustment algorithm based on concentration gradient feedback to dynamically adjust the oxygen supplementation flow distribution according to the oxygen partial pressure deviation of each region to obtain the balanced multi-element gas partial pressure field.

9. A pressure equalization control device for multi-element gas components in a large space under high pressure, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the high-pressure environment large-space multi-element gas partial pressure equalization control method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the high-pressure environment large-space multi-element gas partial pressure equalization control method as described in any one of claims 1 to 7.

Images