A method for separating gas components for industrial oxygen production

Abstract

This invention relates to the field of industrial oxygen production technology and discloses a gas component separation method for industrial oxygen production, comprising: dividing the adsorption bed into axial sampling nodes and collecting a set of mass transfer state parameters; inputting the set of mass transfer state parameters into an adsorption kinetic characteristic mapping table to determine an asymmetric pulse control sequence; driving an exhaust programmable valve to generate a transient expansion wave in the pores of the adsorption bed; utilizing the inertial oscillation of the airflow generated by the transient expansion wave to strip the adsorption boundary layer on the surface of the molecular sieve; achieving deep regeneration of the adsorbent under low flushing gas flow by offsetting intrinsic mass transfer resistance; effectively suppressing nitrogen penetration caused by nonlinear broadening of the temperature gradient; and improving component separation selectivity and yield.

Description

Technical Field

[0001] This invention relates to the field of industrial oxygen production technology, and in particular to a method for separating gas components for industrial oxygen production. Background Technology

[0002] Current industrial oxygen production uses pressure swing adsorption (PSA) technology to separate oxygen and nitrogen from the air. It utilizes the selective adsorption characteristics of zeolite molecular sieves for nitrogen and completes component enrichment and adsorbent regeneration through periodic alternation of pressure potential energy. The exothermic effect generated in the adsorption phase and the endothermic effect generated in the desorption phase inside the adsorption bed form a complex temperature gradient. Since the propagation rate of the heat wave inside the bed lags behind the concentration wave, the effective adsorption capacity dynamically shifts with the evolution of the local thermodynamic state. Conventional technology uses the solidification time step to trigger pressure equalization and venting actions. This static timing logic cannot adapt to the displacement of the mass transfer zone caused by fluctuations in the intake air load. Furthermore, in the desorption phase, the lack of physical disturbance to the fluid boundary layer on the adsorbent surface results in incomplete regeneration of the deep bed.

[0003] Existing regulation methods mostly rely on feedback from the purity of the produced gas at the outlet. However, the lag in gas diffusion makes it difficult for the outlet parameters to reflect the instantaneous physical field distortion inside the bed. Furthermore, existing technologies have the following shortcomings: 1. Fixed switching timing cannot adapt to the asynchronous evolution of thermal waves and concentration waves; 2. The continuous depressurization process is constrained by the fluid dynamic viscosity inside the porous medium, making it difficult to peel off the high-concentration nitrogen adsorption boundary layer on the surface of the adsorbent; 3. The compensation mechanism for pressure fluctuations lacks a deep understanding of the thermodynamic characteristics inside the bed, making it difficult for the system to balance oxygen production yield and regeneration energy consumption.

[0004] Therefore, the technical problem to be solved by this invention is how to extract the displacement law of thermodynamic feature points by monitoring the temperature-time derivative at a specific depth of the adsorption bed, and generate a desorption pulse sequence with asymmetric topological features in a pre-set waveform design constraint library based on the feature vector, thereby destroying mass transfer resistance and locking a safe physical boundary at the physical level. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for separating gas components for industrial oxygen production, comprising the following steps: Step S1: Divide the adsorption bed into multiple sampling nodes along the axial direction, collect the temperature change rate generated by the non-equilibrium mass transfer process at each sampling node and the transient pressure difference between each sampling node, and combine them into a mass transfer state parameter set. Step S2: The mass transfer state parameter set is used as the input vector, and a preset adsorption kinetic characteristic mapping table is matched to determine the asymmetric pulse control sequence. The asymmetric pulse control sequence includes valve opening time, valve closing time, and duty cycle parameters. The adsorption kinetic characteristic mapping table records the correlation data reflecting the nonlinear functional relationship between the mass transfer resistance field and the pulse control parameters. The corresponding asymmetric pulse control sequence is extracted by searching the correlation data point that is closest to the feature value in the input vector. Step S3: Drive the exhaust programmable valve to operate according to the pulse state matrix generated by the asymmetric pulse control sequence. Generate transient expansion waves excited by pressure pulses in the pores of the adsorption bed. Use the inertial oscillation of the airflow generated by the transient expansion waves to strip the nitrogen adsorption boundary layer on the surface of the molecular sieve, break the intrinsic mass transfer resistance limitation, and when the flushing gas volume is lower than the reference flushing gas volume threshold, use the duty cycle parameter in the asymmetric pulse control sequence to adjust the regeneration depth and suppress nitrogen penetration caused by the broadening of the temperature gradient.

[0006] Preferably, step S3 includes: dynamically adjusting the opening and closing frequency and duty cycle of the exhaust programmable valve based on the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed, forming a waveform evolution from high-frequency disturbance at the near end to low-frequency penetration at the far end; in the early stage of desorption, high-frequency pulses with a frequency higher than 10Hz are used to treat the shallow adsorbent, and as the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed decreases, the frequency is switched to a low-frequency pulse sequence with a frequency lower than 5Hz, and the waveform evolution is used to compensate for the dissipation of high-frequency components by the viscosity of the fluid in the adsorption bed, so that the transient expansion wave penetrates the full depth of the adsorption bed.

[0007] Preferably, the method further includes: statistically analyzing the absolute time offset of the thermal wave front across the sampling node to sense the degree of aging of the adsorbent pores; when the absolute time offset exceeds the physical boundary threshold, extracting the corresponding aging compensation weight in the adsorption kinetic characteristic mapping table, and using the aging compensation weight to correct the duty cycle parameter in the asymmetric pulse control sequence.

[0008] Preferably, the adsorption kinetic characteristic mapping table includes pulse sequence topological features preset for different local potential energy boundaries; in step S2, the pulse parameters corresponding to the current mass transfer resistance field are matched in the adsorption kinetic characteristic mapping table using the mass transfer state parameter set, so that the asymmetric pulse control sequence generates feedforward adjustment commands.

[0009] Preferably, the distribution density of sampling nodes is dynamically adjusted according to the change in thermal wave gradient in the middle of the adsorption bed; the sampling frequency is increased in the physical region corresponding to the front edge of the mass transfer zone to improve the spatial resolution of the temperature change rate.

[0010] Preferably, the valve opening time in the asymmetric pulse control sequence is shorter than the valve closing time, triggering boundary layer stripping through transient expansion waves, and maintaining the static pressure gradient within the adsorption bed using the closing interval determined by the valve closing time.

[0011] Preferably, during the waveform evolution process, the duty cycle in the asymmetric pulse control sequence increases as the pressure difference decreases, thereby increasing the energy penetration time of the pulse in the low-pressure stage and enhancing the desorption intensity of the deep adsorbent in a weak dynamic field environment.

[0012] Preferably, the oscillation frequency excited by the transient expansion wave in the pore flow channel of the adsorption bed is matched with the diffusion relaxation time of nitrogen molecules in the molecular sieve. Through the superposition of mechanical disturbance and thermodynamic potential difference, the adsorbed nitrogen molecules are detached from the molecular sieve pores.

[0013] Preferably, in step S3, the peak pressure of the transient expansion wave is maintained at 0.3 MPa to 0.5 MPa by adjusting the duty cycle parameter, so as to protect the structural strength of the adsorption bed while ensuring the peeling effect.

[0014] The beneficial effects of this invention are: 1. In gas component separation, by monitoring the time derivative of the temperature at a specific depth in the adsorption bed, the triggering logic of the equalizing phase and the reverse desorption phase is physically coupled with the dynamic displacement of the thermodynamic characteristic points inside the bed. This mechanism eliminates the lag in mass transfer zone identification caused by the fluctuation of the gas inlet load in the timed switching mode. Since the thermal wave front accurately corresponds to the mass transfer boundary, the method of the present invention keeps the mass transfer zone within the preset physical safety boundary throughout the entire cycle, avoiding the risk of nitrogen penetration caused by the nonlinear broadening of the temperature gradient.

[0015] 2. The adsorption bed is abstracted into a logical grid. By acquiring the real-time boundary state feature vector and addressing it in the pre-set waveform design constraint library, the pulse sequence generated by the reverse desorption phase has asymmetric topological characteristics that match the current physical field. This mechanism replaces the linear regulation logic of conventional pressure feedback. The discrete pulse state matrix drives the exhaust programmable valve. This asymmetric excitation based on physical feature vector triggering generates transient decompression expansion waves in the bed pores. The high-frequency stripping of the high-concentration nitrogen adsorption boundary layer on the surface of the molecular sieve is achieved by using the inertial oscillation of the airflow, breaking the constraint of intrinsic mass transfer resistance, and enabling the system to achieve deep regeneration at a low purging gas volume.

[0016] 3. Based on the pressure difference between the transient pressure in the intake manifold and the residual pressure in the bed, the opening and closing frequency and duty cycle of the pulse sequence are dynamically adjusted to form a waveform evolution from high-frequency disturbance at the near end to low-frequency penetration at the far end. In the early stage of desorption, high-frequency pulses are used to treat the shallow adsorbent. As the pressure difference decreases, the sequence switches to a low-frequency long-wave sequence. This frequency conversion mechanism compensates for the dissipation of high-frequency components due to the viscosity of the fluid dynamics in the large adsorption tower, ensuring that the stripping effect of the pulse expansion wave penetrates the entire depth of the bed. This process eliminates the deep desorption dead zone that exists in traditional continuous depressurization, enhances the consistency of the regeneration state in each region of the bed, and improves the yield of oxygen-producing components. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 This is a flowchart of the deep regeneration method based on asymmetric pulse excitation of the present invention; Figure 2 This is a block diagram of the mass transfer boundary sensing and pulse excitation control system of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0019] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, an embodiment or embodiment referred to herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. An embodiment appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0021] This invention is described in detail with reference to the schematic diagrams. When describing the embodiments of this invention, for ease of explanation, the cross-sectional views of the device structure will be partially enlarged without adhering to the general scale. Moreover, the schematic diagrams are only examples and should not limit the scope of protection of this invention. In addition, in actual manufacturing, the three-dimensional spatial dimensions of length, width and depth should be included.

[0022] Furthermore, in the description of this invention, it should be noted that the terms such as "upper," "lower," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or component referred to has a specific orientation, or is constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0023] Unless otherwise explicitly specified and limited, the terms installation, connection, and linking in this invention should be interpreted broadly. For example, they can refer to fixed connection, detachable connection, or integrated connection; similarly, they can refer to mechanical connection, electrical connection, or direct connection, or indirect connection through an intermediate medium, or internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0024] A method for separating gaseous components for industrial oxygen production includes the following steps: Step S1: Divide the adsorption bed into multiple sampling nodes along the axial direction, collect the temperature change rate generated by the non-equilibrium mass transfer process at each sampling node and the transient pressure difference between each sampling node, and combine them into a mass transfer state parameter set. Step S2: The mass transfer state parameter set is used as the input vector, and a preset adsorption kinetic characteristic mapping table is matched to determine the asymmetric pulse control sequence. The asymmetric pulse control sequence includes valve opening time, valve closing time, and duty cycle parameters. The adsorption kinetic characteristic mapping table records the correlation data reflecting the nonlinear functional relationship between the mass transfer resistance field and the pulse control parameters. The corresponding asymmetric pulse control sequence is extracted by searching the correlation data point that is closest to the feature value in the input vector. Step S3: Drive the exhaust programmable valve to operate according to the pulse state matrix generated by the asymmetric pulse control sequence. Generate transient expansion waves excited by pressure pulses in the pores of the adsorption bed. Use the inertial oscillation of the airflow generated by the transient expansion waves to strip the nitrogen adsorption boundary layer on the surface of the molecular sieve, break the intrinsic mass transfer resistance limitation, and when the flushing gas volume is lower than the reference flushing gas volume threshold, use the duty cycle parameter in the asymmetric pulse control sequence to adjust the regeneration depth and suppress nitrogen penetration caused by the broadening of the temperature gradient.

[0025] Preferably, step S3 includes: dynamically adjusting the opening and closing frequency and duty cycle of the exhaust programmable valve based on the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed, forming a waveform evolution from high-frequency disturbance at the near end to low-frequency penetration at the far end; in the early stage of desorption, high-frequency pulses with a frequency higher than 10Hz are used to treat the shallow adsorbent, and as the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed decreases, the frequency is switched to a low-frequency pulse sequence with a frequency lower than 5Hz, and the waveform evolution is used to compensate for the dissipation of high-frequency components by the viscosity of the fluid in the adsorption bed, so that the transient expansion wave penetrates the full depth of the adsorption bed.

[0026] Preferably, the method further includes: statistically analyzing the absolute time offset of the thermal wave front across the sampling node to sense the degree of aging of the adsorbent pores; when the absolute time offset exceeds the physical boundary threshold, extracting the corresponding aging compensation weight in the adsorption kinetic characteristic mapping table, and using the aging compensation weight to correct the duty cycle parameter in the asymmetric pulse control sequence.

[0027] Preferably, the aging compensation weight is determined in the following way: ,in, β is the aging compensation weight, and β is the preset response gain coefficient. The absolute time it takes for the heat wave to cross the sampling node as monitored in real time. This is the baseline span time for the adsorbent in its initial state.

[0028] Preferably, the adsorption kinetic characteristic mapping table includes pulse sequence topological features preset for different local potential energy boundaries; in step S2, the pulse parameters corresponding to the current mass transfer resistance field are matched in the adsorption kinetic characteristic mapping table using the mass transfer state parameter set, so that the asymmetric pulse control sequence generates feedforward adjustment commands.

[0029] Preferably, the distribution density of sampling nodes is dynamically adjusted according to the change in thermal wave gradient in the middle of the adsorption bed; the sampling frequency is increased in the physical region corresponding to the front edge of the mass transfer zone to improve the spatial resolution of the temperature change rate.

[0030] Preferably, the valve opening time in the asymmetric pulse control sequence is shorter than the valve closing time, triggering boundary layer stripping through transient expansion waves, and maintaining the static pressure gradient within the adsorption bed using the closing interval determined by the valve closing time.

[0031] Preferably, during the waveform evolution process, the duty cycle in the asymmetric pulse control sequence increases as the pressure difference decreases, thereby increasing the energy penetration time of the pulse in the low-pressure stage and enhancing the desorption intensity of the deep adsorbent in a weak dynamic field environment.

[0032] Preferably, the oscillation frequency excited by the transient expansion wave in the pore flow channel of the adsorption bed is matched with the diffusion relaxation time of nitrogen molecules in the molecular sieve. Through the superposition of mechanical disturbance and thermodynamic potential difference, the adsorbed nitrogen molecules are detached from the molecular sieve pores.

[0033] Preferably, in step S3, the peak pressure of the transient expansion wave is maintained at 0.3 MPa to 0.5 MPa by adjusting the duty cycle parameter, so as to protect the structural strength of the adsorption bed while ensuring the peeling effect.

[0034] Example 1: In a large-scale metallurgical field where the hourly oxygen production fluctuates by 20%, a gas component separation method for industrial oxygen production provided by this invention is used. This method includes a multi-tower parallel pressure swing adsorption oxygen production system. The adsorption bed is filled with lithium-based zeolite molecular sieves. Due to fluctuations in the inlet gas load caused by metallurgical process requirements, the leading edge of the nitrogen mass transfer zone inside the adsorption bed and the thermal wave front formed by adsorption exotherm exhibit nonlinear dynamic displacement in the axial position. During the adsorption phase operation, multiple sampling nodes set along the axial direction of the adsorption bed collect transient temperature sequences generated by the non-equilibrium mass transfer process. The system initiates a 50Hz high-speed sampling task, reading data from a 12-bit analog-to-digital converter every 20ms. A first-in-first-out sliding buffer with a length of 5 sample points is established in memory. By calculating the difference between the current sample mean and the sample mean 100ms ago in the buffer, and then dividing by a time span of 0.1s, the control unit calculates the time derivative based on the temperature sequence. When the time derivative When the preset positive threshold of 0.4℃ / s is reached, the characteristic depth of the internal thermal wave front reaching the sampling node is established. At this time, the system terminates the adsorption phase and switches to the equalization phase. The timing of the equalization phase action is physically coupled with the actual physical position of the mass transfer boundary layer, eliminating nitrogen penetration caused by fluctuations in the inlet flow rate. In the porous medium gas-solid mass transfer system, the latent heat of adsorption causes the solid phase heat wave transmission rate to lag behind the gas phase concentration wave propagation. Based on this physical transmission law, after identifying the characteristic depth, the control unit extracts the real-time airflow velocity of the inlet manifold and calculates the lag spatial bias by combining the pre-calibrated porosity of the adsorption bed. The control unit calculates the time compensation difference based on the spatial bias and the transient airflow velocity and issues the equalization phase trigger command in advance to block the penetration of high-concentration nitrogen ahead of the thermal wave front. When constructing the mass transfer state parameter set, the system calls the range normalization function to convert the obtained absolute pressure difference and temperature time derivative to the dimensionless range of 0 to 1 to ensure that each feature component of the combined multidimensional input vector has a level weight benchmark in the matching distance calculation.

[0035] In the reverse desorption phase, the system obtains the pressure difference between the transient pressure in the inlet header and the residual pressure in the adsorption bed, and modulates the pressure difference with the time derivative. The parameters are combined into a mass transfer state parameter set, which serves as the input vector. This set is used to match a pre-defined adsorption kinetics mapping table to determine the asymmetric pulse control sequence. The asymmetric pulse control sequence includes valve opening time, valve closing time, and duty cycle parameters. The exhaust control valve performs opening and closing actions according to the asymmetric pulse control sequence, generating transient expansion waves excited by pressure pulses within the pores of the adsorption bed. The inertial oscillation of the airflow generated by the transient expansion waves strips away the nitrogen adsorption boundary layer on the molecular sieve surface. The exhaust control valve is driven by the asymmetric pulse control sequence. In the initial stage of desorption, the system uses a frequency higher than 10Hz. The system uses high-frequency pulses to treat shallow adsorbents. As the pressure difference decreases, the system switches to a low-frequency pulse sequence with a frequency below 5Hz. Waveform evolution is used to compensate for the dissipation of high-frequency components due to fluid viscosity within the adsorption bed, allowing transient expansion waves to penetrate the full depth of the adsorption bed. During this process, oxygen purity is maintained at 93% and the flushing gas volume is below the baseline flushing gas volume threshold, achieving deep regeneration of the adsorbent. The method senses the degree of adsorbent pore aging by continuously statistically analyzing the absolute time offset of the thermal wave front across the sampling node. When the absolute time offset exceeds the physical boundary threshold, the system extracts the corresponding aging compensation weight from the adsorption kinetic characteristic mapping table. And utilize aging compensation weights The duty cycle parameter in the asymmetric pulse control sequence is modified to achieve adaptive hedging against physical configuration decay. The irreversible shrinkage of the pores in porous adsorbent materials follows a first-order decay kinetic law, and the relative increase in internal mass transfer resistance is directly proportional to the diffusion characteristic time drift. Based on a structural aging evolution model, the system collects dimensionless time offset rate as an evaluation benchmark to generate a modified duty cycle numerical multiplier factor and aging compensation weight. The calculation formula is as follows: ,in, β is the aging compensation weight, and β is the preset response gain coefficient. The absolute time it takes for the heat wave to cross the sampling node as monitored in real time. This is the baseline span time for the adsorbent in its initial state.

[0036] Example 2: In a pilot-scale gas separation test platform, the stability of the gas component separation method for industrial oxygen production provided by this invention under complex intake fluctuation conditions was verified. The test platform includes three parallel adsorption beds, each filled with lithium-based zeolite molecular sieves with an average particle size of 0.5 mm to 0.8 mm. The system is equipped with a fast electromagnetic programmable valve with a response time of 1 ms and a thermistor sensor group with a measurement accuracy of 0.05℃. To simulate flow fluctuations in a real metallurgical scenario, a random disturbance signal with an amplitude of 5% to 25% of the rated flow rate was introduced into the intake manifold. The sampling period was set according to the number of sensors. Based on a trade-off between throughput and the computational load of the control unit, when the monitored pressure fluctuation slope is above 0.1 MPa / s, to avoid aliasing of mass transfer signals at the sampling frequency, the sampling period is adjusted from 100 ms to 20 ms to capture the transient displacement of the thermal wave front. To coordinate with the adjustment of the sampling density, the distribution density of the sampling nodes is dynamically adjusted according to the change in the thermal wave gradient in the middle of the adsorption bed. Specifically, a one-dimensional logic grid is constructed by pre-arranging a fixed thermistor sensor array at equal intervals along the axial direction of the adsorption bed. The control unit monitors the leading edge position of the thermal wave gradient in real time based on dynamic addressing logic. When a signal is detected... When the heat wave front advances to a specific axial physical region, the activation topology matrix reconstruction at the control software level selects only the sensor group within this active region and includes it in the high-frequency inspection sampling channel. Data from adjacent nodes is interpolated and mapped with high spatial resolution. Sensor nodes in non-mass transfer active regions undergo data merging or frequency reduction / sleep processing. This allows for dynamic adjustment of the spatial distribution density of effective working sampling nodes and improves the spatial resolution of the temperature change rate, while maintaining the fixed physical positions of the hardware sensors. Three comparative schemes were set up during the experiment: Comparative scheme A using fixed timing switching and continuous pressure relief. In the control group B, which retains only the adsorption phase thermal wave triggering but uses a continuous flow mode for the desorption phase, and the experimental group using the complete method of this invention, under the condition that the initial inlet flow rate fluctuation rate is 10%, the original adsorption temperature curve collected by the control group A exhibits tooth-shaped fluctuations under background noise interference, which increases the uncertainty in determining the position of the nitrogen mass transfer zone front. The experimental group performs logical smoothing by calculating ∂T / ∂t and triggers pressure equalization switching instantaneously when the time derivative crosses the 0.4℃ / s threshold. At this time, the nitrogen concentration at the top of the adsorption bed is measured to be 0.02%, which reduces the deviation of physical boundary locking compared to 0.85% in the control group A.

[0037] During the desorption stage, the experimental group determined the asymmetric pulse control sequence by matching the adsorption kinetic characteristic mapping table, and used a 12Hz high-frequency pulse in the early stage of desorption. Under the rated adsorption conditions, the intrinsic diffusion relaxation time of nitrogen molecules in the micropores and mesopores of the lithium-based zeolite molecular sieve is 83ms to 200ms. The mechanical disturbance period corresponding to the 12Hz high-frequency pulse is 83.3ms. At this time, the alternating oscillation frequency of the pressure pulse waveform is controlled between 0.5 times and 1.2 times the reciprocal of the diffusion relaxation time, which meets the quantitative matching standard that the physical and mechanical disturbance period of the pulse waveform coincides with the characteristic time scale of nitrogen molecule desorption and outflow from the micropores. The transient expansion wave generated forms a micro-pressure vibration of 20Pa to 30Pa in the pores of the adsorption bed. The micro-pressure oscillation of 20Pa to 30Pa refers to the local micro-pressure pulsation amplitude of the reverse fluid decompression expansion wave with a total amplitude of 0.3MPa to 0.5MPa in the macroscopic mainstream channel. When the wave penetrates the molecular sieve particle accumulation layer, it is affected by the viscous dissipation and frictional damping low-pass filtering effect of the solid microporous channel, resulting in wave energy attenuation. Ultimately, a local micro-pressure pulsation amplitude of micro-alternating shear wave is induced and superimposed in the micro-fluid boundary layer on the surface of the molecular sieve particles. The macroscopic total pressure drop serves as the initial driving force, which is transformed into this local micro-pressure pulsation in the micro-pores. The two are unified in macroscopic dynamic driving and micro-scale energy dissipation, and their physical evolution logic is completely consistent. The system activates the built-in duty cycle follow-up compensation logic, and as the residual pressure inside the bed decreases from the initial 0.5MPa to close to the normal... As the driving force of the pressure difference between the fluid inside the adsorption bed and the external atmosphere gradually decreases, the control unit monitors the transient pressure difference in real time and issues a hardware timing adjustment command every time the pressure difference decreases by 0.1 MPa. This extends the opening time of the exhaust programmable valve by 8 ms and shortens the closing time by 8 ms, causing the duty cycle to increase stepwise from the initial 20% to the final 45%, extending the physical time for gas molecules to flow out of the pores under a weak driving force field. Based on the decreasing pressure difference trend, the frequency is lowered to 4.5 Hz to enhance deep penetration. Data shows that when the flushing gas volume of the experimental group is only 75% of that of the control group B, the benchmark flushing gas volume threshold used as the control is determined through a pre-calibration process. That is, during the initialization stage after the device is commissioned at the factory or after the molecular sieve is replaced, the adsorption system is first subjected to conventional continuous non-pulsed flushing. In the wash-regeneration mode, while keeping other inlet parameters constant, the amount of product oxygen introduced for wash-regeneration is gradually reduced, and the oxygen purity at the system outlet is monitored simultaneously. The minimum continuous wash oxygen volume flow rate required to stably maintain the oxygen purity at 93% of the rated value is defined as the baseline wash gas flow rate threshold. In this process unit, this threshold is calibrated as 15% of the rated oxygen production volume flow rate. When the wash gas flow rate in the actual operation of the test group decreases to 11.25% of the rated oxygen production volume flow rate, thus being forced to fall below the 15% baseline threshold, the control unit generates a transient expansion wave by adjusting the duty cycle parameter in the asymmetric pulse control sequence to deepen the regeneration depth, so that the specific surface area activity recovery rate of the adsorbent reaches 98% under low wash gas flow rate.To establish the physical boundaries of key parameters, an out-of-range control group C with a pulse frequency higher than 15Hz was set up. The experimental results showed that as the pulse frequency increased, the temperature rise inside the adsorption bed increased due to fluid viscosity. When the frequency reached 18Hz, the energy of the transient expansion wave was thermally dissipated in the front section of the mass transfer zone, resulting in a 15% decrease in the desorption activity of the deep adsorbent. This confirmed that the frequency range limitation constituted the optimal working window for this process. In the gradient test where the inlet gas fluctuation rate increased from 5% to 25%, the oxygen purity of the control group A decreased from 93.1% to 85.4%, while the experimental group, with its aging compensation weight, achieved a decrease. Dynamic correction of the pulse duty cycle maintained the purity between 92.5% and 93.2%. The unit oxygen production energy consumption of the experimental group was reduced by 12.6% compared with the control group A. It solved the problem of energy consumption and efficiency balance in the oxygen production process by quantitatively sensing the mass transfer boundary and intervening in the asymmetric waveform. Among them, ∂T / ∂t is the derivative of temperature with respect to time at the sampling node, which is used to characterize the movement rate of the thermal wave front. The aging compensation weight is calculated according to the following formula: Where β is the response gain coefficient. This refers to the absolute time as monitored in real time. Used as the base time.

[0038] Example 3: In the maintenance scenario of an industrial oxygen generator unit with a continuous operating time of 8000 hours, the lithium-based zeolite molecular sieve undergoes physical particle breakage due to long-term airflow pressure impact, resulting in non-uniform displacement of the packing density at the bottom of the adsorption bed. The change in packing density causes a local shift in the mass transfer resistance field, and the trajectory of the nitrogen mass transfer zone leading edge inside the adsorption bed is distorted accordingly, causing the oxygen production purity to deteriorate unidirectionally as the adsorbent ages. The method provided by this invention counteracts the above physical deviation by refining the mass transfer addressing logic. The control unit collects the temperature change rate generated by the non-equilibrium mass transfer process at each sampling node, and combines it with the transient pressure difference between each sampling node to form a mass transfer state parameter set. The system uses the mass transfer state parameter set as an input vector and matches it with the adsorption kinetic characteristic mapping table. The adsorption kinetic characteristic mapping table is constructed based on a one-dimensional axial logic grid, and the logic grid step size is 5% of the total height of the adsorption bed.

[0039] When searching for the optimal control command suitable for the current aging resistance field, the system calculates the Euclidean distance between the input vector and each associated data point in the logic grid, and selects the four neighboring nodes with the smallest Euclidean distance. Linear interpolation is then performed using the pulse parameters corresponding to these four neighboring nodes to generate an asymmetric pulse control sequence. The calculation formula is as follows: ,in, To match distance, These are the physical feature components in the input vector. To determine the response gain coefficient β, based on the baseline physical components stored in the adsorption kinetics characteristic mapping table, the system initiates the following calibration procedure: With the adsorption bed at the baseline operating pressure, the β value is adjusted incrementally in increments of 0.05 within the range of 0.05 to 0.5, while simultaneously monitoring the sensitivity of oxygen production purity to changes in flushing gas volume. When the fluctuation of oxygen production purity is less than 0.2% when the flushing gas volume decreases by 10%, the current β value is determined as the optimal response coefficient. Under these calibration conditions, when the system monitors the real-time time span... Compared to the baseline span time When the offset reaches 15%, the aging compensation weight extracted by the system Correcting the duty cycle parameter in the asymmetric pulse control sequence and aging compensation weights. The calculation formula is as follows: Where β is the response gain coefficient. The absolute time it takes for the heat wave to cross the sampling node as monitored in real time. This is the baseline span time in the initial state.

[0040] Example 4: During the on-site commissioning of a newly deployed industrial oxygen production air separation unit, due to the discreteness of different batches of lithium-based zeolite molecular sieves in terms of pore distribution and specific surface area activity, there is a risk of mismatch between the nonlinear mass transfer resistance field inside the adsorption bed and the preset benchmark data in the adsorption kinetic characteristic mapping table. If the system starts the separation cycle without calibrating the benchmark, it will cause a mismatch between the temperature change rate threshold captured by the thermal wave feedforward mechanism and the boundary locking of the mass transfer zone, resulting in fluctuations in the purity of oxygen production during the initial operation stage.

[0041] Before oxygen production, a standardized pre-calibration process was initiated to establish an operational baseline. This process included introducing high-purity nitrogen at a constant baseline inlet pressure of 0.5 MPa, sampling baseline data using the first five complete cycles after system startup, calculating the average time for thermal wave crossing nodes, and storing this data in non-volatile memory. Simultaneously, based on the pore decay physical model of lithium-based molecular sieves after 8000 hours of operation, the response gain coefficient was initially calibrated to 0.35. This value defines the linear correction ratio of the time offset to the pulse duty cycle. Sensor arrays located at the axial sampling nodes of the adsorption bed recorded the transient temperature step signal generated by the release of adsorption heat. The time corresponding to the maximum value of the first derivative of the transient temperature step signal was extracted and determined as the baseline crossing time in the initial state. The system adjusts the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed in increments of 0.05 MPa. At each pressure equilibrium point, the sensor array collects real-time feature vectors and calculates the pulse frequency required to peel off the boundary layer. The calculated on-time, off-time, and duty cycle parameters are written into the logical grid nodes of the adsorption kinetic characteristic mapping table to complete the initial data filling, so that the energy density of the transient expansion wave generated by the asymmetric pulse control sequence generated in subsequent operation is locked with the kinetic frequency in the adsorption bed.

[0042] Example 5: In the solidification of operating parameters of a newly built industrial oxygen production unit, the system conducts offline calibration and data filling processes to construct the underlying dataset of the adsorption kinetic characteristic mapping table, targeting the physical distribution of the mass transfer resistance field inside the lithium-based zeolite molecular sieve adsorption bed. This process includes using computational fluid dynamics simulation based on the Navier-Stokes equations to simulate the flow field distribution of the adsorption bed in the pressure range of 0.1 MPa to 0.8 MPa, obtaining the adsorption boundary layer thickness corresponding to different nitrogen loads by solving the porous media mass transfer model under each pressure gradient, determining the transient expansion wave energy density required to generate physical stripping, and encapsulating the corresponding parameters into logical key-value pairs. Multiple sets of logical key-value pairs generated by the simulation are written into the storage module of the control unit to complete the gridding filling of the adsorption kinetic characteristic mapping table.

[0043] When the system faces on-site calibration conditions regarding sensor deployment locations and mechanism response characteristics, the system initiates a pre-deployment calibration process to determine the physical spacing of the axial sampling nodes. In addition to the programmable valve action delay parameter, the system presets multiple test points along the adsorption bed axis and introduces high-purity nitrogen gas with controlled flow to generate a reference thermal wave. The spatial density of physical signal acquisition is determined by recording the time difference of the peak value of the first derivative of the temperature captured at each test point. When the time difference is within the range of 0.8s to 1.2s, the spacing at this time is determined as the physical spacing of the axial sampling nodes. The lag time of the pressure waveform jump in the pipeline after the exhaust programmable valve receives the opening command is monitored and defined as the valve action delay. The valve action delay is written into the generation operator of the asymmetric pulse control sequence to achieve time-domain compensation of the physical and mechanical response, so that the generated digital control sequence is phase-matched with the physical mass transfer process inside the adsorption bed.

[0044] In the logic construction scenario targeting the exfoliation of the adsorption boundary layer of lithium-based zeolite molecular sieves, the system deconstructs the spatiotemporal distribution of the Reynolds number of the fluid within the pores of the adsorption bed to determine the valve opening time in the asymmetric pulse control sequence. and valve closing time The non-isochronous proportional relationship means that when the calculated adsorption boundary layer thickness exceeds the intrinsic resistance critical value of 1.5 μm, the control unit retrieves the corresponding asymmetric operator from the adsorption kinetic characteristic mapping table and... Set to 45ms, and The time was set to 155ms to create an asymmetric pressure excitation with a duty cycle of 22.5%, causing the transient expansion wave generated inside the adsorption bed to exhibit a waveform characteristic of steep initial phase followed by a gentler phase. The high pressure gradient at the leading edge overcomes the viscous drag of gas molecules, while a relatively long shut-off interval maintains the inertial oscillation effect within the pores. This improves the mechanical work conversion efficiency of boundary layer stripping while keeping the total cycle time constant. To verify the stripping depth of the physical boundary layer by the transient expansion wave, a porous medium acoustic attenuation compensation algorithm was introduced during the offline calibration phase. This was achieved by measuring the amplitude attenuation coefficient of the pressure pulse during its axial propagation in the adsorption bed. To correct the pulse intensity parameter in the adsorption kinetics mapping table, when simulation results show that the wave energy loss due to fluid damping at the depth of the adsorption bed exceeds 30%, the control unit, in the process of generating the asymmetric pulse control sequence, counteracts the axial energy attenuation by increasing the peak pressure span of the initial pulse. The large-mass particle bed in the overall industrial adsorption tower exhibits an inherent hydrodynamic low-pass filtering effect on high-frequency aerodynamic disturbances, preventing the millisecond-level initial exhaust action from being completely dissipated by gas viscosity when transmitted to the depth of the bed. The control unit directly reconstructs the electrical drive signal based on the measured amplitude attenuation coefficient. The control unit extracts the valve opening time from the original pulse control sequence and multiplies it by a coefficient 1 + After obtaining the target opening range after stretching, the target opening range is sent to the servo drive circuit of the exhaust programmable valve. The signal command causes the valve core of the exhaust programmable valve to increase the physical opening by a corresponding proportion in each pulse cycle. The residence time of the additional release of transient gas flow is converted into long-wave mechanical thrust that penetrates the porous medium to the full depth. This causes the surface of the adsorbent particles at the bottom of the adsorption bed to also obtain tangential shear force sufficient to destroy the nitrogen molecule barrier. This physical-level kinetic energy replenishment mechanism eliminates the non-uniformity of the axial regeneration degree of the adsorption bed, ensuring that the average nitrogen residual concentration in the entire bed range is stable below 0.01% in the operation mode of lower than the reference flushing gas volume.

[0045] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the protection scope of the present invention.

Claims

1. A method for separating gaseous components in industrial oxygen production, characterized in that, Includes the following steps: Step S1: Divide the adsorption bed into multiple sampling nodes along the axial direction, collect the temperature change rate generated by the non-equilibrium mass transfer process at each sampling node and the transient pressure difference between each sampling node, and combine them into a mass transfer state parameter set. Step S2: The mass transfer state parameter set is used as the input vector, and a preset adsorption kinetic characteristic mapping table is matched to determine the asymmetric pulse control sequence. The asymmetric pulse control sequence includes valve opening time, valve closing time, and duty cycle parameters. The adsorption kinetic characteristic mapping table records the correlation data reflecting the nonlinear functional relationship between the mass transfer resistance field and the pulse control parameters. The corresponding asymmetric pulse control sequence is extracted by searching the correlation data point that is closest to the feature value in the input vector. Step S3: Drive the exhaust programmable valve to operate according to the pulse state matrix generated by the asymmetric pulse control sequence. Generate transient expansion waves excited by pressure pulses in the pores of the adsorption bed. Use the inertial oscillation of the airflow generated by the transient expansion waves to strip the nitrogen adsorption boundary layer on the surface of the molecular sieve, break the intrinsic mass transfer resistance limitation, and when the flushing gas volume is lower than the reference flushing gas volume threshold, use the duty cycle parameter in the asymmetric pulse control sequence to adjust the regeneration depth and suppress nitrogen penetration caused by the broadening of the temperature gradient.

2. The method for separating gas components in industrial oxygen production according to claim 1, characterized in that, Step S3 includes: dynamically adjusting the opening and closing frequency and duty cycle of the exhaust programmable valve based on the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed, forming a waveform evolution from high-frequency disturbance at the near end to low-frequency penetration at the far end; in the early stage of desorption, high-frequency pulses with a frequency higher than 10Hz are used to treat the shallow adsorbent, and as the pressure difference between the transient pressure in the intake manifold and the residual pressure in the adsorption bed decreases, the frequency is switched to a low-frequency pulse sequence with a frequency lower than 5Hz. The waveform evolution is used to compensate for the dissipation of high-frequency components by the viscosity of the fluid in the adsorption bed, so that the transient expansion wave penetrates the full depth of the adsorption bed.

3. The method for separating gas components for industrial oxygen production according to claim 1, characterized in that, The method also includes: statistically analyzing the absolute time offset of the thermal wave front across the sampling node to sense the degree of aging of the adsorbent pores; when the absolute time offset exceeds the physical boundary threshold, extracting the corresponding aging compensation weight in the adsorption kinetic characteristic mapping table, and using the aging compensation weight to correct the duty cycle parameter in the asymmetric pulse control sequence.

4. The gas component separation method for industrial oxygen production according to claim 1, characterized in that, The adsorption kinetics characteristic mapping table contains the pulse sequence topological features preset for different local potential energy boundaries; in step S2, the pulse parameters corresponding to the current mass transfer resistance field are matched in the adsorption kinetics characteristic mapping table using the mass transfer state parameter set, so that the asymmetric pulse control sequence generates feedforward adjustment commands.

5. The method for separating gas components for industrial oxygen production according to claim 1, characterized in that, The distribution density of sampling nodes is dynamically adjusted according to the change in thermal wave gradient in the middle of the adsorption bed; the sampling frequency is increased in the physical region corresponding to the front edge of the mass transfer zone to improve the spatial resolution of the temperature change rate.

6. The method for separating gas components for industrial oxygen production according to claim 1, characterized in that, In the asymmetric pulse control sequence, the valve opening time is shorter than the valve closing time, which triggers boundary layer stripping through transient expansion waves, and maintains the static pressure gradient in the adsorption bed by utilizing the closing interval determined by the valve closing time.

7. A method for separating gas components in industrial oxygen production according to claim 2, characterized in that, During waveform evolution, the duty cycle in the asymmetric pulse control sequence increases as the pressure difference decreases. By increasing the energy penetration time of the pulse in the low-pressure stage, the desorption intensity of the deep adsorbent in a weak dynamic field environment is enhanced.

8. A method for separating gas components in industrial oxygen production according to claim 1, characterized in that, The oscillation frequency excited by the transient expansion wave in the pore flow channel of the adsorption bed matches the diffusion relaxation time of nitrogen molecules in the molecular sieve. Through the superposition of mechanical disturbance and thermodynamic potential difference, the adsorbed nitrogen molecules are detached from the molecular sieve pores.

9. A method for separating gas components in industrial oxygen production according to claim 1, characterized in that, In step S3, the peak pressure of the transient expansion wave is maintained between 0.3 MPa and 0.5 MPa by adjusting the duty cycle parameter, so as to protect the structural strength of the adsorption bed while ensuring the stripping effect.

Images