Molecular sieve regeneration safety control method, system, device and medium based on thermal wave feedforward

Abstract

The present application relates to the technical field of industrial data processing and safety monitoring, and specifically provides a molecular sieve regeneration safety control method, system, device and medium based on thermal wave feedforward, which comprises the following steps: obtaining the mass flow of contaminated nitrogen gas, the inlet temperature and the actual output power of the electric heater under the regeneration state of the adsorber; constructing a thermal wave deduction model based on the mass flow of contaminated nitrogen gas and the physical property parameters of the multi-layer adsorbent, and generating the theoretical thermal wave arrival time of the thermal wave array surface to the predetermined measuring point position; collecting the measured temperature time series data of the measuring point, and extracting the temperature initial rising time point; comparing the temperature initial rising time point with the theoretical thermal wave arrival time to determine the time series deviation, determining the abnormal type according to the deviation direction; and performing intervention control according to the abnormal type before the measured temperature reaches the safety threshold. The present application replaces the absolute temperature amplitude criterion with the phase characteristics of the thermal wave arrival time, realizes early warning of local over-temperature, and avoids the collapse and pulverization of the molecular sieve skeleton.

Description

Technical Field

[0001] This invention relates to the field of industrial data processing and security monitoring technology, specifically to a method, system, equipment, and medium for the safety control of molecular sieve regeneration based on thermal wave feedforward. Background Technology

[0002] In cryogenic separation processes, the molecular sieve purification system is a critical safety barrier at the front end of the cold box, used to remove impurities such as water vapor, carbon dioxide, carbon monoxide, hydrocarbons, and nitrous oxide from the air. The purification system is typically equipped with dual adsorbers. A high-power electric heater heats the waste nitrogen gas (usually to around 200°C) to purge and desorb the adsorbers in the regeneration process at high temperature, thereby restoring the adsorption capacity of the adsorbent.

[0003] Existing purification systems rely primarily on single-point temperature sensors installed at the outlets of the electric heater and the molecular sieve bed. These sensors employ hard-interlock logic with a fixed temperature threshold (e.g., setting the alarm temperature to 220°C). When the detected temperature exceeds the threshold, an alarm is triggered, and the interlock action is executed. However, industrial-grade adsorbers are bulky, filled with tens of tons of activated alumina, molecular sieves, and catalysts, resulting in significant thermal inertia in the bed. When uneven airflow distribution or adsorbent collapse causes localized airflow short-circuiting within the adsorber, high-temperature airflow directly blows onto the affected area. Under the cover of thermal inertia, the bed outlet temperature probe initially fails to produce a significant amplitude jump. By the time the outlet temperature alarm is triggered, the internal bed has often already suffered from prolonged overheating, leading to molecular sieve framework collapse and pulverization. In other words, existing post-hoc comparison logic based on absolute temperature amplitude is insufficient to identify anomalies in the initial stages of localized overheating.

[0004] Furthermore, if the electric heater experiences localized dry burning or internal agglomeration leading to abnormal heat accumulation in a localized area, traditional temperature measurement points cannot detect it. If the air separation system inhales trace amounts of flammable hydrocarbons and these accumulates locally within the bed, undetected localized thermal runaway can easily trigger an internal fire in the adsorber, threatening the safety of the unit. To mitigate these risks, one existing approach is to densely distribute a matrix of temperature probes inside the adsorption bed to directly monitor the internal temperature field. However, this method is difficult to implement, the probes are easily damaged, and it increases the risk of bed leakage, resulting in high engineering costs.

[0005] To address these issues, the present invention provides a method, system, device, and medium for safe control of molecular sieve regeneration based on thermal wave feedforward. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention provides a method, system, equipment and medium for safety control of molecular sieve regeneration based on thermal wave feedforward, so as to solve the problems in the prior art.

[0007] One embodiment of the present invention provides a molecular sieve regeneration safety control method based on thermal wave feedforward, comprising the following steps: The operating condition data of the adsorber in the regeneration state are obtained, including the mass flow rate of waste nitrogen, the inlet temperature of waste nitrogen, and the actual output power of the electric heater. Based on the mass flow rate of waste nitrogen and the physical properties of the multilayer adsorbent in the adsorption bed, a thermal wave simulation model is constructed to simulate the propagation process of the thermal wave front in the adsorption bed. The theoretical thermal wave arrival time of the thermal wave front to the predetermined measuring point is generated based on the inlet temperature of the waste nitrogen using the thermal wave simulation model. Collect measured temperature time series data at predetermined measurement points, and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. The timing deviation is determined by comparing the initial temperature rise time with the theoretical arrival time of the heat wave. The type of anomaly in the adsorption bed is determined based on the direction of the time-series deviation. Specifically, when the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. Before the measured temperature reaches the preset safety threshold, intervention control is performed according to the type of abnormality.

[0008] This invention also relates to a molecular sieve regeneration safety control system based on thermal wave feedforward, comprising: The data acquisition module is used to acquire the operating condition data of the adsorber in the regeneration state. The operating condition data includes the mass flow rate of waste nitrogen, the inlet temperature of waste nitrogen, and the actual output power of the electric heater. The time extrapolation module is used to construct a thermal wave extrapolation model based on the mass flow rate of waste nitrogen gas and the physical property parameters of the multilayer adsorbent in the adsorption bed to extrapolate the process of the thermal wave front advancing in the adsorption bed, and to use the thermal wave extrapolation model to generate the theoretical thermal wave arrival time of the thermal wave front reaching the predetermined measuring point position based on the inlet temperature of waste nitrogen gas. The temperature rise determination module is used to collect measured temperature time series data at predetermined measurement points and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. The deviation determination module is used to compare the initial temperature rise time with the theoretical heat wave arrival time to determine the timing deviation. The anomaly detection module is used to determine the anomaly type of the adsorption bed based on the direction of the timing deviation. Specifically, when the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. The intervention control module is used to perform intervention control according to the type of abnormality before the measured temperature reaches the preset safety threshold.

[0009] The present invention also relates to a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for safe control of molecular sieve regeneration based on thermal wave feedforward.

[0010] The present invention also relates to a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for safe control of molecular sieve regeneration based on thermal wave feedforward.

[0011] The above embodiments provide a molecular sieve regeneration safety control method, system, equipment, and medium based on thermal wave feedforward, which have the following beneficial effects: This invention generates the theoretical arrival time of the heat wave front at a predetermined measuring point by constructing a heat wave simulation model. It then compares the initial temperature rise time extracted from measured temperature time-series data with the theoretical heat wave arrival time to determine the time-series deviation. The phase characteristics of the heat wave arrival time replace the traditional post-hoc comparison logic based on absolute temperature amplitude as the regeneration safety criterion. Therefore, even in the initial stage of local overheating under the masking effect of the adsorption bed's significant thermal inertia and before a significant amplitude jump in the bed outlet temperature, the direction of the time-series deviation can be used to identify flow field anomalies, adsorbent agglomeration, and insufficient electric heater power in advance. This invention identifies abnormal types such as foot temperature and enables early warning of localized thermal runaway. Intervention and control are implemented based on the abnormal type before the measured temperature reaches a preset safety threshold, fundamentally avoiding the problem of the molecular sieve skeleton collapsing and pulverizing due to prolonged overheating when the outlet temperature triggers an alarm. Furthermore, this invention eliminates the need for a dense matrix of temperature probes inside the adsorption bed; instead, it indirectly senses the internal heat transfer state of the bed using temperature sensors at predetermined measurement points combined with a thermal wave simulation model. This reduces construction difficulty and the risk of probe damage, avoids bed leakage hazards, and minimizes engineering implementation costs. Attached Figure Description

[0012] Figure 1 A flowchart of a molecular sieve regeneration safety control method based on thermal wave feedforward provided for an embodiment of the present invention; Figure 2This is a schematic block diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0013] The technical solutions in the embodiments of the present invention will now be clearly and completely described in conjunction with the accompanying drawings.

[0014] Reference Figure 1 One embodiment of the present invention provides a molecular sieve regeneration safety control method based on thermal wave feedforward, comprising the following steps: S10. Obtain operating condition data of the adsorber in regeneration state, including waste nitrogen mass flow rate, waste nitrogen inlet temperature and actual output power of electric heater. S20. Based on the mass flow rate of waste nitrogen and the physical properties of the multilayer adsorbent in the adsorption bed, a thermal wave simulation model is constructed to simulate the propagation process of the thermal wave front in the adsorption bed. The thermal wave simulation model is used to generate the theoretical thermal wave arrival time of the thermal wave front to the predetermined measuring point position based on the inlet temperature of waste nitrogen. S30. Collect measured temperature time series data at predetermined measurement points, and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. S40. Compare the initial temperature rise time with the theoretical heat wave arrival time to determine the timing deviation; S50. Determine the abnormality type of the adsorption bed based on the direction of the time sequence deviation. When the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. S60. Before the measured temperature reaches the preset safety threshold, perform intervention control according to the type of abnormality.

[0015] In this embodiment, specifically: As described in steps S10-S20 above, this scheme first acquires the operating condition data of the adsorber under regeneration conditions, and then constructs a thermal wave simulation model based on the operating condition data and physical property parameters to generate the theoretical thermal wave arrival time. Taking the molecular sieve purification system of a cryogenic air separation unit as an example, this purification system is equipped with dual adsorbers, one of which is in a regeneration heating state. The adsorber is filled from top to bottom with an activated alumina layer, a molecular sieve layer, and a catalyst layer, with the filling thicknesses of the three layers along the airflow direction being approximately 1.5m, 3m, and 1.5m, respectively. During the regeneration heating stage, the electric heater heats the waste nitrogen gas to a high temperature and then introduces it into the adsorber from top to bottom. Multiple temperature measuring points are set at the bed outlet to continuously collect the outlet temperature time-series data. Specifically, the control system first acquires the operating condition data of the adsorber under regeneration conditions, including the waste nitrogen gas mass flow rate, the waste nitrogen gas inlet temperature, and the actual output power of the electric heater. For example, in the above example scenario, the control system collects a waste nitrogen gas mass flow rate of approximately 30,000 N. The inlet temperature of the waste nitrogen gas is approximately 200℃, and the actual output power of the electric heater is approximately 800kW. The waste nitrogen gas mass flow rate reflects the transport rate of the heat carrier during regeneration and is a key factor determining the speed of heat wave propagation. The waste nitrogen gas inlet temperature is the initial temperature boundary condition for the regeneration heating process. The actual output power of the electric heater is used as a comparison reference for subsequent anomaly type determination.

[0016] After acquiring operational data, this scheme constructs a thermal wave simulation model based on the waste nitrogen mass flow rate and the physical properties of the multilayer adsorbents within the adsorption bed to simulate the propagation process of the thermal wave front within the adsorption bed. In the example scenario described above, the activated alumina layer, molecular sieve layer, and catalyst layer each have corresponding specific heat capacity and porosity parameters. Combined with the waste nitrogen mass flow rate, the thermal wave simulation model can quantitatively calculate the propagation rate of the thermal wave front through each layer, and thus estimate the total time for the thermal wave front to advance from the bed inlet to the outlet measuring point. The principle is that during the regeneration heating stage, high-temperature waste nitrogen gas flows downward through the bed and undergoes convective heat transfer with the low-temperature adsorbent, forming a temperature interface, i.e., a thermal wave front, within the bed. As the heating process continues, the thermal wave front gradually advances from top to bottom. Under normal operating conditions, the advancement speed of the thermal wave front is determined by parameters such as the mass flow rate of waste nitrogen gas, the heat capacity characteristics of the gas and adsorbent, and the porosity of the bed. Therefore, the time it takes for the thermal wave front to reach a predetermined position within the bed can be quantitatively predicted. By constructing a thermal wave prediction model and generating a theoretical thermal wave arrival time, this scheme provides a comparable benchmark value for subsequent measured thermal wave arrival times, replacing the traditional post-hoc comparison logic based on a fixed temperature threshold.

[0017] After the thermal wave simulation model is constructed, the control system uses this model, with the inlet temperature of the polluted nitrogen gas as the boundary condition, to generate the theoretical arrival time of the thermal wave front at the predetermined measuring point. Taking the example scenario above, based on the boundary condition of an inlet temperature of approximately 200℃, the thermal wave simulation model calculates that the theoretical time for the thermal wave front to advance from the bed inlet to the outlet measuring point is approximately 30 minutes. This theoretical arrival time serves as a reference time under normal operating conditions, used for comparison with the subsequently measured thermal wave arrival time, providing a quantifiable comparison benchmark for subsequent early identification of anomalies based on time phase deviation.

[0018] As described in step S30 above, this scheme collects measured temperature time-series data at predetermined measuring points. Based on the measured temperature time-series data and the waste nitrogen inlet temperature time-series data, the initial temperature rise time point is determined. Taking the above example scenario, the bed outlet measuring point continuously records the temperature change sequence data over time. Under normal operating conditions, the outlet measuring point temperature shows an initial rise inflection point approximately 30 minutes after the start of regeneration heating, which is basically consistent with the theoretical heat wave arrival time. When the heat wave front advances from top to bottom and reaches the outlet position, the outlet measuring point temperature sequence will show an initial rise inflection point. The control system, in conjunction with the waste nitrogen inlet temperature time-series data, identifies the time corresponding to this inflection point from the temperature sequence, which is the initial temperature rise time point. The principle is that industrial-grade adsorbers are bulky, filled with tens of tons of adsorbent, and the bed as a whole has enormous thermal inertia. When local airflow short-circuiting occurs within the bed due to uneven airflow distribution or adsorbent collapse, although the high-temperature airflow blows directly to the local area, the temperature at the bed outlet does not initially show a significant amplitude jump due to thermal inertia. However, by the time the outlet temperature reaches the alarm threshold (e.g., 220℃), the internal local bed has often already suffered molecular sieve framework collapse and pulverization failure due to prolonged overheating. In contrast, the arrival time of the thermal wave front is more sensitive to changes in the flow field and heat transfer state within the bed. When local airflow short-circuiting occurs, the effective velocity of the high-temperature airflow increases, the advancement speed of the thermal wave front accelerates, and the arrival time at the outlet measuring point is advanced. When the adsorbent agglomerates, the local heat transfer performance of the bed deteriorates or the equivalent heat capacity increases, the advancement speed of the thermal wave front slows down, and the arrival time at the outlet measuring point is delayed. This change in time phase occurs before a significant increase in temperature amplitude; therefore, using time phase as a monitoring feature can achieve early identification of anomalies. By extracting the initial temperature rise time point and comparing it with the theoretical heat wave arrival time, this scheme establishes an early sensing channel for changes in the internal flow field and heat transfer state of the bed, effectively breaking through the thermal inertia blind spot of traditional ex-post comparison logic based on temperature amplitude.

[0019] As described in steps S40-S50 above, this scheme compares the initial temperature rise time with the theoretical heat wave arrival time to determine the timing deviation, and determines the anomaly type of the adsorption bed based on the direction of the deviation. Taking the example scenario above, the control system extracts the initial temperature rise time from the measured temperature timing at the outlet measuring point and compares it with the theoretical heat wave arrival time generated by the heat wave simulation model, calculating the time difference between the two, which is the timing deviation. For example, in the example scenario above, the theoretical heat wave arrival time is approximately 30 minutes; if the measured heat wave arrival time at the outlet measuring point is 20 minutes, earlier than the theoretical heat wave arrival time, the timing deviation is -10 minutes; if the measured heat wave arrival time at the outlet measuring point is 45 minutes, later than the theoretical heat wave arrival time, the timing deviation is +15 minutes. After obtaining the timing deviation, this scheme determines the anomaly type of the adsorption bed based on the direction of the timing deviation. The principle is that the positive and negative directions of the time-series deviation correspond to the changes in the speed of the heat wave propagation, which are directly related to the flow field and heat transfer state inside the bed. When an anomaly occurs in the flow field inside the bed, the effective flow channel of the high-temperature polluted nitrogen changes, the local airflow velocity increases, and the propagation speed of the heat wave front increases accordingly, causing the measured heat wave to arrive at the measuring point earlier than the theoretical time, i.e., the time-series deviation is negative. When adsorbent agglomeration occurs inside the bed or the actual output power of the electric heater is insufficient, the local heat transfer performance of the bed deteriorates or the input heat decreases, the propagation speed of the heat wave front slows down, causing the measured heat wave to arrive at the measuring point later than the theoretical time, i.e., the time-series deviation is positive. By using the positive and negative directions of the time-series deviation, this scheme decouples the judgment of flow field anomalies from heat transfer anomalies, realizing the rapid identification of different types of anomalies and providing a basis for subsequent targeted intervention and control.

[0020] As described in step S60 above, this scheme performs intervention control based on the anomaly type before the measured temperature reaches the preset safety threshold. Taking the example scenario above, when the control system determines that there is an abnormal flow field in the adsorption bed, adsorbent agglomeration, and / or insufficient actual output power of the electric heater, the corresponding intervention control action is triggered before the measured temperature at the outlet measuring point reaches the preset safety threshold (e.g., 220℃). The specific intervention control methods include: adaptively reducing the output power of the electric heater and / or increasing the purging flow rate of ambient temperature polluted nitrogen to the adsorption bed according to the anomaly type and the severity of the timing deviation. For example, when it is determined that there is an abnormal flow field and the timing deviation is negative 10 minutes, the control system reduces the output power of the electric heater from about 800kW to about 400kW to reduce the heat input to the bed; at the same time, it opens the ambient temperature polluted nitrogen bypass valve to increase the purging flow rate of ambient temperature polluted nitrogen from about 30000N. Increased to approximately 45000N This method accelerates bed cooling. The principle is that traditional regeneration safety interlocks rely on the temperature amplitude reaching a fixed threshold before triggering alarms and interlock actions, by which time irreversible damage often occurs inside the bed. This solution identifies anomalies in advance through timing deviations, intervening before the temperature amplitude reaches a dangerous level, cutting off the thermal runaway chain at the source of energy input. Reducing the output power of the electric heater decreases the heat input to the bed, curbing temperature rise at its source; increasing the flow rate of ambient temperature nitrogen purging increases the cold input, accelerating the removal of heat from inside the bed. These two intervention methods can be used individually or in combination, depending on the type and severity of the anomaly, achieving precise and efficient intervention control. By implementing intervention control based on the anomaly type before the temperature amplitude reaches the safety threshold, this solution shifts from post-event alarms to pre-event intervention, effectively avoiding irreversible damage such as molecular sieve framework collapse and pulverization failure caused by prolonged overheating inside the bed. It also avoids the impact on the operational stability of the entire air separation unit caused by emergency shutdowns triggered by traditional threshold interlocks.

[0021] In summary, this solution achieves theoretical quantitative extrapolation of the thermal wave front's propagation process within the bed by acquiring operational data during the adsorber regeneration process and constructing a thermal wave simulation model based on the physical properties of the multilayer adsorbent. It also achieves phase feature extraction of the thermal wave front's arrival time by collecting measured temperature time-series data from measurement points and extracting the initial temperature rise time. Furthermore, it enables rapid decoupling and identification of flow field and heat transfer anomalies by comparing the measured arrival time with the theoretical arrival time to determine the time-series deviation and identifying the anomaly type based on the deviation direction. Finally, it transforms the approach from post-event alarm based on temperature amplitude to pre-event intervention based on time and phase by implementing intervention control based on the anomaly type before the measured temperature reaches the preset safety threshold. This effectively solves the engineering challenges of existing regeneration safety interlocks, which rely on absolute temperature amplitude comparisons, suffer from delayed local over-temperature identification due to bed thermal inertia, and require densely packed probes within the bed. This solution realizes a closed-loop data-driven safety control system for the entire molecular sieve regeneration process, from operational data acquisition, thermal wave simulation, phase comparison, anomaly identification to early intervention.

[0022] It should be noted that the specific values ​​such as the mass flow rate of polluted nitrogen, the output power of the electric heater, the filling thickness of each layer of adsorbent, and the theoretical heat wave arrival time involved in the above example scenarios are all exemplary descriptions given to facilitate those skilled in the art to understand the technical solution of the present invention, and are not intended to limit the scope of protection of the present invention. In practical applications, those skilled in the art can make corresponding adjustments to the above parameters according to the specific specifications of the adsorber, the adsorbent material, and the actual operating conditions.

[0023] In one embodiment, step S20, which involves constructing a thermal wave simulation model based on the mass flow rate of the polluted nitrogen gas and the physical property parameters of the multilayer adsorbent in the adsorption bed, specifically includes: S211. The adsorption bed is divided into multiple adsorbent sub-layers along the gas flow direction, and each adsorbent sub-layer has corresponding physical property parameters. S212. Obtain historical measured temperature time series data at predetermined measurement points in the historical regeneration cycle, and correct the physical property parameters of each adsorbent sublayer based on the historical measured temperature time series data to obtain the physical property correction parameters. S213. Based on the physical property correction parameters corresponding to each adsorbent sublayer and the mass flow rate of the polluted nitrogen gas, establish the unsteady heat transfer equations corresponding to each adsorbent sublayer. S214. Using the heat flow continuity condition and temperature continuity condition at the interface of adjacent adsorbent sublayers as boundary conditions, the various unsteady heat transfer equations are combined into a set of equations describing the propagation process of the heat wave front in the adsorption bed, thus obtaining the heat wave deduction model.

[0024] In this embodiment, specifically: As described in steps S211-S214 above, this scheme divides the adsorption bed into multiple adsorbent sub-layers along the gas flow direction, corrects the physical property parameters of each sub-layer based on historical measured data, establishes unsteady heat transfer equations for each sub-layer, and obtains a thermal wave extrapolation model by combining the interlayer continuity conditions as the boundary.

[0025] Taking the above example scenario, the adsorber is filled from top to bottom with an activated alumina layer, a molecular sieve layer, and a catalyst layer. The adsorbent materials in each layer are different, resulting in varying specific heat capacity, porosity, density, and other physical properties. The control system divides the adsorption bed along the gas flow direction (from top to bottom) into three sub-layers corresponding to each adsorbent layer. Each sub-layer has corresponding physical properties, including density, specific heat capacity, porosity, and effective thermal conductivity. These initial values ​​are derived from the design parameters of the adsorbent material or the physical property data provided by the supplier, and can be corrected based on measured data during subsequent regeneration cycles. Through layering, this scheme discretizes the bed, whose physical properties change continuously along the bed height, into several sub-layers with uniform physical properties. Each sub-layer can be considered a homogeneous medium, thus simplifying the complex three-dimensional heat transfer problem into a one-dimensional heat transfer problem within each sub-layer.

[0026] Based on the layered structure, the control system acquires historical measured temperature time-series data at predetermined measurement points throughout the historical regeneration cycle. It then corrects the physical property parameters of each adsorbent sublayer based on this data, obtaining corrected parameters. Specifically, the control system compares the theoretical temperature time-series predicted by the model with the measured temperature time-series from the historical regeneration cycle. Based on the deviation between the two, it uses the discrepancy to infer and correct the physical property parameters of each sublayer, making the corrected model predictions more consistent with the measured data. The principle is that the physical property parameters of the adsorbent are not static but slowly drift with the increase in regeneration cycles, adsorbent aging, and pulverization. If the physical property parameters from the initial design stage are always used for heat transfer model simulation, the theoretical heat wave arrival time predicted by the model will gradually deviate from the actual situation. By using measured temperature data from the historical regeneration cycle to correct the physical property parameters, this scheme enables the heat wave simulation model to adaptively track the slow changes in adsorbent properties, ensuring the accuracy of the theoretical heat wave arrival time prediction.

[0027] After obtaining the property correction parameters, the control system establishes the unsteady-state heat transfer equations for each adsorbent sublayer based on the property correction parameters and the mass flow rate of the polluted nitrogen gas. For the first... The general form of the unsteady-state heat transfer equation for an adsorbent sublayer is: Among them, subscript Indicates the first Each adsorbent sublayer; For the first Temperature within each sublayer; For the first The density of the adsorbent in each sublayer; For the first Specific heat capacity of each sublayer adsorbent; For the first Porosity of each sublayer; Density of polluted nitrogen gas; The specific heat capacity of polluted nitrogen gas; The mass flow rate of polluted nitrogen; The cross-sectional area of ​​the adsorption bed; For the first The effective thermal conductivity of each sublayer.

[0028] The physical meanings of each term in the equation are as follows: The first term on the left is the bed heat storage term, reflecting the change in the total heat capacity of the adsorbent and gas per unit volume of the bed over time; the second term on the left is the convection term, representing the heat carried by the downward flow of polluted nitrogen gas, and its coefficient... The product of the mass flow rate of waste nitrogen gas and the specific heat capacity at constant pressure, divided by the cross-sectional area, reflects the heat carrier's ability to transport heat; the right side is the thermal conductivity term, reflecting the conduction and diffusion of heat along the bed axis.

[0029] After establishing the equations for each sublayer, the control system uses the heat flow continuity condition and temperature continuity condition at the interface between adjacent adsorbent sublayers as boundary conditions to assemble the various unsteady heat transfer equations into a set of equations describing the propagation process of the thermal wave front within the adsorption bed, thus obtaining the thermal wave derivation model. Specifically, for the first... Sub-layer and the first Sub-layer interface At this point, the following two boundary conditions are satisfied: Temperature continuity conditions: Continuous heat flow conditions: The temperature continuity condition mentioned above indicates that the temperature on both sides of the interface is the same, and there is no temperature jump; the heat flow continuity condition indicates that the heat flow conducted from the upper layer at the interface is equal to the heat flow flowing into the lower layer, and there is no accumulation or loss of heat at the interface. By combining the unsteady heat transfer equations of each sublayer with the interlayer temperature continuity and heat flow continuity conditions as boundaries, a complete set of equations describing the spatiotemporal distribution of the temperature field within the entire adsorption bed can be obtained.

[0030] The principle is that the adsorption bed is a continuous heat transfer whole, and there is no sudden accumulation or disappearance of heat or sudden temperature jump at the interface of adjacent sub-layers. By combining the equations of each sub-layer with the interlayer continuity condition as the boundary, this scheme integrates the original independent heat transfer equations of each layer into a complete model that can describe the thermal wave propagation process of the entire bed. This not only preserves the accurate description of the differences in physical properties of each layer, but also ensures the physical consistency of the interlayer heat transfer process.

[0031] By employing layered modeling, self-correction of physical properties, and joint establishment of interlayer boundaries, the thermal wave projection model constructed in this scheme can accurately reflect the influence of differences in the physical properties of different adsorbent layers on the thermal wave propagation speed, and can also adaptively track the slow drift of adsorbent properties, providing an accurate and reliable model basis for the generation of theoretical thermal wave arrival time.

[0032] In one embodiment, step S212 specifically includes: S2121. Based on the time series data of the inlet temperature of waste nitrogen gas in the historical regeneration cycle and the time series data of the historical measured temperature, determine the historical measured thermal wave propagation time from the inlet of the adsorption bed to the predetermined measuring point. S2122. Based on the physical property parameters of each adsorbent sublayer, historical operating condition data corresponding to historical regeneration cycles, and the correspondence between thermal wave propagation rate and physical property parameters, the historical theoretical thermal wave propagation time from the adsorption bed inlet to the predetermined measuring point position is deduced. S2123. Compare the historical measured thermal wave propagation time with the historical theoretical thermal wave propagation time to determine the propagation time deviation; S2124. Based on the correspondence between the propulsion time deviation and the thermal wave propulsion rate and the physical property parameters, the physical property parameters of each adsorbent sublayer are corrected to obtain the corrected physical property parameters.

[0033] In this embodiment, specifically: As described in steps S2121-S2124 above, this scheme is based on the comparison between measured data and theoretical deduction data of historical regeneration cycles to invert and correct the physical property parameters of each adsorbent sublayer.

[0034] Taking the above example scenario, the control system acquires the time-series data of the inlet temperature of the waste nitrogen gas from the previous regeneration cycle and the historical measured temperature time-series data of the outlet measuring point. Specifically, the control system extracts the moment of the initial rise in inlet temperature from the waste nitrogen gas inlet temperature time-series data as the starting moment of the thermal wave front from the bed inlet; simultaneously, it extracts the moment of the initial rise in outlet temperature from the historical measured temperature time-series data of the outlet measuring point as the moment the thermal wave front arrives at the outlet measuring point; the difference between the two moments is taken as the historical measured thermal wave propagation time from the adsorption bed inlet to the predetermined measuring point position. The reason for extracting the measured thermal wave propagation time from the historical regeneration cycle is that the physical properties of the adsorbent will slowly drift during long-term operation. If the initial design values ​​or the physical properties provided by the supplier are always used, the theoretical thermal wave arrival time predicted by the model will gradually deviate from the actual situation. The historical measured thermal wave propagation time is an objective reflection of the actual heat transfer state of the bed and is not affected by the inaccuracy of physical properties. Therefore, it can be used as an independent reference benchmark for correcting physical properties. For example, in the previous regeneration cycle, the inlet temperature of the waste nitrogen gas began to rise 5 minutes after the start of regeneration, and the outlet measuring point temperature began to rise 35 minutes after the start of regeneration. Therefore, the historical measured heat wave propagation time is 30 minutes.

[0035] After obtaining the historically measured thermal wave propagation time, the control system, based on the physical properties of each adsorbent sublayer, historical operating data corresponding to historical regeneration cycles, and the correspondence between the thermal wave propagation rate and physical properties, extrapolates the historical theoretical thermal wave propagation time from the adsorbent bed inlet to the predetermined measuring point. Specifically, the control system uses the analytical correspondence between the thermal wave propagation rate and physical properties to calculate the propagation rate of the thermal wave front in each adsorbent sublayer, then calculates the propagation time of the thermal wave front traversing each sublayer based on the packing thickness of each sublayer along the airflow direction, and finally sums up the propagation times of each sublayer to obtain the historical theoretical thermal wave propagation time. The reason for using an analytical correspondence rather than a complete numerical solution model to extrapolate the historical theoretical time is to break the circular dependency that "correcting physical properties requires relying on a complete model, and a complete model requires accurate physical properties." There is a clear analytical relationship between the thermal wave propagation rate and physical properties, which can be directly calculated using only physical properties and operating data, without the need for complex partial differential equation numerical solutions. The computational load is small and the logic is independent, making it suitable as a reference benchmark for physical property correction. The principle lies in the clear analytical correspondence between the propagation rate of the thermal wave front within each adsorbent sublayer and the physical properties of that sublayer. For the... Each adsorbent sublayer has a thermal wave propagation rate The calculation formula is: in, This represents the mass flow rate of waste nitrogen gas during the historical regeneration cycle. The cross-sectional area of ​​the adsorption bed is... , , The first The density, specific heat capacity, and porosity of each sublayer adsorbent , These represent the density and specific heat capacity of the polluted nitrogen gas, respectively.

[0036] For example, the thickness of the activated alumina layer is 1.5m, the thickness of the molecular sieve layer is 3m, and the thickness of the catalyst layer is 1.5m. Based on the physical properties of each layer and the mass flow rate of the historical regeneration cycle, the propagation rate of the heat wave in each layer is calculated to be 0.15m / min, 0.2m / min, and 0.3m / min, respectively. The propagation time of each layer is 10 minutes, 15 minutes, and 5 minutes, respectively, and the total historical theoretical heat wave propagation time is 30 minutes.

[0037] After obtaining the historical theoretical thermal wave propagation time, the control system compares the historical measured thermal wave propagation time with the historical theoretical thermal wave propagation time to determine the propagation time deviation. Specifically, the difference between the historical measured thermal wave propagation time and the historical theoretical thermal wave propagation time is the propagation time deviation. Because neither the measured time nor the theoretical time alone can directly indicate whether the physical property parameter is inaccurate, it is necessary to compare the measured time with the theoretical time. Only by comparing the two can the degree of deviation between the current physical property parameter and the true value be quantified. The positive or negative direction and magnitude of the propagation time deviation directly reflect whether the physical property parameter is too large or too small, and how large the deviation is.

[0038] For example, if the historically measured thermal wave propagation time is 30 minutes and the historically theoretical thermal wave propagation time is also 30 minutes, then the propagation time deviation is 0, indicating that the current physical property parameters are accurate. If the adsorbent ages after several regeneration cycles, the actual propagation rate slows down, and the historically measured thermal wave propagation time becomes 35 minutes, then the propagation time deviation is positive 5 minutes, indicating that the actual total heat capacity is greater than the theoretical calculated value.

[0039] After obtaining the propulsion time deviation, the control system corrects the physical properties of each adsorbent sublayer based on the correspondence between the propulsion time deviation, the thermal wave propulsion rate, and the physical property parameters, thus obtaining the corrected physical property parameters. Specifically, the control system calculates the correction coefficient based on the propulsion time deviation, and then corrects the physical properties of each sublayer based on the correction coefficient. The reason why the time ratio can be directly used as the correction coefficient is that the thermal wave propulsion rate is inversely proportional to the total heat capacity of the bed, and the propulsion time is inversely proportional to the propulsion rate; therefore, the propulsion time is directly proportional to the total heat capacity. This means that the ratio of propulsion times is directly equal to the ratio of total heat capacities, and the correction coefficient can be directly calculated without complex iterative inversion, with clear physical meaning and simple calculation.

[0040] The specific correction method is as follows: correction coefficient For physical properties that affect total heat capacity (such as density and specific heat capacity), multiply their original values ​​by a correction factor. This yields the corrected physical property parameters. For example, if the historical measured propagation time is 35 minutes and the historical theoretical thermal wave propagation time is 30 minutes, then the correction factor... =35 / 30≈1.17; Multiply the density and specific heat capacity of each sublayer by 1.17 to obtain the corrected physical property parameters.

[0041] By using measured data from historical regeneration cycles to perform closed-loop correction of physical property parameters, this scheme enables the thermal wave extrapolation model to adaptively track the slow drift of adsorbent properties with the number of regeneration cycles, avoiding the gradual accumulation of prediction errors in theoretical thermal wave arrival time due to inaccurate physical property parameters.

[0042] In one embodiment, step S20, which involves using a thermal wave simulation model to generate the theoretical arrival time of the thermal wave front at a predetermined measuring point based on the inlet temperature of the polluted nitrogen gas, specifically includes: S221. Using the inlet temperature of polluted nitrogen as the boundary condition, solve the system of equations obtained by combining the unsteady heat transfer equations corresponding to each adsorbent sublayer to obtain the theoretical temperature rise time series data of each adsorbent sublayer. S222. Determine the theoretical propagation rate of the thermal wave front within each adsorbent sublayer based on the theoretical temperature rise time series data of each adsorbent sublayer. S223. Based on the loading geometric parameters of each adsorbent sublayer along the gas flow direction and the theoretical propulsion rate, determine the layer-by-layer propulsion time of the thermal wave front through each adsorbent sublayer. S224. The time of each adsorbent sublayer is accumulated to obtain the theoretical arrival time of the thermal wave at the predetermined measuring point.

[0043] In this embodiment, specifically: As described in steps S221-S224 above, this scheme uses a thermal wave simulation model to solve the problem with the inlet temperature of the polluted nitrogen gas as the boundary condition, and generates the theoretical arrival time of the thermal wave front at the predetermined measuring point.

[0044] Taking the above example scenario, after the thermal wave simulation model is constructed, the control system uses the inlet temperature of the polluted nitrogen gas as the boundary condition to solve the system of equations obtained by combining the unsteady heat transfer equations corresponding to each adsorbent sublayer, thus obtaining the theoretical temperature rise time series data for each adsorbent sublayer. Specifically, the control system uses the time series data of the polluted nitrogen gas inlet temperature as the temperature boundary condition at the bed inlet, substitutes it into the system of equations composed of the unsteady heat transfer equations of each sublayer and the interlayer continuity condition, and uses the finite difference method to numerically discretize and solve the partial differential equation system.

[0045] The solution process includes the following steps: First, the spatial domain is discretized: each adsorbent sublayer is divided into several computational nodes along the gas flow direction (axial direction), and the distance between adjacent nodes is the spatial step size. The total number of computational nodes for the entire bed is the sum of the number of nodes in each sublayer. For example, an activated alumina layer with a thickness of 1.5m is divided into 10 nodes, a molecular sieve layer with a thickness of 3m is divided into 20 nodes, and a catalyst layer with a thickness of 1.5m is divided into 10 nodes, for a total of 40 computational nodes for the entire bed, with a spatial step size of... It is 0.15m.

[0046] Secondly, the time domain is discretized: the total time of the regeneration heating process is divided into several time steps. The temperature of each node is calculated once within each time step. For example, taking a time step of... The calculation time is 10 seconds, and the total calculation time covers the entire regeneration heating process.

[0047] Then, the unsteady-state heat transfer equation is differentially discretized: the partial derivative of temperature with respect to time at each node is approximated by forward difference, and the first and second partial derivatives of temperature with respect to space are approximated by central difference, transforming the original partial differential equation into an algebraic relationship between the temperatures at each node. For example, for node... The unsteady heat transfer equation at point A is obtained by forward difference discretization of the time partial derivatives: The first-order partial derivatives in space are discretized using central difference: The second-order partial derivatives in space are discretized using central difference: Among them, superscript Indicates the first Each time step Indicates the first Each time step.

[0048] Within each time step, the inlet temperature of the polluted nitrogen gas at that moment is used as the temperature boundary condition for the bed inlet node, and the interlayer heat flow continuity and temperature continuity conditions are used as the coupling conditions for the interface nodes of adjacent sublayers. The algebraic relationships of all nodes are combined into a linear equation system, and the temperature values ​​of each node within that time step are obtained by solving the system. By progressively advancing the time step, from the start of regeneration heating to the end of regeneration heating, a complete sequence of temperature changes of each node within the entire bed over time can be obtained, i.e., the theoretical temperature rise time series data of each adsorbent sublayer. The reason for numerically solving the complete equation system is that the analytical formula used above to estimate the thermal wave propagation rate only yields the average propagation rate, while the numerical solution of the complete model can accurately reflect the influence of the differences in physical properties of each layer on the thermal wave propagation process, as well as the fine characterization of the spatiotemporal distribution of the temperature field by the interlayer boundary conditions, thus obtaining a more accurate theoretical temperature rise time series.

[0049] After obtaining the theoretical temperature rise time series data for each sublayer, the control system determines the theoretical propagation rate of the thermal wave front within each adsorbent sublayer based on this data. Specifically, for each adsorbent sublayer, the control system selects several observation points along the axial direction within that sublayer and extracts the initial temperature rise time from the theoretical temperature rise time series data for each observation point. Then, based on the axial position of each observation point and the corresponding initial temperature rise time, the propagation rate of the thermal wave front within that sublayer is fitted. For example, three observation points are selected within the molecular sieve layer at distances of 0.5m, 1.5m, and 2.5m from the bottom of the layer. The initial temperature rise times for each observation point are extracted from the theoretical temperature rise time series data at the 10th minute, 17.5 minutes, and 25th minute, respectively. Based on the linear fitting of the axial position and time, the theoretical propagation rate of the thermal wave within the molecular sieve layer is approximately 0.2m / min.

[0050] After determining the theoretical propulsion rate of each sublayer, the control system determines the propulsion time of the thermal wave front through each sublayer based on the loading geometry parameters and theoretical propulsion rate of each adsorbent sublayer along the gas flow direction. Specifically, the propulsion time of the thermal wave front through each sublayer is obtained by dividing the loading thickness of each adsorbent sublayer by the theoretical propulsion rate within that sublayer. For example, the propulsion time is 10 minutes for an activated alumina layer with a thickness of 1.5 m and a theoretical propulsion rate of 0.15 m / min; 15 minutes for a molecular sieve layer with a thickness of 3 m and a theoretical propulsion rate of 0.2 m / min; and 5 minutes for a catalyst layer with a thickness of 1.5 m and a theoretical propulsion rate of 0.3 m / min.

[0051] After obtaining the advancement time of each sublayer, the control system sums up the advancement times of each adsorbent sublayer to obtain the theoretical thermal wave arrival time for the thermal wave front to reach the predetermined measuring point. For example, by summing up the 10 minutes for the activated alumina layer, the 15 minutes for the molecular sieve layer, and the 5 minutes for the catalyst layer, the theoretical thermal wave arrival time from the bed inlet to the outlet measuring point is obtained as 30 minutes.

[0052] By numerically solving the complete model with inlet temperature as the boundary condition and calculating and accumulating the propagation time layer by layer, the theoretical heat wave arrival time generated by this scheme not only considers the influence of the differences in physical properties of each layer on the heat wave propagation speed, but also accurately reflects the fine characteristics of the interlayer heat transfer process through numerical solution, providing an accurate and reliable theoretical benchmark for subsequent comparison with the measured heat wave arrival time.

[0053] In one embodiment, step S30 specifically includes: S311. Obtain time-series data of waste nitrogen inlet temperature, and fit a dynamic baseline of inlet temperature based on the waste nitrogen inlet temperature time-series data. S312. Determine the thermal inertia time constant and desorption thermal buffer coefficient of the temperature sensor; S313. Based on the thermal inertia time constant of the temperature sensor, perform sensor hysteresis compensation on the measured temperature time series data, and based on the desorption thermal buffer coefficient, perform desorption thermal buffer compensation on the measured temperature time series data after sensor hysteresis compensation to obtain equivalent fluid temperature time series data. S314. Calculate the temperature change rate of the equivalent fluid temperature time series data relative to the inlet temperature dynamic baseline. When the temperature change rate exceeds the preset thermal wave decoupling threshold, determine the moment when the temperature change rate exceeds the preset thermal wave decoupling threshold as the initial temperature rise time point.

[0054] In this embodiment, specifically: As described in steps S311-S314 above, this scheme performs sensor hysteresis compensation and desorption thermal buffer compensation on the measured temperature time series data to obtain equivalent fluid temperature time series data, and then determines the initial temperature rise time point based on the comparison between the temperature change rate and the dynamic baseline.

[0055] Taking the above example scenario, the control system acquires time-series data of the waste nitrogen inlet temperature and fits a dynamic baseline for the inlet temperature based on this data. Specifically, the control system acquires time-series data of the waste nitrogen inlet temperature changing over time during the regeneration heating process, fits this data to obtain a trend line of the inlet temperature changing over time, i.e., the dynamic baseline for the inlet temperature. For example, when the inlet temperature gradually rises from room temperature to approximately 200℃, the fitted dynamic baseline for the inlet temperature is an upward curve, reflecting the upward trend of the inlet temperature itself. The reason for fitting the dynamic baseline for the inlet temperature is that the inlet temperature itself is not a sudden jump, but rather a gradual increase; the temperature rise observed at the measuring point includes both the temperature jump caused by the arrival of the heat wave front and the influence of the inlet temperature rise itself; by fitting the dynamic baseline for the inlet temperature, the influence of the upward trend of the inlet temperature itself can be subtracted in subsequent calculations, accurately identifying the temperature change caused by the arrival of the heat wave front.

[0056] After obtaining the dynamic baseline of the inlet temperature, the control system determines the thermal inertia time constant of the temperature sensor and the desorption thermal buffer coefficient. Specifically, the thermal inertia time constant of the temperature sensor... The thermal inertia of the temperature sensor itself, i.e., the degree of lag in the sensor's measurement following changes in the actual gas temperature, is characterized by this time constant. This time constant can be predetermined based on the sensor's model parameters or calibration test results. Desorption thermal buffer coefficient. This coefficient characterizes the buffering effect of adsorbent desorption endothermic heat on the bed temperature rise during regeneration heating. It can be pre-calculated based on the desorption heat parameters of the adsorbent material. The reason for determining these two parameters is that the temperature collected by the measured temperature sensor is affected by both the sensor's own thermal inertia and the slowed temperature rise due to adsorbent desorption endothermic heat. Without compensating for these two effects, the extracted initial temperature rise time point will lag behind the actual arrival time of the thermal wave front, leading to inaccurate calculations of subsequent timing deviations.

[0057] After determining the thermal inertia time constant and the desorption thermal buffer coefficient, the control system performs sensor hysteresis compensation on the measured temperature time series data based on the thermal inertia time constant of the temperature sensor, and performs desorption thermal buffer compensation on the measured temperature time series data after sensor hysteresis compensation based on the desorption thermal buffer coefficient, thus obtaining equivalent fluid temperature time series data. Specifically, the sensor hysteresis compensation adopts the method of first-order inertial element inversion: based on the sensor thermal inertia time constant... Measured temperature time series data Inversion to compensated temperature The compensation formula is: in, This represents the rate of change of the measured temperature over time. The principle behind this compensation formula is that the lag of the sensor measurement relative to the actual fluid temperature is related to the time constant. The hysteresis is proportional to the rate of temperature change. By superimposing this hysteresis on the measured value, the temperature after sensor hysteresis compensation can be obtained.

[0058] Based on sensor hysteresis compensation, the control system further performs desorption thermal buffer compensation: according to the desorption thermal buffer coefficient. The amount of temperature rise reduction caused by desorption endothermic heat is added to the temperature after sensor compensation to obtain the equivalent fluid temperature. The compensation formula is: in, This represents the rate of temperature change over time after sensor compensation. The principle behind this compensation is that desorption heat absorption consumes some heat, causing the bed temperature to rise more slowly than during pure convection heat transfer. The purpose of compensation is to eliminate the influence of desorption heat absorption and restore the temperature rise process of the fluid itself without desorption heat absorption.

[0059] Through the above two compensation steps, the control system obtains the equivalent fluid temperature time-series data. This time series data reflects the actual fluid temperature change after deducting the effects of sensor thermal inertia and desorption heat absorption.

[0060] After obtaining the equivalent fluid temperature time-series data, the control system calculates the rate of temperature change of the equivalent fluid temperature time-series data relative to the inlet temperature dynamic baseline. When the rate of temperature change exceeds a preset thermal wave decoupling threshold, that moment is determined as the initial temperature rise time point. Specifically, the control system calculates the equivalent fluid temperature time-series data... Relative to the dynamic baseline of inlet temperature Temperature change rate When the rate of temperature change exceeds the preset thermal wave decoupling threshold, it is considered that the thermal wave front has reached the predetermined measurement point, and this moment is determined as the initial temperature rise time point.

[0061] For example, under normal operating conditions, the rate of temperature change of the equivalent fluid temperature relative to the inlet dynamic baseline remains at a low level before the arrival of the heat wave. When the heat wave front reaches the outlet measuring point, the rate of temperature change suddenly increases and exceeds the preset heat wave decoupling threshold. The control system determines the moment when the rate of temperature change first exceeds the threshold as the initial temperature rise time point. The reason for using the comparison between the rate of temperature change and the threshold to identify the arrival time of the heat wave is that after sensor compensation and desorption compensation, the temperature rise characteristic caused by the arrival of the heat wave in the equivalent fluid temperature will be manifested in the form of a sudden increase in the rate of change. By setting a reasonable heat wave decoupling threshold, the inflection point of the heat wave arrival can be accurately identified in the background noise, avoiding misjudging the slow rise of the inlet temperature itself as the arrival of the heat wave.

[0062] By performing sensor hysteresis compensation and desorption thermal buffer compensation on the measured temperature time series data, and identifying the arrival time of the heat wave based on the comparison of the temperature change rate and the dynamic baseline, this scheme accurately extracts the time phase characteristics of the heat wave front arriving at the measuring point from the measured temperature, eliminating the interference of sensor thermal inertia and desorption heat absorption on the time phase extraction, and providing an accurate and reliable measured time reference for subsequent time series deviation calculation.

[0063] In one embodiment, step S40 specifically includes: S411. Obtain the outlet dew point data or outlet impurity concentration data of the adsorber in the previous adsorption cycle, and determine the initial adsorption load of the adsorbent in the adsorption bed based on the outlet dew point data or outlet impurity concentration data. S412. Based on the initial adsorption load, the theoretical heat wave arrival time is corrected to obtain the load-corrected theoretical arrival time. S413. Determine the desorption endothermic hysteresis compensation value based on the initial adsorption load, and use the desorption endothermic hysteresis compensation value to compensate for the initial temperature rise time point to obtain the equivalent measured arrival time. S414. Divide the difference between the equivalent measured arrival time and the load-corrected theoretical arrival time by the load-corrected theoretical arrival time to obtain the timing deviation.

[0064] In this embodiment, specifically: As described in steps S411-S414 above, this scheme determines the initial adsorption load based on the outlet dew point or impurity concentration data of the previous adsorption cycle, and performs load correction and desorption hysteresis compensation on the theoretical heat wave arrival time and the measured initial temperature rise time point, respectively, and finally calculates the normalized time series deviation.

[0065] Taking the above example scenario, the control system acquires the outlet dew point data or outlet impurity concentration data of the adsorber from the previous adsorption cycle, and determines the initial adsorption load of the adsorbent in the adsorption bed based on the outlet dew point data or outlet impurity concentration data. Specifically, the control system collects the curve of the change of dew point temperature or impurity concentration at the adsorber outlet over time during the previous adsorption cycle, and determines the level of outlet dew point or impurity concentration at the end of the adsorption cycle based on this curve; combined with the adsorption isotherm characteristics of the adsorbent, the amount of impurities adsorbed on the adsorbent in each adsorbent sublayer at the end of the adsorption cycle is calculated, i.e., the initial adsorption load. For example, at the end of the previous adsorption cycle, the outlet dew point temperature rises from the initial -70℃ to -50℃; based on the adsorption isotherm of the molecular sieve, the average adsorption load of the adsorbent in the molecular sieve layer can be calculated.

[0066] The reason for determining the initial adsorption load is that a certain amount of impurities such as water vapor and carbon dioxide have been adsorbed onto the adsorbent at the end of the previous adsorption cycle. During the subsequent regeneration heating process, these adsorbed impurities will desorb, absorbing heat and slowing down the advance speed of the heat wave front. The larger the initial adsorption load, the more heat is absorbed during desorption, and the more significant the impact on the heat wave advance speed. If the initial adsorption load is not considered, there will be deviations in the theoretical heat wave arrival time and the measured initial temperature rise time, leading to inaccurate calculations of subsequent timing deviations.

[0067] After determining the initial adsorption load, the control system corrects the theoretical heat wave arrival time based on the initial adsorption load, obtaining the load-corrected theoretical arrival time. Specifically, as the initial adsorption load increases, the desorption endothermic effect increases, the heat wave frontal propagation speed decreases, and the theoretical heat wave arrival time is correspondingly prolonged. Therefore, a correction time proportional to the initial adsorption load is superimposed on the original theoretical heat wave arrival time to obtain the load-corrected theoretical arrival time. For example, if the initial adsorption load is... The load-time correction factor is The theoretical arrival time of load correction is then... ;in The theoretical arrival time of the thermal wave is not considered in terms of load. The reason for correcting the theoretical arrival time of the thermal wave is that the aforementioned theoretical arrival time is calculated based on the adsorbent being unloaded or under standard load conditions; however, in the actual regeneration process, a certain amount of impurities have been adsorbed on the adsorbent, and the desorption endothermic effect will change the thermal wave propagation speed; by correcting the load, the theoretical arrival time is made closer to the actual working conditions.

[0068] While correcting the theoretical arrival time of the thermal wave, the control system determines the desorption endothermic lag compensation value based on the initial adsorption load. This compensation value is then used to compensate for the initial temperature rise time, yielding the equivalent measured arrival time. Specifically, since desorption endothermic heating also slows down the measured temperature rise at the measuring point, the initial temperature rise time lags behind the actual arrival time of the thermal wave front at the measuring point. The larger the initial adsorption load, the more desorption endothermic heating occurs, and the more severe the lag in the measured time becomes. Therefore, the desorption endothermic lag compensation value is determined based on the initial adsorption load. This compensation value is proportional to the initial adsorption load; subtracting this hysteresis compensation value from the measured initial temperature rise time point yields the equivalent measured arrival time. For example, in the above scenario, if the load-hysteresis coefficient is... The desorption endothermic hysteresis compensation value Equivalent measured arrival time ,in This represents the initial temperature rise time point. The reason for compensating for the initial temperature rise time point is that the measured temperature rise at the measuring point is affected by both the arrival of the thermal wave front and the desorption heat absorption. Desorption heat absorption slows down the temperature rise, causing the measured initial temperature rise time point to be later. By subtracting the desorption lag compensation value, the equivalent measured arrival time unaffected by desorption heat absorption is restored, allowing the measured time to be compared with the theoretical time on the same benchmark.

[0069] After obtaining the theoretical arrival time and equivalent measured arrival time of the load correction, the control system divides the difference between the equivalent measured arrival time and the theoretical arrival time of the load correction by the theoretical arrival time of the load correction to obtain the timing deviation. Specifically, the timing deviation... This timing deviation is a dimensionless relative deviation value. For example, in the above scenario, if the theoretical arrival time of the load correction is 30 minutes and the equivalent measured arrival time is 32 minutes, then the timing deviation is... =(32-30) / 30≈6.7%, which is a positive value, indicating that the heat wave propagation is delayed. The reason for using the normalized method of dividing the difference by the theoretical arrival time to calculate the timing deviation is that the operating conditions (such as waste nitrogen flow rate, inlet temperature, etc.) of different regeneration cycles may differ, resulting in different absolute values ​​of the theoretical heat wave arrival time; directly comparing the absolute time difference makes it difficult to make a unified anomaly judgment across cycles; the normalized timing deviation eliminates the influence of absolute time, making the deviations under different operating conditions comparable, which facilitates the subsequent setting of a unified judgment threshold.

[0070] By correcting the theoretical arrival time and compensating for the measured arrival time based on the initial adsorption load, and using the normalized relative deviation as the time series deviation, this scheme eliminates the interference of the difference in the initial adsorption load on the calculation of the time series deviation. This makes the time series deviation under different regeneration cycles and different operating conditions have a unified comparison benchmark, providing an accurate and consistent deviation quantification index for subsequent anomaly type determination.

[0071] In one embodiment, step S50, which involves determining the type of anomaly in the adsorption bed based on the direction of the timing deviation, specifically includes: S511. Determine the major category of the adsorption bed anomaly based on the direction and magnitude of the time deviation: when the time deviation is negative and its absolute value exceeds a first preset threshold, it is determined that there is a flow field anomaly; when the time deviation is positive and exceeds a second preset threshold, it is determined that there is a regeneration lag anomaly. S512. When a regeneration lag-type anomaly is determined, the power deviation between the actual output power of the electric heater and the set output power is obtained: when the power deviation is less than zero and the absolute value exceeds the third preset threshold, it is determined that the actual output power of the electric heater is insufficient; when the power deviation is greater than or equal to zero, it is determined that the adsorbent is clumped. S513. Obtain measured temperature time series data at at least two predetermined measurement points along the airflow direction, and calculate the thermal wave front thickness based on the measured temperature time series data at at least two predetermined measurement points; when the flow field is determined to be abnormal and the thermal wave front thickness is less than a preset thickness threshold, confirm the determination of flow field abnormality.

[0072] In this embodiment, specifically: As described in steps S511-S513 above, this scheme determines the major category of anomalies based on the direction and magnitude of the timing deviation, further subdivides the specific causes of regeneration hysteresis anomalies, and confirms the flow field anomalies by measuring the thickness of the thermal wave front.

[0073] Taking the above example scenario, the control system determines the major category of anomalies in the adsorption bed based on the direction and magnitude of the time-series deviation. Specifically, the control system classifies anomalies into two main categories based on the positive or negative direction and magnitude of the time-series deviation: flow field anomalies and regeneration lag anomalies. When the time-series deviation is negative and its absolute value exceeds a first preset threshold, a flow field anomaly is determined to exist; when the time-series deviation is positive and exceeds a second preset threshold, a regeneration lag anomaly is determined to exist. For example, the first preset threshold is set to -10%, and the second preset threshold is set to +10%. If the normalized time-series deviation is -15% (the measured heat wave arrival time is 15% earlier than the theoretical time) and its absolute value exceeds the first preset threshold, a flow field anomaly is determined to exist; if the normalized time-series deviation is +15% (the measured heat wave arrival time is 15% later than the theoretical time) and its absolute value exceeds the second preset threshold, a regeneration lag anomaly is determined to exist. The reason for first identifying the major anomaly categories is that the positive or negative direction of the time-series deviation corresponds to the change in the rate of thermal wave propagation, and the change in the thermal wave propagation rate is directly related to different types of anomalies within the bed: a negative time-series deviation indicates accelerated thermal wave propagation, corresponding to flow field anomalies; a positive time-series deviation indicates slowed thermal wave propagation, corresponding to regeneration hysteresis anomalies. By first classifying the major categories, the general direction of the anomalies can be quickly identified, and then the specific causes can be further subdivided within each major category.

[0074] After identifying a regeneration lag anomaly, the control system acquires the power deviation between the actual output power and the set output power of the electric heater, and further subdivides the specific cause of the regeneration lag anomaly based on the power deviation. Specifically, the control system acquires the actual output power and the set output power of the electric heater, and calculates the power deviation between the two. When the power deviation is less than zero and the absolute value exceeds a third preset threshold, it is determined that the actual output power of the electric heater is insufficient; when the power deviation is greater than or equal to zero, it is determined that the adsorbent is agglomerated. For example, if the set output power of the electric heater is 800kW, and the actual output power is only 600kW, the power deviation is -200kW, and the absolute value exceeds the third preset threshold, then it is determined that the actual output power of the electric heater is insufficient; if the actual output power is 800kW (consistent with the set value), and the power deviation is 0, then it is determined that the regeneration lag is caused by adsorbent agglomeration. The reason for further subdividing regeneration lag anomalies based on power deviation is that these anomalies can be caused by two different reasons: first, insufficient actual output power of the electric heater, resulting in insufficient heat input to the bed and slowed heat wave propagation; second, adsorbent agglomeration, leading to localized deterioration of bed heat transfer performance and obstruction of heat wave propagation. The intervention methods for these two reasons differ: insufficient heater power requires heater inspection; adsorbent agglomeration requires scheduled maintenance. By comparing power deviations, these two causes can be quickly distinguished.

[0075] After determining that the flow field is abnormal, the control system acquires measured temperature time-series data at at least two predetermined measuring points along the airflow direction, and calculates the thermal wave front thickness based on the measured temperature time-series data at these two predetermined measuring points. When the thermal wave front thickness is less than a preset thickness threshold, the determination of the flow field abnormality is confirmed. Specifically, the control system sets at least two predetermined measuring points near the bed outlet along the airflow direction, with a certain distance between each measuring point along the axial direction; it collects measured temperature time-series data at the two measuring points respectively, and extracts the moment when the temperature begins to rise and the moment when the temperature reaches a steady state from the temperature time-series data at each measuring point; based on the time span of the temperature rise and the thermal wave propagation velocity, the thickness of the thermal wave front is calculated. The thermal wave front thickness is calculated as follows: Thermal wave front thickness = Thermal wave propagation velocity × Temperature rise time span. Wherein, the temperature rise time span is the time taken for the temperature at the measuring point to rise from the start of the rise to a steady state. For example, if the distance between two adjacent measuring points is 0.3m; the time span for the temperature at the upstream measuring point to rise from the start to the steady state is 2 minutes; and the heat wave propagation speed is 0.2m / min, then the heat wave front thickness = 0.2m / min × 2min = 0.4m. If the preset thickness threshold is 0.5m, and the calculated front thickness of 0.4m is less than the preset thickness threshold, then the judgment of flow field anomaly is confirmed. The reason for confirming flow field anomaly through heat wave front thickness is that although a negative time deviation (accelerated heat wave propagation) mainly corresponds to flow field anomaly, it may also be caused by other accidental factors; while true flow field anomalies (such as airflow short-circuiting, flow deviation) not only cause accelerated heat wave propagation but also make the heat wave front steeper, that is, the front thickness thinner; by calculating the heat wave front thickness and comparing it with the preset thickness threshold, the judgment of flow field anomaly can be further confirmed, reducing the false alarm rate.

[0076] By employing a three-tiered judgment process—first classifying anomalies into broad categories, then subdividing them into specific causes, and finally confirming flow field anomalies—this approach enables rapid and accurate identification of adsorption bed anomaly types. This ensures the timeliness of anomaly judgment and improves the accuracy of judgment through power deviation comparison and array thickness confirmation, providing a reliable basis for subsequent targeted intervention and control.

[0077] In one embodiment, a molecular sieve regeneration safety control system based on thermal wave feedforward is provided, which corresponds to the molecular sieve regeneration safety control method based on thermal wave feedforward described in the above embodiments. The molecular sieve regeneration safety control system based on thermal wave feedforward includes: The data acquisition module is used to acquire the operating condition data of the adsorber in the regeneration state. The operating condition data includes the mass flow rate of waste nitrogen, the inlet temperature of waste nitrogen, and the actual output power of the electric heater. The time extrapolation module is used to construct a thermal wave extrapolation model based on the mass flow rate of waste nitrogen gas and the physical property parameters of the multilayer adsorbent in the adsorption bed to extrapolate the process of the thermal wave front advancing in the adsorption bed, and to use the thermal wave extrapolation model to generate the theoretical thermal wave arrival time of the thermal wave front reaching the predetermined measuring point position based on the inlet temperature of waste nitrogen gas. The temperature rise determination module is used to collect measured temperature time series data at predetermined measurement points and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. The deviation determination module is used to compare the initial temperature rise time with the theoretical heat wave arrival time to determine the timing deviation. The anomaly detection module is used to determine the anomaly type of the adsorption bed based on the direction of the timing deviation. Specifically, when the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. The intervention control module is used to perform intervention control according to the type of abnormality before the measured temperature reaches the preset safety threshold.

[0078] Specific limitations regarding the molecular sieve regeneration safety control system based on thermal wave feedforward can be found in the above description of the molecular sieve regeneration safety control method based on thermal wave feedforward, and will not be repeated here. Each module in the aforementioned molecular sieve regeneration safety control system based on thermal wave feedforward can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0079] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows. Figure 2 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database is used for data storage, data processing, and data analysis. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a molecular sieve regeneration safety control method based on thermal wave feedforward.

[0080] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a molecular sieve regeneration safety control method based on thermal wave feedforward.

[0081] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements a molecular sieve regeneration safety control method based on thermal wave feedforward.

[0082] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0083] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0084] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for safe control of molecular sieve regeneration based on thermal wave feedforward, characterized in that, Includes the following steps: The operating condition data of the adsorber in the regeneration state are obtained, including the mass flow rate of waste nitrogen, the inlet temperature of waste nitrogen, and the actual output power of the electric heater. Based on the mass flow rate of waste nitrogen and the physical properties of the multilayer adsorbent in the adsorption bed, a thermal wave simulation model is constructed to simulate the propagation process of the thermal wave front in the adsorption bed. The theoretical thermal wave arrival time of the thermal wave front to the predetermined measuring point is generated based on the inlet temperature of the waste nitrogen using the thermal wave simulation model. Collect measured temperature time series data at predetermined measurement points, and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. The timing deviation is determined by comparing the initial temperature rise time with the theoretical arrival time of the heat wave. The type of anomaly in the adsorption bed is determined based on the direction of the time-series deviation. Specifically, when the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. Before the measured temperature reaches the preset safety threshold, intervention control is performed according to the type of abnormality.

2. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 1, characterized in that, The step of constructing a thermal wave simulation model based on the mass flow rate of polluted nitrogen gas and the physical property parameters of the multilayer adsorbent in the adsorption bed to simulate the propagation process of the thermal wave front within the adsorption bed specifically includes: The adsorption bed is divided into multiple adsorbent sub-layers along the gas flow direction, and each adsorbent sub-layer has corresponding physical property parameters. Historical measured temperature time series data at predetermined measurement points in the historical regeneration cycle are obtained, and the physical property parameters of each adsorbent sublayer are corrected based on the historical measured temperature time series data to obtain the physical property correction parameters. Based on the physical property correction parameters corresponding to each adsorbent sublayer and the mass flow rate of the polluted nitrogen gas, unsteady heat transfer equations corresponding to each adsorbent sublayer are established respectively. Using the heat flow continuity condition and temperature continuity condition at the interface of adjacent adsorbent sublayers as boundary conditions, the various unsteady heat transfer equations are combined into a set of equations describing the propagation process of the thermal wave front within the adsorption bed, thus obtaining the thermal wave derivation model.

3. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 2, characterized in that, The step of acquiring historical measured temperature time-series data at predetermined measurement points during historical regeneration cycles, and correcting the physical property parameters of each adsorbent sublayer based on the historical measured temperature time-series data to obtain the corrected physical property parameters specifically includes: Based on the time series data of waste nitrogen inlet temperature in the historical regeneration cycle and the historical measured temperature time series data, the historical measured thermal wave propagation time from the inlet of the adsorption bed to the predetermined measuring point position was determined. Based on the physical properties of each adsorbent sublayer, historical operating data corresponding to historical regeneration cycles, and the relationship between thermal wave propagation rate and physical properties, the historical theoretical thermal wave propagation time from the adsorption bed inlet to the predetermined measuring point is estimated. By comparing the historically measured thermal wave propagation time with the historically theoretical thermal wave propagation time, the propagation time deviation is determined. Based on the correspondence between the propulsion time deviation and the thermal wave propulsion rate and the physical property parameters, the physical property parameters of each adsorbent sublayer are corrected to obtain the corrected physical property parameters.

4. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 2, characterized in that, The step of generating the theoretical arrival time of the thermal wave at the predetermined measuring point based on the inlet temperature of the polluted nitrogen gas using a thermal wave simulation model specifically includes: Using the inlet temperature of polluted nitrogen as the boundary condition, the equations obtained by combining the unsteady heat transfer equations corresponding to each adsorbent sublayer are solved to obtain the theoretical temperature rise time series data of each adsorbent sublayer. Based on the theoretical temperature rise time series data of each adsorbent sublayer, the theoretical propagation rate of the thermal wave front within each adsorbent sublayer is determined. Based on the loading geometric parameters of each adsorbent sublayer along the gas flow direction and the theoretical propulsion rate, the layer-by-layer propulsion time of the thermal wave front through each adsorbent sublayer is determined. By summing up the advancement times of each adsorbent sublayer, the theoretical arrival time of the thermal wave at the predetermined measuring point is obtained.

5. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 1, characterized in that, The step of collecting measured temperature time-series data at predetermined measurement points and determining the initial temperature rise time point based on the measured temperature time-series data and the waste nitrogen inlet temperature time-series data specifically includes: Acquire time-series data of waste nitrogen inlet temperature, and fit a dynamic baseline of inlet temperature based on the waste nitrogen inlet temperature time-series data; Determine the thermal inertia time constant and desorption thermal buffer coefficient of the temperature sensor; Based on the thermal inertia time constant of the temperature sensor, sensor hysteresis compensation is performed on the measured temperature time series data. Based on the desorption thermal buffer coefficient, desorption thermal buffer compensation is performed on the measured temperature time series data after sensor hysteresis compensation to obtain equivalent fluid temperature time series data. Calculate the rate of change of equivalent fluid temperature time series data relative to the inlet temperature dynamic baseline. When the rate of change of temperature exceeds a preset thermal wave decoupling threshold, the moment when the rate of change of temperature exceeds the preset thermal wave decoupling threshold is determined as the initial temperature rise time point.

6. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 1, characterized in that, The step of comparing the initial temperature rise time with the theoretical arrival time of the heat wave to determine the timing deviation specifically includes: Obtain the outlet dew point data or outlet impurity concentration data of the adsorber from the previous adsorption cycle, and determine the initial adsorption load of the adsorbent in the adsorption bed based on the outlet dew point data or outlet impurity concentration data. The theoretical heat wave arrival time is corrected based on the initial adsorption loading to obtain the loading-corrected theoretical arrival time. The desorption endothermic hysteresis compensation value is determined based on the initial adsorption load, and the initial temperature rise time point is compensated using the desorption endothermic hysteresis compensation value to obtain the equivalent measured arrival time. The timing deviation is obtained by dividing the difference between the equivalent measured arrival time and the load-corrected theoretical arrival time by the load-corrected theoretical arrival time.

7. The molecular sieve regeneration safety control method based on thermal wave feedforward as described in claim 1, characterized in that, The step of determining the anomaly type of the adsorption bed based on the direction of the time-series deviation specifically includes: The major category of anomalies in the adsorption bed is determined based on the direction and magnitude of the time-series deviation: when the time-series deviation is negative and its absolute value exceeds a first preset threshold, an anomaly in the flow field is determined to exist; when the time-series deviation is positive and exceeds a second preset threshold, an anomaly of regeneration lag is determined to exist. When a regeneration lag-type anomaly is identified, the power deviation between the actual output power and the set output power of the electric heater is obtained: when the power deviation is less than zero and the absolute value exceeds the third preset threshold, it is determined that the actual output power of the electric heater is insufficient; when the power deviation is greater than or equal to zero, it is determined that the adsorbent is clumped. Acquire measured temperature time-series data at at least two predetermined measurement points along the airflow direction, and calculate the thermal wave front thickness based on the measured temperature time-series data at at least two predetermined measurement points; when the flow field is determined to be abnormal and the thermal wave front thickness is less than a preset thickness threshold, confirm the determination of flow field abnormality.

8. A molecular sieve regeneration safety control system based on thermal wave feedforward, used to implement the steps of the molecular sieve regeneration safety control method based on thermal wave feedforward as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire the operating condition data of the adsorber in the regeneration state. The operating condition data includes the mass flow rate of waste nitrogen, the inlet temperature of waste nitrogen, and the actual output power of the electric heater. The time extrapolation module is used to construct a thermal wave extrapolation model based on the mass flow rate of waste nitrogen gas and the physical property parameters of the multilayer adsorbent in the adsorption bed to extrapolate the process of the thermal wave front advancing in the adsorption bed, and to use the thermal wave extrapolation model to generate the theoretical thermal wave arrival time of the thermal wave front reaching the predetermined measuring point position based on the inlet temperature of waste nitrogen gas. The temperature rise determination module is used to collect measured temperature time series data at predetermined measurement points and determine the initial temperature rise time point based on the measured temperature time series data and the waste nitrogen inlet temperature time series data. The deviation determination module is used to compare the initial temperature rise time with the theoretical heat wave arrival time to determine the timing deviation. The anomaly detection module is used to determine the anomaly type of the adsorption bed based on the direction of the timing deviation. Specifically, when the initial temperature rise time is earlier than the theoretical heat wave arrival time, it is determined that there is an abnormal flow field in the adsorption bed. When the initial temperature rise time is later than the theoretical heat wave arrival time, it is determined that there is adsorbent agglomeration and / or insufficient actual output power of the electric heater in the adsorption bed. The intervention control module is used to perform intervention control according to the type of abnormality before the measured temperature reaches the preset safety threshold.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the molecular sieve regeneration safety control method based on thermal wave feedforward as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the molecular sieve regeneration safety control method based on thermal wave feedforward as described in any one of claims 1-7.

Images