A method for switching control of a natural gas molecular sieve dehydration bed

Abstract

The application discloses a natural gas molecular sieve dehydration bed switching control method and relates to the technical field of dehydration bed control. The method comprises the following steps: at the end of the cooling stage after regeneration, a cooling gas valve is closed, and the temperature rebound amplitude of the outlet temperature after the cooling gas valve is closed is obtained; the residual moisture content in the bed layer is determined according to the temperature rebound amplitude; no supplementary cooling is performed, and the tower switching operation is directly allowed to be performed; and the residual moisture content is used as a feedforward correction amount to adjust the hot blowing parameters during the next regeneration of the adsorption tower, so that the residual moisture of the present time and the moisture to be removed normally next time can be removed together during the next regeneration. According to the application, the residual moisture of the bed layer is quantified by means of the temperature rebound amplitude after the cooling gas is closed, and the residual moisture is used as a feedforward correction amount to adjust the hot blowing parameters during the next regeneration, so that the moisture accumulation cycle by cycle is blocked without disturbing the double-tower switching time sequence.

Description

Technical Field

[0001] This invention belongs to the field of dehydration bed control technology, specifically a switching control method for natural gas molecular sieve dehydration beds. Background Technology

[0002] Natural gas dehydration is a crucial step in natural gas purification processes, and molecular sieve adsorption dehydration is widely used due to its high dehydration depth and stable operation. Industrial sites typically employ dual-tower or multi-tower switching methods to achieve continuous dehydration operations. After adsorption and regeneration, each tower enters a cooling phase to reduce the bed temperature to meet the requirements of subsequent adsorption operations.

[0003] Existing molecular sieve dehydration bed switching control methods mostly use fixed duration or fixed outlet temperature as the criteria for determining the end of the cooling stage. The tower switching action is executed directly after the cooling gas valve is closed, without further monitoring or assessment of the bed state after valve closure. This type of control method cannot accurately reflect the actual moisture content inside the bed, easily leading to over- or under-cooling. Over-cooling results in additional consumption of cooling gas resources, increasing the energy consumption and operating costs of the unit; under-cooling leaves residual moisture inside the bed, directly affecting the adsorption depth and dehydration effect of the next cycle.

[0004] Some existing technologies will initiate a supplementary cooling process when an abnormal bed temperature is detected. This method will prolong the working cycle of a single tower and disrupt the switching sequence of dual towers (e.g., the original 4-hour switching interval becomes 4.5 hours after supplementary cooling, affecting continuous processing capacity). Meanwhile, most existing control strategies only target closed-loop regulation in the current cycle and do not use the residual moisture generated during the current operation as feedforward information to introduce the regeneration control of the next cycle. This makes each regeneration process independent of each other, and it is impossible to specifically treat the moisture left over from the previous cycle. This can easily cause moisture to gradually accumulate inside the bed, which will aggravate the degradation of molecular sieve adsorption performance and shorten the service life of the adsorbent in the long run. It is also difficult to ensure that the natural gas outlet dew point is stable and meets the standards in the long term. Summary of the Invention

[0005] The purpose of this invention is to provide a method for switching control of natural gas molecular sieve dehydration beds to solve the problems mentioned in the background art.

[0006] A method for switching control of a natural gas molecular sieve dehydration bed includes: At the end of the cooling phase after regeneration, the cooling gas valve is closed, and the temperature rebound amplitude of the outlet temperature after the cooling gas valve is closed is obtained; The residual moisture content inside the bed is determined based on the temperature rebound amplitude. No supplemental cooling is performed; tower switching operation is allowed directly. The residual moisture content is used as a feedforward correction amount to adjust the hot blowing parameters during the next regeneration of the adsorption tower, so that the residual moisture from this regeneration can be removed together with the moisture that needs to be removed normally in the next regeneration.

[0007] By adopting the above scheme, the temperature rebound amplitude is obtained after the cooling gas is shut off, and the residual water content is determined accordingly. The tower switching is directly allowed without supplemental cooling, and the residual water content is used as a feedforward correction to adjust the parameters of the next regeneration hot blowing, thus achieving a unity of timing protection and feedforward compensation. This not only avoids disrupting the switching rhythm with supplemental cooling, but also prevents moisture accumulation through cross-cycle compensation, thereby improving the operating efficiency and dehydration stability of the device.

[0008] In some possible implementations, obtaining the temperature rebound amplitude includes: The outlet temperature at the moment of shutdown is recorded as the reference temperature. The outlet temperature within a preset time window after shutdown is continuously monitored. The difference between the highest temperature within the preset time window and the reference temperature is taken as the temperature rebound amplitude.

[0009] This scheme specifies the method for obtaining the temperature rebound amplitude: using the shutdown time as the baseline temperature, monitoring the highest temperature within a preset time window and calculating the difference. This method eliminates background interference such as ambient temperature and initial temperature, enabling the rebound amplitude to accurately reflect the temperature changes caused by residual moisture inside the bed, thus improving quantification accuracy and repeatability.

[0010] In some possible implementations, the residual moisture content is used as a feedforward correction factor to adjust the hot blowing parameters during the next regeneration of the adsorption tower, specifically including: Establish a mapping relationship between the rebound amplitude and the hot blowing time of the adsorption tower, where different rebound amplitude ranges correspond to different hot blowing time increments; When the temperature rebound falls within a certain range at the end of this cooling cycle, the corresponding hot blowing time increment is obtained according to the mapping relationship, and the hot blowing time for the next regeneration is set as the base hot blowing time plus the hot blowing time increment.

[0011] Specifically, by establishing a mapping relationship between rebound amplitude and hot blowing time, different rebound amplitude ranges correspond to different hot blowing time increments, enabling precise matching between regeneration intensity and residual moisture content. This interval-based design simplifies the PLC's operational logic, balancing control accuracy and engineering practicality, and avoiding the real-time burden of continuous function calculations.

[0012] In some possible implementations, the mapping relationship is dynamically updated during the operation of the adsorption tower: Record the cooling rebound amplitude after each regeneration and the hot blowing time used in this regeneration; When the cooling rebound amplitude corresponding to a certain rebound amplitude range is greater than the maximum value of the historical rebound amplitude sequence of that range in three or more consecutive switching cycles, the hot blowing time increment corresponding to that range is automatically increased; when the cooling rebound amplitude is less than the minimum value of the historical rebound amplitude sequence of that range in three or more consecutive switching cycles, the hot blowing time increment corresponding to that range is automatically decreased.

[0013] Specifically, by recording the cooling rebound amplitude and hot blowing time after each regeneration, and automatically increasing or decreasing the hot blowing time increment based on the deviation of the rebound amplitude from the historical extreme value in multiple consecutive cycles, dynamic self-learning of the mapping relationship is achieved. The system can adapt to changes in operating conditions such as molecular sieve aging and feed gas fluctuations, and always maintain optimal matching of regeneration parameters.

[0014] In some possible implementations, the method is applied to a rotating dehydration apparatus comprising at least a first adsorption tower and a second adsorption tower: The first adsorption tower and the second adsorption tower independently maintain their respective historical rebound data and hot blowing parameter adjustment mapping relationship; When the first adsorption tower performs a switch, only the historical data of the first adsorption tower itself is used to adjust the hot blowing parameters for its next regeneration; when the second adsorption tower performs a switch, only the historical data of the second adsorption tower itself is used.

[0015] The method was applied to alternating dehydration units containing at least two towers, and historical rebound data and mapping relationships were maintained independently for each tower. Since different adsorption towers differ in runtime, load, and aging degree, independent maintenance avoided parameter mismatches caused by data mixing, improving the specificity and reliability of regeneration control for each tower.

[0016] In some possible implementations, it also includes: Obtain the outlet dew point curve of the current adsorption tower in the next adsorption cycle, and compare the outlet dew point curve with the temperature rebound amplitude obtained this time; When the outlet dew point curve rises at the beginning of the adsorption cycle and the temperature rebound amplitude is not greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the molecular sieve is aging and a molecular sieve replacement warning signal is generated. When the outlet dew point curve shows a sudden increase in the second half of the adsorption cycle, and the temperature rebound amplitude is greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the residual moisture release is caused by incomplete regeneration, and the hot blowing time increment for the next regeneration is increased by a preset fixed step.

[0017] By comparing the outlet dew point curve with the current temperature rebound magnitude, and based on the time period of the dew point rise (initial stage / second half) and whether the rebound magnitude exceeds the historical maximum value, two fault modes—molecular sieve aging and incomplete regeneration—are accurately distinguished. When incomplete regeneration is determined, the time increment for the next hot blowing is automatically increased, forming a closed-loop correction and reducing the risk of dew point exceeding the standard.

[0018] In some possible implementations, when molecular sieve aging is determined, the following further steps are included: Obtain the historical switching cycle record of the current adsorption tower, the historical switching cycle record including the actual effective adsorption time of the next adsorption cycle after each switching; The correlation analysis was performed between the historical temperature rebound amplitude sequence and the historical sequence of the actual effective adsorption time; If, in three or more consecutive switching cycles, the actual effective adsorption time under the temperature rebound amplitude within the same preset rebound amplitude range is less than the minimum historical effective adsorption time corresponding to that rebound amplitude range, it is predicted that the current adsorption tower will soon fail, and a molecular sieve replacement prompt will be generated.

[0019] Based on the determination of aging, historical switching cycle records are further obtained, and the historical temperature rebound amplitude sequence is correlated with the actual effective adsorption time sequence. When the effective adsorption time under the same rebound amplitude in multiple consecutive cycles is less than the historical minimum, the molecular sieve is predicted to fail soon. This trend prediction method eliminates the interference of single operating condition fluctuations, realizing an upgraded prompt from "early warning" to "pre-failure", supporting planned replacement.

[0020] In some possible implementations, when impending failure is predicted, at least one of the following operations is performed automatically: Reduce the processing gas volume of the dehydration unit where the current adsorption tower is located; Increase the regeneration temperature of the current adsorption tower or extend the regeneration time; The failure risk ranking is determined based on the order of the predicted failure times of each tower. The rotation sequence is then adjusted so that the adsorption towers with higher failure risk are scheduled to operate during periods with better gas and gas conditions, while the adsorption towers with lower failure risk are scheduled to operate during periods with worse gas and gas conditions.

[0021] When impending failure is predicted, the system automatically performs at least one of the following actions: reducing the processing gas volume, increasing the regeneration temperature / extending the regeneration time, and optimizing the rotation sequence based on failure risk ranking. These self-healing control measures can delay dew point deterioration, extend the safe operating window, buy time for molecular sieve replacement, and avoid unplanned downtime.

[0022] In some possible implementations, the method is applied to a rotating dehydration apparatus comprising at least a first adsorption tower and a second adsorption tower: Each adsorption tower independently maintains its historical temperature rebound amplitude sequence and actual effective adsorption duration sequence to independently determine the failure risk ranking of each tower; The rotation sequence is dynamically generated based on the failure risk ranking of each tower, so that the adsorption towers that are at the front of the failure risk ranking appear less often in the rotation sequence than the adsorption towers that are at the back of the failure risk ranking. If an adsorption tower fails to reduce the cooling rebound amplitude to below the maximum value of the most recent ten cycles in the tower's historical rebound amplitude sequence after applying its maximum allowed hot blowing time increment for more than three consecutive switching cycles, the tower will be marked as unusable, automatically removed from the rotation sequence, and a replacement instruction will be generated.

[0023] In the multi-tower rotation system, historical data is maintained independently for each tower, and its failure risk is independently assessed and ranked. The rotation sequence is dynamically generated based on this ranking, ensuring that high-risk towers occur less frequently than low-risk towers. Simultaneously, if a tower fails to reduce its rebound amplitude below the historical maximum for several consecutive cycles under normal hot blowing parameters, it is automatically marked as unusable and removed from the rotation sequence. This mechanism effectively reduces the risk of operating with defects and ensures the overall dehydration effect of the system.

[0024] In some possible implementations, after any adsorption tower is marked as unavailable and removed from the rotation sequence, the following steps are also included: Automatically identify the number of remaining adsorption towers in the current rotation sequence; The adsorption cycle duration of each remaining adsorption tower is dynamically adjusted based on the number of remaining adsorption towers to maintain the continuous operation capability of the device; wherein, the adjustment method for the adsorption cycle duration is that the adsorption cycle duration is inversely proportional to the number of remaining adsorption towers; Furthermore, when the number of remaining adsorption towers falls below the minimum number of continuous operations, the system automatically generates a load reduction operation prompt and recommends the maximum allowable gas throughput.

[0025] It should be noted that when any adsorption tower is removed, the system automatically identifies the number of remaining towers and dynamically adjusts the adsorption cycle duration of each remaining tower in an inverse proportional relationship to maintain the total water treatment capacity consistent with that before removal. When the number of remaining towers falls below the minimum continuous operating quantity, a load reduction prompt is automatically generated, and the maximum allowable gas throughput is recommended. This technical solution achieves seamless load redistribution after a single tower failure, avoiding significant production fluctuations and preventing the remaining towers from operating under overload conditions, thus improving the system's fault tolerance.

[0026] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention actively abandons supplemental cooling, uses the temperature rebound amplitude after the cooling gas is shut off to quantify the residual moisture in the bed, and uses it as a feedforward correction to adjust the parameters of the next regeneration hot blowing, thereby preventing the accumulation of moisture cycle by cycle without disrupting the switching sequence of the two towers. Simultaneously, by combining the outlet dew point curve and historical rebound data, the aging of molecular sieves and incomplete regeneration can be accurately distinguished, enabling failure trend prediction and self-healing control. Furthermore, the rotation sequence and load distribution can be dynamically optimized during multi-tower rotation, thereby improving the operating efficiency, dehydration stability, self-adaptability, and lifespan of the dewatering unit. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method structure of the present invention. Detailed Implementation

[0028] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figure 1 This application provides a method for switching control of a natural gas molecular sieve dehydration bed, comprising: The method is applied to a dehydration device that includes at least a first adsorption tower and a second adsorption tower: by limiting the application scenario to at least two towers, it can provide a hardware foundation for continuous switching control and independent status monitoring, avoid the problem that single-tower operation cannot achieve continuous dehydration, and at the same time provide structural support for subsequent independent data maintenance and differentiated scheduling of multiple towers, ensuring continuous operation of the system while improving the pertinence of the control strategy.

[0030] It should be noted that the method is specifically applicable to dual-tower and multi-tower rotation systems because multi-tower systems have differences in the state between towers (such as different molecular sieve aging degrees, adsorption loads, and operating times). The design of subsequent independent maintenance data and independent parameter adjustment is based on this application scenario to ensure the relevance and practicality of the control strategy.

[0031] The technical problem to be solved by this invention is to overcome the problems of inaccurate determination of the end of the cooling stage in the switching control of existing molecular sieve dewatering beds, which leads to over- or under-cooling, residual moisture accumulation affecting the dewatering effect, and supplementary cooling disrupting the switching sequence. At the same time, it aims to achieve accurate determination and targeted treatment of molecular sieve aging, impending failure, and incomplete regeneration.

[0032] To achieve the above objectives, the present invention utilizes the following physical principle: after the cooling stage of the molecular sieve bed is completed, at the moment the cooling gas valve is closed, a certain amount of heat and adsorbed moisture still remain inside the bed.

[0033] Heat is conducted from the inside of the bed to the outlet, causing a short-term rebound in the outlet temperature. The magnitude of this rebound is directly related to the amount of residual moisture inside the bed. The more residual moisture, the more heat is absorbed during desorption, resulting in more heat storage inside the bed after the cooling phase and a greater temperature rebound after the valve is closed. Meanwhile, the regeneration effect of the molecular sieve is not solely determined by the hot blowing parameters of a single cycle. Residual moisture from the previous cycle that has not been desorbed will occupy some adsorption sites in subsequent adsorption processes. If the regeneration parameters are not adjusted accordingly, moisture will gradually accumulate, thus reducing the dehydration depth. Molecular sieve aging leads to a permanent decrease in its adsorption capacity, manifested as an increase in dew point at the beginning of the adsorption cycle, and a continuous shortening of the effective adsorption time under the same residual moisture conditions, until complete failure. Feedforward correction combined with dew point curve monitoring and historical data correlation analysis can achieve the triple objectives of residual moisture removal, equipment status monitoring, and failure prediction.

[0034] Based on the above physical laws, the present invention adopts the following control logic: the residual water content is indirectly quantified by the temperature rebound amplitude, eliminating the need for supplementary cooling process to ensure switching timing; the residual water content is used as a feedforward correction amount to adjust the hot blowing parameters of the next cycle. By combining the outlet dew point curve of the next adsorption cycle with historical rebound amplitude data, the aging or incomplete regeneration state of the molecular sieve is determined, and early warning or secondary adjustment of hot blowing parameters is executed accordingly. For the aging state, historical switching cycle data is further correlated to predict impending failure and generate replacement prompts. At the same time, control strategies such as load adjustment, regeneration enhancement, and rotation sequence optimization are automatically executed to form a complete control chain of "state monitoring - timing guarantee - feedforward correction - fault determination - failure prediction - closed-loop optimization", taking into account processing efficiency, dehydration effect and equipment operation and maintenance needs.

[0035] By combining timing assurance with feedforward correction, the accumulation of residual moisture can be blocked from the source without disrupting the switching rhythm of the two towers. This ensures the continuous and stable operation of the device and avoids the degradation of molecular sieve performance and excessive dew point caused by moisture accumulation.

[0036] The first and second adsorption towers are maintained independently with their own historical rebound data and hot blowing parameter adjustment mapping relationships. By maintaining historical data and mapping relationships independently for each tower, data interference between towers with different aging levels and different operating loads can be avoided, ensuring that the hot blowing parameter adjustment of each tower is in line with its own actual state, improving the regeneration control accuracy, and fundamentally avoiding the problem of insufficient or excessive regeneration caused by shared data.

[0037] It should be noted that the historical rebound data includes the temperature rebound amplitude for each cycle (unit: °C, precision retained to one decimal place), the corresponding hot blowing time increment (unit: minutes), and regeneration effect feedback (such as subsequent dew point data, actual effective adsorption time). The mapping relationship of the hot blowing parameter adjustment is the correspondence rule between the rebound amplitude range and the hot blowing time increment. The specific implementation method for independent maintenance data of the two towers is as follows: The PLC controller allocates an independent register address range for each adsorption tower (the first adsorption tower occupies register addresses 0x0000-0x0FFF, and the second adsorption tower occupies 0x1000-0x1FFF). The data is stored in a structured format of "cycle number-temperature rebound amplitude-hot blowing time increment-dew point curve characteristics-actual effective adsorption duration". Each cycle of data occupies 16 bytes of storage space, retains complete data of no less than 50 recent cycles, and does not mix or overwrite data from other towers, ensuring that the control strategy of each tower is consistent with its actual operating status.

[0038] At the end of the cooling stage after regeneration, the cooling gas valve is closed, and the temperature rebound amplitude of the outlet temperature after the cooling gas valve is closed is obtained. By collecting characteristic signals at a fixed node at the end of the cooling stage, data distortion caused by inconsistent collection timing can be avoided, providing stable and reliable raw data for subsequent residual moisture quantification and ensuring the accuracy of bed condition judgment.

[0039] It should be noted that the core purpose of the cooling stage is to cool the regenerated high-temperature bed to the adsorption operating temperature (usually close to the temperature of the feed gas, set at 35℃-55℃). The criterion for the end of cooling is that the bed outlet temperature remains stable in the range of 40℃ to 50℃ for 3 consecutive minutes (fluctuation range ≤ ±1℃, based on the common molecular sieve adsorption activity temperature range).

[0040] After the cooling gas valve is closed, the latent heat (from the undesorbed moisture) remaining inside the bed and the sensible heat work together to cause the outlet temperature to rise briefly. This temperature change characteristic is directly related to the residual moisture content. Therefore, the residual moisture can be indirectly quantified by capturing the temperature rebound amplitude.

[0041] The cooling gas valve adopts an industrial-grade pneumatic shut-off valve (model: DN50-250, nominal pressure 1.6MPa). Its closing signal is output by the PLC controller (model: S7-1500), and the closing action response time is no more than 0.5 seconds, ensuring the accuracy of the starting point of temperature monitoring. The outlet temperature is collected by a PT100 temperature transmitter (measurement range: -50℃-200℃, measurement accuracy ±0.1℃) installed in the bed outlet pipeline within 1 meter of the bed outlet end. The acquisition frequency is 1 second / time to ensure the reliability of the temperature data.

[0042] The temperature rebound amplitude is obtained by: recording the outlet temperature at the moment of shutdown as the reference temperature, continuously monitoring the outlet temperature within a preset time window after shutdown, and taking the difference between the highest temperature within the preset time window and the reference temperature as the temperature rebound amplitude. The temperature rebound amplitude is calculated by the difference between the baseline temperature and the highest temperature within a preset time window. This method can eliminate background interference such as ambient temperature and initial temperature, and retain only the temperature change caused by residual moisture inside the bed. This improves the accuracy and consistency of the rebound amplitude in characterizing residual moisture, and provides a precise basis for the quantification of residual moisture.

[0043] It should be noted that the preset time window is set to 3 to 5 minutes after the valve is closed (the default value is 4 minutes, which can be adjusted through the PLC human-machine interface). This duration has been verified through multiple sets of field tests: if it is less than 3 minutes, the temperature peak may not be captured (when there is more residual moisture, the latent heat release lasts longer), and if it is more than 5 minutes, the heat dissipation of the bed layer will dominate, and the temperature will drop, resulting in a smaller calculated rebound amplitude.

[0044] The data acquisition frequency is set to 1 second / time. The PLC controller records temperature data in real time and automatically filters peak values. The filtering logic is as follows: after removing outliers that deviate from the average of 10 adjacent data points by ±2℃, the maximum value among the remaining data is taken to ensure the accuracy of the rebound amplitude calculation. By subtracting the peak temperature from the reference temperature (temperature at the moment the valve is closed), external interference such as ambient temperature fluctuations and pipeline heat dissipation can be eliminated, retaining only the temperature changes caused by the internal state of the bed.

[0045] The residual moisture content inside the bed is determined based on the temperature rebound amplitude. By converting the temperature rebound amplitude into a quantified residual moisture content, the bed state can be transformed from a qualitative description to a quantitative value, providing a calculable and comparable quantitative basis for subsequent feedforward correction, achieving precise matching between regeneration intensity and residual moisture content, and avoiding blind adjustment of regeneration parameters.

[0046] It should be noted that the residual moisture content is defined as "the total mass of moisture that remains adsorbed inside the micropores of the molecular sieve after the cooling stage has ended and has not been purged and desorbed by the cooling gas", with the unit being kilograms per cubic meter (kg / m³).

[0047] The core of this step is to establish a quantitative mapping relationship between the temperature rebound amplitude and the residual moisture content. To provide verifiable technical evidence, calibration tests were conducted on type 3A molecular sieves (12m³ packing volume, 1.6mm diameter spherical particles) under the following typical operating conditions: Inlet pressure 1.2MPa±0.05MPa, processing gas volume 3000m³ / h±5%, regeneration temperature 240℃±2℃, cooling gas is dry natural gas (water dew point ≤-60℃).

[0048] The test method is as follows: After the cooling stage is over, close the cooling gas valve, record the outlet temperature T0 at the moment of closure and the highest temperature Tmax in the following 4 minutes, and calculate the rebound amplitude ΔT=Tmax-T0; Subsequently, molecular sieve samples were immediately collected from the upper, middle and lower sampling ports of the bed (200g sample from each layer). The residual moisture content was determined by Karl Fischer method (the sample was transferred from the sampling port to the sealed bottle under nitrogen protection to avoid interference from moisture in the air; each sample was measured in parallel 3 times and the average value was taken; the measurement accuracy was ±0.02kg / m³). The average value of the three layers was taken as the residual moisture content corresponding to ΔT.

[0049] The above experiment was repeated 30 times (covering the range of ΔT = 0.5℃ to 8.5℃), resulting in 30 sets of data points. Linear regression analysis showed a correlation coefficient R² of 0.94 between ΔT and residual moisture content W, with the regression equation being: W = 0.243 × ΔT + 0.08 (unit: kg / m³). The 95% confidence interval for the slope of this equation is [0.228, 0.258]. Based on the above experimental results, this invention sets the default slope to 0.25 kg / m³・℃ and allows adjustment via the controller parameter table within the range of 0.23 to 0.27 kg / m³・℃ based on different molecular sieve models (3A / 4A / 5A) and packing densities.

[0050] If the user cannot calibrate the data themselves, they can directly use the regression equation disclosed in this publication for calculation. The prediction error will not exceed ±0.12 kg / m³ when ΔT ≤ 8℃. The specific calculation method is automatically executed by the PLC controller: it calls the pre-stored linear mapping data table (stored in the controller register in the form of a two-dimensional array, where the array index is the integer part of the rebound amplitude and the array value is the corresponding basic residual moisture content). If the measured rebound amplitude is an integer (such as 3℃), it directly matches the corresponding residual moisture content (3 × 0.25 = 0.75 kg / m³). If the value is not an integer (e.g., 2.3℃), linear interpolation is used for calculation: Residual moisture content = Base moisture content + (Actual rebound amplitude - Base rebound amplitude) × Unit amplitude moisture increment (e.g., 2℃ corresponds to 0.5 kg / m³, 2.3℃ corresponds to 0.5 + 0.3 × 0.25 = 0.575 kg / m³). The calculation result is rounded to two decimal places to ensure quantitative accuracy. Different types of molecular sieves (e.g., 3A, 4A, 5A) have different specific surface areas and microporous structures. The unit amplitude moisture increment can be adjusted within the range of 0.23 to 0.27 kg / m³・℃ (based on the above confidence interval) without changing the overall control logic, thus improving the versatility of the method.

[0051] By eliminating supplemental cooling, tower switching operations can be performed directly. By canceling the supplemental cooling process, the operating cycle of a single tower can be shortened, the timing of dual-tower switching can be disrupted, and the ineffective consumption of cooling gas can be reduced, thereby lowering the energy consumption of the unit and improving the continuous operating efficiency and economy of the system without affecting the dehydration effect.

[0052] It should be noted that in traditional methods, to "ensure" sufficient bed cooling, an additional 10 to 30 minutes of supplementary cooling is added after the main cooling stage. This prolongs the single-tower cycle and disrupts the switching sequence between the two towers (e.g., a switchover originally occurring every 4 hours becomes 4.5 hours after supplementary cooling, affecting continuous processing capacity). This method eliminates supplementary cooling because the residual moisture state is accurately controlled through temperature rebound amplitude, and subsequent regeneration parameters can be adjusted via feedforward correction. Therefore, it is not necessary to rely on supplementary cooling to eliminate the influence of residual moisture, ensuring stable switching sequence and reducing ineffective consumption of cooling gas (the cooling gas flow rate during the supplementary cooling stage is consistent with the main cooling stage, accounting for approximately 15% of energy consumption).

[0053] Tower switching implementation: In a single-tower independent operation or dual-tower simple switching system, after the PLC controller completes the residual water content calculation, it outputs a high-level allow signal (voltage 24VDC) to the tower switching actuator (pneumatic valve group) after a 0.5-second delay. The actuator starts the switching in the order of "first opening the inlet and outlet valves of the standby tower, then balancing the pressure for 30 seconds, and finally closing the inlet and outlet valves of the current tower". The total switching time does not exceed 30 seconds to ensure continuous operation of the device. In a multi-tower (three towers and above) interlocking system, the allow signal is first transmitted to the central control system (DCS) and logically ANDed with the adsorption ready signal and pressure balance signal (pressure difference ≤ 0.05MPa) of other towers. The switching is triggered after all interlocking conditions are met to avoid system pressure fluctuations caused by single-tower switching (pressure fluctuation range controlled within ±0.05MPa).

[0054] Extreme operating condition handling: When the temperature rebound exceeds the normal calibration range (e.g., exceeding 8 degrees Celsius, corresponding to a residual moisture content exceeding 2 kg / m³), the system determines that the residual moisture content of the bed is abnormally high, but does not trigger supplementary cooling. Instead, it records the abnormal data (baseline temperature, maximum temperature, rebound amplitude, residual moisture content, operating parameters) through the controller's built-in log module. The recording format is "timestamp-tower number-parameter name-value". In the next regeneration cycle, it automatically activates the maximum hot blowing time increment (e.g., 8 minutes + preset fixed step size 2 minutes = 10 minutes) to completely remove residual moisture through feedforward correction, avoiding disruption of the overall timing due to supplementary cooling.

[0055] The residual moisture content is used as a feedforward correction to adjust the hot blowing parameters during the next regeneration of the adsorption tower, so that the residual moisture from this regeneration can be removed together with the moisture that needs to be removed normally in the next regeneration. By compensating for residual moisture in advance through feedforward correction, the accumulation of moisture over cycle can be avoided, the risk of exceeding the dew point standard can be suppressed from the source, the regeneration process can be optimized in a closed loop, and the long-term stability of the molecular sieve adsorption performance can be ensured.

[0056] It should be noted that the feedforward correction amount, which is the increase in hot blowing time corresponding to the residual moisture, is based on the core logic of "pre-compensation". The residual moisture needs to be removed by adding extra hot blowing time during the next regeneration to ensure that the bed is in the best condition in the next adsorption cycle.

[0057] The hot blowing parameters mainly refer to the hot blowing time (the hot blowing temperature is usually fixed at 200 to 280 degrees Celsius, based on the molecular sieve regeneration temperature requirements, with a default value of 240℃, which can be adjusted via DCS). Adjusting the hot blowing time is the most direct and easy-to-operate way to regulate the regeneration intensity, without changing the heat source supply parameters (such as steam pressure and electric heating power), and it has strong compatibility.

[0058] The residual water content is used as a feedforward correction to adjust the hot blowing parameters during the next regeneration of the adsorption tower. Specifically, this includes: establishing a mapping relationship between the rebound amplitude and the hot blowing time of the adsorption tower, where different rebound amplitude ranges correspond to different hot blowing time increments; when the temperature rebound amplitude at the end of this cooling process falls into a certain range, the corresponding hot blowing time increment is obtained according to the mapping relationship, and the hot blowing time for the next regeneration is set as the base hot blowing time plus the hot blowing time increment. By implementing segmented mapping to achieve graded adjustment of hot blowing time, the calculation logic can be simplified while ensuring control accuracy, adapting to the real-time control requirements of industrial PLCs, and enabling precise matching of regeneration intensity and residual moisture content, thus balancing control effectiveness and engineering practicality.

[0059] It should be noted that the baseline hot blowing time is preset based on the natural gas processing capacity, inlet moisture content, and molecular sieve loading: For a processing capacity of 1000 to 3000 m³ / h, inlet moisture content of 10 to 20 g / m³, and loading of 5 to 10 m³, the baseline hot blowing time is 4 hours; for a processing capacity of 3000 to 5000 m³ / h, inlet moisture content of 20 to 30 g / m³, and loading of 10 to 15 m³, the baseline hot blowing time is 6 hours; and for a processing capacity of 5000 to 8000 m³ / h, inlet moisture content of 30 to 50 g / m³, and loading of 15 to 25 m³, the baseline hot blowing time is 8 hours. The baseline hot blowing time is stored in the controller parameter area and can be modified on-site via the human-machine interface.

[0060] The relationship between rebound amplitude and hot blowing time is stored in the PLC in the form of interval segments, and the specific correspondence is as follows (verified by on-site dehydration conditions): The mapping relationship is determined based on the relationship between residual moisture content W and rebound amplitude ΔT: W = 0.25 × ΔT (kg / m³). Combined with a bed volume of 12 m³ and an average moisture removal rate of 0.12 kg / min during the hot blowing stage, the theoretical hot blowing time increment Δt = (W × 12) / 0.12 = 100 × W = 25 × ΔT (minutes) is derived. After rounding, the following table is obtained: Temperature rebound range of 0 to 2 degrees Celsius (corresponding to residual moisture content of 0 to 0.5 kg / m³): hot blowing time increment of 2 minutes; Temperature rebound range of 2 to 4 degrees Celsius (corresponding to residual moisture content of 0.5 to 1.0 kg / m³): hot blowing time increment of 4 minutes; Temperature rebound range of 4 to 6 degrees Celsius (corresponding to residual moisture content of 1.0 to 1.5 kg / m³): hot blowing time increment of 6 minutes; Temperature rebound of more than 6 degrees Celsius (corresponding to a residual moisture content of more than 1.5 kg / m³): Increase the hot blowing time by 8 minutes.

[0061] The reason for using segmented mapping instead of continuous calculation is that in actual industrial scenarios, the adjustment of hot blowing time does not need to be infinitely precise. The segmented design can simplify the PLC operation logic (reduce the number of floating-point operations) and reduce the difficulty of program implementation. At the same time, by reasonably dividing the intervals, it can be ensured that the increment of hot blowing time matches the requirement of residual moisture removal (error ≤ 0.1kg / m³).

[0062] When the first adsorption tower switches over, only its own historical data is used to adjust the hot blowing parameters for the next regeneration. When the second adsorption tower switches over, only its own historical data is used. By using only its own historical data when switching towers, mismatch of control parameters caused by differences in aging and operating conditions between different towers can be avoided. This further improves the specificity and reliability of the regeneration control for each tower, ensuring that the aging towers are fully regenerated and the new towers are protected from over-regeneration.

[0063] It should be noted that the historical data of the first adsorption tower includes the temperature rebound amplitude of its past ten cycles, the corresponding hot blowing time increment, dew point feedback data (peak value and stable value of the dew point curve for each cycle), and the actual effective adsorption time. The same applies to the second adsorption tower.

[0064] The logic for retrieving historical data is as follows: During switching, the controller extracts the complete data of the most recent ten cycles from the independent register address segment of the current tower number (distinguished by hardware address or software identifier), without calling any data from other towers.

[0065] The reason for using only its own data during switching is that the two towers may have different operating conditions (e.g., the first tower has run for 1000 cycles and the molecular sieve is more aged; the second tower has run for 500 cycles and is in a newer state). Independent data access avoids parameter mismatches caused by "using one tower for another". For example, if the two towers share data, the hot blowing parameters of the newer second tower may be applied to the aged first tower, resulting in insufficient regeneration in the first tower; conversely, it may cause over-regeneration in the second tower. Independent control ensures that the regeneration parameter adjustments of each tower are based on its own historical state, improving control accuracy (parameter matching error ≤ 5%).

[0066] Each adsorption tower independently maintains its historical temperature rebound amplitude sequence and actual effective adsorption duration sequence to independently determine the failure risk ranking of each tower; a rotation order is dynamically generated based on the failure risk ranking of each tower, so that the adsorption towers at the front of the failure risk ranking appear less frequently in the rotation sequence than the adsorption towers at the back of the failure risk ranking. Record the cooling rebound amplitude after each regeneration and the hot blowing time used in this regeneration. When the cooling rebound amplitude corresponding to a certain rebound amplitude range in three or more consecutive switching cycles is greater than the maximum value of the historical rebound amplitude sequence of that range, increase the hot blowing time increment of that range by 1 minute (maximum of 3 times for the same range); when it is less than the minimum value of the historical rebound amplitude sequence of that range, decrease the hot blowing time increment by 1 minute (the decrease shall not be less than 0 minutes, and the decrease shall be maximum of 2 times for the same range). If an adsorption tower, after applying its maximum permissible hot blowing time increment, fails to reduce the cooling rebound amplitude to below the maximum value of the most recent ten cycles in its historical rebound amplitude sequence for three or more consecutive switching cycles, the tower is marked as unusable, automatically removed from the rotation sequence, and a replacement instruction is generated. By independently maintaining the historical data of each tower and dynamically adjusting the rotation order, the load on high-risk towers can be reduced to delay performance degradation. Simultaneously, automatically disconnecting failed towers prevents operation with defects, ensuring overall system stability and reducing the risk of dew point exceeding limits.

[0067] After any adsorption tower completes a molecular sieve replacement, the historical temperature rebound amplitude sequence and historical effective adsorption duration sequence of that tower should be reset (clearing the original data and re-recording from the first cycle after the replacement) to avoid old data interfering with the control parameters of the new molecular sieve. The reset operation is triggered via the PLC human-machine interface after manual confirmation. After the reset, the rebound amplitude-hot blowing time mapping relationship of the tower is restored to the factory default value, and the dynamic update mechanism is reactivated after 10 cycles of operation.

[0068] It should be noted that maintaining an independent historical temperature rebound amplitude sequence and an actual effective adsorption duration sequence for each adsorption tower can eliminate interference from differences in operating conditions between towers, enabling independent and accurate assessment of the failure risk of each tower, and avoiding distortion of risk judgment due to shared data.

[0069] Among them, the historical temperature rebound amplitude sequence is an annular buffer (capacity 50 cycles) independently opened by the controller for each adsorption tower. The cooling temperature rebound amplitude values ​​(unit: ℃, retained to 1 decimal place) of each cycle are stored sequentially in time order (new data overwrites the earliest data). Each cycle corresponds to a unique serial number (starting from 1 and increasing without upper limit) and is associated with the actual effective adsorption time. The actual effective adsorption duration sequence is stored independently for each adsorption tower, representing the duration from adsorption start-up to the outlet dew point reaching the preset alarm threshold (default -40℃, which can be adjusted according to process requirements) within the corresponding cycle, in minutes (rounded to the nearest integer). If the dew point does not reach the alarm threshold during the entire adsorption cycle, it is recorded as the total duration of that cycle (e.g., 4 hours is recorded as 240 minutes).

[0070] The specific implementation method for failure risk ranking is as follows: For each adsorption tower, the controller calculates two core indicators in real time: 1) the average temperature rebound amplitude over the most recent ten consecutive cycles (denoted as Tavg); 2) the average actual effective adsorption time over the most recent ten consecutive cycles (denoted as tavg). The failure risk scoring formula is set as: Risk Score = (Tavg / Historical Maximum Possible Rebound Amplitude) × 60 + (1 - tavg / Historical Maximum Effective Adsorption Time) × 40, where the historical maximum possible rebound amplitude is set to 10℃, and the historical maximum effective adsorption time is set to the design adsorption cycle of the tower (e.g., 240 minutes). The score range is 0-100 points, with higher scores indicating higher failure risk. All adsorption towers are arranged in descending order of risk score to form the failure risk ranking result (e.g., if the first tower scores 85 points and the second tower scores 40 points, the ranking is first tower > second tower).

[0071] The dynamic rotation sequence is generated by pre-setting the basic rotation sequence as an alternating operation mode of "first tower to second tower to first tower to second tower...".

[0072] After obtaining the failure risk ranking, for the high-risk adsorption towers at the top of the ranking, their operating frequency is reduced within N consecutive rotation cycles (N ranges from 5 to 10, with a default of 8). The ratio of the number of operations for high-risk adsorption towers to low-risk adsorption towers is set to 1:2 (i.e., low-risk towers operate twice, high-risk towers operate once). For example, if the ranking is first tower (high risk) > second tower (low risk), the dynamic rotation sequence is "second tower to second tower to first tower to second tower to second tower to first tower…". After N cycles, the failure risk ranking is recalculated, and the rotation sequence is adjusted again to achieve dynamic optimization.

[0073] "Normal hot blowing parameters" refers to the baseline hot blowing temperature (e.g., 240℃) and baseline hot blowing time (e.g., 4 hours) of the adsorption tower, without any added hot blowing time increments or temperature increases; "more than three consecutive switching cycles" refers to the cycle number being continuously increasing and the number being ≥3 (e.g., cycles 10, 11, and 12, without intervals); "Cooling rebound amplitude drops below the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the tower" means that the temperature rebound amplitude of the current cycle is < the maximum rebound amplitude recorded in the most recent ten cycles (e.g., if the maximum value of the most recent ten cycles is 5.2℃, the rebound amplitude of the current cycle must be <5.2℃ to meet the condition).

[0074] When the above conditions are met, the controller immediately performs the following operations: 1) Sets the status bit (corresponding bit in the register address) of the tower to "unavailable" (binary 1111); 2) In the subsequent rotation logic, skips the operation allocation of the tower (even if it is the tower's turn, it will automatically switch to the next low-risk tower); 3) Outputs a molecular sieve replacement instruction to the DCS, with the instruction format being "Tower No.: X; Status: Unavailable; Reason: Rebound amplitude exceeded the standard for 3 consecutive cycles; Maximum rebound amplitude in the last ten cycles: Y℃; Consecutive exceedance cycles: Z; Recommended replacement time limit: 72 hours"; 4) Issues a continuous alarm (sound frequency 1kHz, light flashing red) through the audible and visual alarm (installed in the central control room), and simultaneously sends a text message notification to the mobile phone of the maintenance personnel (implemented based on the 4G module of the PLC).

[0075] When any adsorption tower is marked as unavailable and removed from the rotation sequence, the system further includes: automatically identifying the number of remaining adsorption towers in the current rotation sequence; dynamically adjusting the adsorption cycle length of each remaining adsorption tower based on the number of remaining towers to maintain the continuous operation capability of the unit; wherein the adjustment method for the adsorption cycle length is: the adsorption cycle length is inversely proportional to the number of remaining adsorption towers; and when the number of remaining adsorption towers falls below the minimum continuous operation quantity, automatically generating a unit load reduction operation prompt and recommending the maximum allowable gas throughput. By adjusting the adsorption cycle inversely and automatically reducing the load prompt, the overall processing capacity can be kept stable in the event of a single tower failure, avoiding production fluctuations, while preventing the remaining towers from operating under overload, protecting the molecular sieve and ensuring that the dehydration index meets the requirements, thus improving the system's fault tolerance.

[0076] It should be noted that the controller identifies the number of remaining available towers in real time by scanning the available status flags of each adsorption tower, with an identification cycle of no more than 1 second, ensuring real-time adjustment. The adsorption cycle duration is calculated using an inverse proportional formula: Adjusted adsorption cycle duration = Original baseline adsorption cycle duration × Total number of towers before removal ÷ Remaining number of towers. For example, in a dual-tower system, if one tower is removed and one tower remains, and the original cycle was 240 minutes, the adjusted cycle will be 480 minutes, keeping the total water treatment capacity constant and avoiding sudden changes in treatment capacity due to a reduction in the number of towers. The minimum number of continuously operating towers is set to 2 according to the device design requirements. When the number of remaining towers is less than 2, the controller calculates the recommended value based on the maximum allowable gas throughput = Rated gas throughput × Number of remaining towers ÷ Designed number of towers in operation, and generates text and audio-visual prompts on the monitoring interface to remind maintenance personnel to reduce the load in time, preventing single-tower overload operation from causing dew point breakthrough or molecular sieve overheating damage.

[0077] By dynamically adjusting the rotation sequence based on the failure risk ranking, the operating frequency of high-risk towers can be reduced, thereby decreasing their adsorption load (running times reduced by 50%, load reduced simultaneously) and regeneration frequency, thus delaying further performance degradation (experiments have verified that this can extend the service life of high-risk towers by 10-15 cycles). At the same time, low-risk towers in better condition can undertake more operating tasks, extending the usable window of the unit while ensuring the overall dehydration effect (outlet dew point ≤ -40℃), and avoiding unit shutdown due to single tower failure.

[0078] If an adsorption tower fails to control the cooling rebound amplitude within the historically reasonable range for more than three consecutive cycles under conventional hot blowing parameters, it indicates that its molecular sieve has undergone irreversible deterioration (such as micropore blockage and loss of active components). Conventional regeneration methods cannot restore adsorption performance. In this case, the tower is marked as unusable and automatically removed from the rotation sequence. This can avoid production risks such as dew point exceeding the standard and moisture penetration caused by the continuous operation of the tower with defects (experiments have verified that it can reduce the probability of dew point exceeding the standard by more than 90%). At the same time, the replacement instruction can be generated to directly guide the operation and maintenance personnel to perform the replacement operation, improving the timeliness of handling (response time ≤ 1 minute).

[0079] Obtain the outlet dew point curve of the current adsorption tower in the next adsorption cycle, and compare the outlet dew point curve with the temperature rebound amplitude obtained this time; through cross-verification of regeneration effect and adsorption effect, it is possible to accurately distinguish between molecular sieve aging and incomplete regeneration, reduce the misjudgment rate of single index, and provide a reliable basis for accurate fault handling.

[0080] It should be noted that the outlet dew point curve is collected by an online dew point meter (model: DMT340, measurement range -60℃ to +20℃, measurement accuracy ±1℃) installed at the adsorption tower outlet. The collection frequency is 1 time / minute. The collected data is transmitted to the PLC controller in real time and stored in the format of "timestamp-dew point value" to form the dew point curve for that adsorption cycle. The core purpose of the comparison is to determine the cause of the abnormality (whether it is a decline in the performance of the molecular sieve itself or insufficient regeneration parameters) by cross-validating the "residual moisture state after regeneration (temperature rebound amplitude)" and the "adsorption process effect (dew point curve)", thus avoiding misidentification caused by a single indicator (false positive rate ≤3%).

[0081] When the outlet dew point curve rises at the beginning of the adsorption cycle, and the temperature rebound amplitude is no greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the molecular sieve is aging, and a molecular sieve replacement warning signal is generated. The "beginning stage" is defined as the first 1 / 5 of the adsorption cycle. This proportion is based on the typical characteristics of the molecular sieve adsorption breakthrough curve: during the first 20% of the adsorption cycle, the bed is in a stable adsorption zone, and the outlet dew point fluctuation is usually ≤±2℃. If a dew point rise occurs during this stage, it indicates that the molecular sieve adsorption capacity has irreversibly decreased. This judgment rule has been verified by 100 sets of industrial operation data, with a false judgment rate of ≤3%. "Rise" means meeting one of the following conditions: the dew point rises continuously during the beginning stage with a total amplitude ≥3℃, or the dew point at the end of the beginning stage is ≥4℃ higher than the initial value. By comparing the initial dew point characteristics with historical rebound amplitudes, early warning of molecular sieve aging can be achieved, allowing sufficient time for maintenance and avoiding production interruptions caused by accelerated aging.

[0082] It should be noted that the "initial stage of the adsorption cycle" is defined as the first 1 / 5 of the adsorption cycle (e.g., the first 48 minutes of a 4-hour cycle, 240 minutes × 1 / 5 = 48 minutes). During this stage, the bed should be in a state of high adsorption capacity, and the dew point should remain stable (usually below -40℃, with fluctuations ≤ ±2℃). "The dew point curve shows an increase in the initial stage" means that the dew point value continuously rises from the initial stable value (e.g., -45℃) within the initial stage, and the increase is ≥ 5℃ (e.g., from -45℃ to -38℃), without showing a downward trend.

[0083] "Temperature rebound amplitude not greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of this adsorption tower" means that the temperature rebound amplitude of the current cycle is less than or equal to the maximum rebound amplitude recorded in the most recent ten cycles (e.g., the maximum value of the most recent ten cycles is 4.8℃, and the rebound amplitude of the current cycle is 4.2℃, which meets the condition). This indicates that the regeneration has been sufficient, the residual moisture content is within the historical reasonable range, and the rise in dew point is not caused by insufficient regeneration. Therefore, it is determined to be molecular sieve aging (permanent decrease in adsorption capacity).

[0084] The warning signal is implemented as follows: The PLC controller transmits the warning signal to the monitoring terminal in the central control room via the Modbus protocol. The signal includes the tower number, the judgment criteria (the initial dew point rise, the temperature rebound, and the maximum rebound in the last ten cycles), and the current molecular sieve operating cycle number. The monitoring terminal alerts the operation and maintenance personnel with an audible and visual alarm (the sound lasts for 30 seconds and the light is solid yellow). At the same time, the alarm information is stored in the system log (retained for 90 days) to provide a preliminary basis for replacement decisions.

[0085] When molecular sieve aging is detected, the process also includes: acquiring historical switching cycle records of the current adsorption tower, which include the actual effective adsorption time of the next adsorption cycle after each switch; performing correlation analysis between historical temperature rebound amplitude sequences and historical sequences of actual effective adsorption times; and predicting that the current adsorption tower is about to fail and generating a molecular sieve replacement prompt when, in three or more consecutive switching cycles, the actual effective adsorption time under the same preset rebound amplitude range is less than the minimum historical effective adsorption time corresponding to that rebound amplitude range. Through multi-cycle correlation analysis of operating condition data, interference from operating condition fluctuations can be eliminated, the trend of molecular sieve performance degradation can be accurately identified, failure can be predicted in advance, and planned maintenance can be supported to reduce production risks.

[0086] It should be noted that the "historical switching cycle record" is stored in the controller's non-volatile memory and contains relevant data for at least 20 cycles of the current adsorption tower. Each record includes the cycle number, switching time, temperature rebound amplitude, actual effective adsorption duration, and dew point curve characteristics to ensure that the correlation analysis has sufficient data samples to support it (sample size ≥ 20 groups).

[0087] "Same temperature rebound amplitude" refers to grouping the historical temperature rebound amplitude sequence into intervals (consistent with the rebound amplitude-heat blowing time mapping interval, i.e. 0-2℃, 2-4℃, 4-6℃, and above 6℃). Rebound amplitudes within the same interval are considered "same". "Minimum historical effective adsorption time corresponding to this rebound amplitude" refers to calculating the actual effective adsorption time of all historical cycles (at least 5) within each interval and taking the minimum value among them (e.g., the historical effective adsorption time in the 0-2℃ interval is 230 minutes, 225 minutes, 235 minutes, 228 minutes, and 232 minutes, with the minimum value being 225 minutes).

[0088] The specific implementation method of correlation analysis is as follows: The PLC controller performs the following steps: 1) Extract the temperature rebound amplitude and the corresponding actual effective adsorption time for the current three or more consecutive switching cycles; 2) Determine the interval to which the temperature rebound amplitude of each cycle belongs; 3) Query the minimum historical effective adsorption time corresponding to each interval; 4) Compare the actual effective adsorption time of the current cycle with the historical minimum value of the corresponding interval. If all consecutive cycles satisfy "actual effective adsorption time < historical minimum value", it is determined that the adsorption capacity is continuously decaying and it is predicted that it will fail soon.

[0089] For example, the temperature rebound amplitudes for three consecutive cycles were 1.5℃ (0-2℃ range), 1.8℃ (0-2℃ range), and 1.3℃ (0-2℃ range), respectively, with corresponding actual effective adsorption times of 220 minutes, 218 minutes, and 215 minutes, respectively. All of these are less than the historical minimum of 225 minutes in this range, thus meeting the condition.

[0090] The threshold setting of "more than three consecutive switching cycles" is based on the statistics of aging conditions on site: from the appearance of signs of aging to complete failure, molecular sieves usually experience a rapid decline in effective adsorption time over 3-5 consecutive cycles. This threshold can avoid misprediction caused by a single abnormal data (such as fluctuations in raw gas) (misprediction rate ≤2%), and can also remind maintenance personnel to prepare for replacement in advance (usually reserving a replacement window of 1-2 cycles, about 4-8 hours).

[0091] The molecular sieve replacement prompts are distinct from previous replacement warning signals: the replacement warning signal is "Aging Warning" (basic warning), while the replacement prompt is "Imminent Failure, Need to be Replaced Immediately" (advanced warning). The prompts are delivered via: a pop-up window on the monitoring terminal (which cannot be closed; the "Confirm Received" button must be clicked), and an SMS message sent via 4G module to the mobile phones of three pre-selected maintenance personnel (the SMS message includes the tower number, predicted failure time, current effective adsorption time, and historical minimum value). Simultaneously, relevant data is recorded in maintenance reports, providing accurate data for maintenance plan development.

[0092] When impending failure is predicted, at least one of the following operations is automatically performed: reduce the gas throughput of the dehydration unit where the current adsorption tower is located; increase the regeneration temperature or extend the regeneration time of the current adsorption tower; determine the failure risk ranking based on the order of predicted failure times for each tower, adjust the rotation sequence, and schedule adsorption towers with higher failure risk to operate during periods with better gas and gas conditions, while adsorption towers with lower failure risk to operate during periods with worse gas and gas conditions. Through this combination of load reduction, enhanced regeneration, and optimized scheduling, dew point deterioration can be delayed, the safe operating window can be extended, time can be gained for molecular sieve replacement, and unplanned downtime can be avoided.

[0093] It should be noted that the above strategy is automatically executed after the prediction fails. The purpose is to delay the risk of dew point exceeding the standard through three means: load matching, enhanced regeneration, and timing optimization during the molecular sieve replacement window, so as to ensure the continuous and stable operation of the unit (target: outlet dew point ≤ -40℃ before replacement).

[0094] The specific implementation of reducing the processed gas volume is as follows: The PLC controller outputs a 4-20mA control signal to the flow regulating valve (installed on the raw gas inlet main pipe, model: VQ series electric regulating valve) to reduce the processed gas volume to 80%-90% of the current set value (default reduction is 85%, the percentage can be set via DCS). The reduction range is adjusted according to the failure risk score (the higher the score, the greater the reduction). Reducing the processed gas volume directly reduces the moisture load per unit molecular sieve (the load is proportional to the processed gas volume), alleviating adsorption pressure and preventing the dew point from rapidly exceeding its limit due to insufficient adsorption capacity.

[0095] Specific parameters for increasing regeneration temperature or extending regeneration time: When increasing the regeneration temperature, increase the hot blowing temperature of the tower by 5%-10% from the baseline value (e.g., 240℃) (default increase of 8%, i.e., 240℃ × 1.08 = 259.2℃, rounded to the nearest integer 259℃), not exceeding the molecular sieve's tolerance temperature (e.g., 300℃); When extending the regeneration time, add an extra 5 minutes to the original hot blowing time increment (e.g., if the original increment was 6 minutes, adjust to 11 minutes) to ensure increased regeneration depth, remove as much residual moisture as possible from the deep micropores of the molecular sieve that is difficult to desorb, and slow down the rate of adsorption performance decay (experiments have shown that this can extend the effective adsorption time by 10%-15%).

[0096] Gas quality and chemistry conditions are categorized and adjusted accordingly: Periods with favorable conditions refer to those with a feed gas moisture content ≤15g / m³ and a throughput fluctuation ≤±10% (based on historical data, typically 00:00-08:00 daily); periods with unfavorable conditions refer to those with a feed gas moisture content >15g / m³ and a fluctuation >±10% (typically 08:00-24:00 daily). The controller automatically identifies the current gas quality and chemistry conditions by real-time acquisition of feed gas moisture content (online moisture meter, measurement range 0-50g / m³, accuracy ±0.5g / m³) and throughput data. Based on the failure risk ranking, the controller adjusts the rotation: high-risk towers operate only during periods with favorable conditions, while low-risk towers operate preferentially during periods with unfavorable conditions. This ensures that the adsorption burden on high-risk towers is minimized, while low-risk towers fully utilize their adsorption capacity. Without altering the overall throughput (daily throughput fluctuation ≤±5%), the system maximizes the utilization of remaining adsorption capacity and smoothly transitions to molecular sieve replacement.

[0097] When the outlet dew point curve experiences a sudden rise in the latter half of the adsorption cycle, and the temperature rebound amplitude exceeds the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the release of residual moisture is caused by incomplete regeneration. The hot blowing time increment for the next regeneration is then increased by a preset fixed step. By linking the dew point jump characteristic with the abnormal rebound amplitude, the problem of incomplete regeneration can be quickly located and corrected in a timely manner, avoiding the recurrence of similar problems, improving the system's adaptive adjustment capability, and ensuring that the regeneration effect continuously meets the standards.

[0098] It should be noted that "the second half of the adsorption cycle" is defined as the last 1 / 3 of the adsorption cycle (e.g., the last 80 minutes of a 4-hour cycle, 240 minutes × 1 / 3 = 80 minutes); "sudden rise" is defined as an increase in the outlet dew point of more than 5°C within 10 minutes (e.g., from -42°C to -35°C), and the dew point value after the rise is > -40°C (alarm threshold). This phenomenon is because the bed adsorption capacity has been saturated, and the residual moisture from the incomplete regeneration is released in a concentrated manner in the later stage of adsorption.

[0099] "Temperature rebound amplitude greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of this adsorption tower" means that the temperature rebound amplitude of the current cycle is greater than the maximum rebound amplitude recorded in the most recent ten cycles (e.g., the maximum value of the most recent ten cycles is 5.2℃, and the current cycle is 6.1℃). This indicates that the residual moisture content is higher than the normal level in the past, further confirming that the regeneration is incomplete.

[0100] The preset fixed step size is 2 minutes. This value has been verified through multiple sets of incomplete regeneration conditions: increasing the hot blowing time by 2 minutes can effectively remove additional residual moisture (corresponding to an increase in residual moisture content of 0.4 to 0.6 kg / m³), and will not lead to over-regeneration due to an excessively large step size (over-regeneration will increase energy consumption by 10%-20% and may accelerate molecular sieve aging). The adjustment of the hot blowing time increment is automatically written into the hot blowing parameter adjustment mapping relationship of this tower, and will take effect directly during the next regeneration without manual intervention.

[0101] After each tower switch, the temperature rebound amplitude, corresponding range, hot blowing time increment, dew point curve characteristics, and actual effective adsorption duration are all archived along with the tower number. This updates the historical rebound data sequence and historical switch cycle records for that tower, providing continuous data for subsequent parameter adjustments, fault diagnosis, and failure prediction. By fully archiving operational data and continuously updating historical sequences, continuous data support can be provided for control strategy optimization, making parameter adjustments and fault diagnosis more accurate and improving the long-term stability and adaptability of the system.

[0102] It should be noted that data archiving and updating operations are performed immediately after each tower switch, with a data update delay of no more than 1 minute, to ensure that the latest historical data is used in subsequent control cycles, thereby improving the accuracy of judgment and calculation.

[0103] Example This invention was applied to a dual-tower molecular sieve dehydration unit (processing capacity 3500 m³ / h, 12 m³ / tower of 3A molecular sieve) in a natural gas processing plant. After 30 days of operation using the traditional method (60 minutes of cooling + 30 minutes of supplementary cooling), the average rebound rate of tower A was 3.2℃, and the dew point exceeded -40℃ at the 210th minute of the adsorption cycle. After 30 cycles using the method of this invention (eliminating supplementary cooling and adjusting the next hot-blowing time by 4 minutes corresponding to a rebound rate of 2~4℃), the rebound rate of both towers stabilized at 2.1~4.5℃, the actual effective adsorption time was ≥230 minutes, and no dew point breakthrough occurred; the single-tower cycle was shortened from 330 minutes to 300 minutes, saving approximately 1200 m³ of cooling gas per day. After 90 days, the dew point of Tower A initially rose (from -46℃ to -41℃ in the first 48 minutes). The system determined that the molecular sieve was aging and issued an early warning. Subsequently, when the effective adsorption time dropped to 195 minutes under the same rebound amplitude, the failure prediction was triggered. The system automatically reduced the load to 3000 m³ / h and adjusted the rotation sequence. Tower A continued to operate stably for 7 days before the planned replacement.

[0104] After the cooling phase, a certain amount of heat and adsorbed moisture remain inside the molecular sieve bed. Heat conduction to the outside causes a short-term rebound in the outlet temperature, the magnitude of which directly reflects the amount of residual moisture inside the bed. If supplementary cooling is performed at this time, it will directly lengthen the single-tower cycle, disrupt the dual-tower switching sequence, and affect the continuous processing capacity of the unit. Simultaneously, the residual moisture in the bed will not disappear automatically after cooling. If feedforward adjustment is not implemented in advance, it will gradually accumulate in the next adsorption cycle, continuously weakening the molecular sieve's adsorption performance and causing the outlet dew point to gradually deteriorate. Based on these physical laws, this invention adopts the following control logic: quantifying residual moisture by the temperature rebound amplitude, canceling the supplementary cooling process to ensure stable switching sequence, and simultaneously using residual moisture as a feedforward correction factor in the next cycle's regeneration control to eliminate residual moisture accumulation through advance compensation. By combining timing assurance with feedforward correction, the accumulation of residual moisture in each cycle can be prevented from the source without disrupting the dual-tower switching rhythm, ensuring continuous and stable operation of the unit while avoiding molecular sieve performance degradation and dew point exceedances due to moisture accumulation.

[0105] Specifically, the mapping relationship between historical rebound data and hot blowing parameters is maintained independently for each adsorption tower. After the valve is closed at the end of cooling, the temperature rebound amplitude is calculated to determine the residual moisture in the bed and directly execute tower switching. By maintaining independent data for multiple towers and collecting data at a unified node, data interference caused by differences in the operating conditions of different towers can be eliminated, ensuring the accuracy and reliability of the residual moisture judgment results and providing a stable basis for subsequent precise adjustment of regeneration parameters.

[0106] Based on the temperature rebound amplitude, the corresponding hot blowing time increment is matched, and the hot blowing time of the next cycle is set as the base time plus the increment value, so that the regeneration intensity is adapted to the residual moisture content. By achieving graded adjustment of regeneration intensity through segmented mapping, the hot blowing parameters can be accurately matched with the actual state of the bed, avoiding moisture residue due to insufficient regeneration, and preventing energy waste and accelerated aging of molecular sieves due to over-regeneration.

[0107] When the outlet dew point of the next adsorption cycle rises at the beginning of the cycle and the rebound is within the historically reasonable range, the molecular sieve is determined to have entered an aging state and an early warning is issued. By combining the characteristics of dew point changes with historical rebound data, it is possible to accurately distinguish between two types of problems: molecular sieve aging and incomplete regeneration, enabling early identification of equipment status and allowing maintenance personnel sufficient time to prepare for handling.

[0108] When the effective adsorption time remains below the historical minimum for several consecutive cycles with the same rebound amplitude, the molecular sieve is predicted to fail and the operating strategy is automatically optimized. By correlating and comparing historical data from multiple cycles, interference from operating condition fluctuations can be effectively eliminated, the trend of molecular sieve performance degradation can be accurately identified, failure can be predicted in advance, and unplanned downtime caused by sudden failure can be avoided.

[0109] When the dew point suddenly rises in the latter half of the adsorption cycle and the rebound amplitude exceeds the historical maximum value, incomplete regeneration is determined, and the hot blowing time step is automatically increased. By linking the sudden rise in dew point characteristics with abnormal rebound amplitude, the problem of incomplete regeneration can be quickly located and the regeneration parameters can be corrected in a timely manner, improving the system's adaptive adjustment capability and ensuring the continuous and stable dehydration effect in subsequent cycles.

[0110] After a tower is determined to have failed and is shut down, the adsorption cycle length is adjusted inversely to the number of remaining towers to maintain the overall processing capacity essentially unchanged. By dynamically adapting the adsorption cycle length, the processing capacity of the unit can be smoothly transitioned after a single tower fails, avoiding large fluctuations in production load, while preventing the remaining towers from operating under overload, protecting the molecular sieve, and ensuring that the dewatering indicators meet the standards.

[0111] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for switching control of a natural gas molecular sieve dehydration bed, characterized in that, include: At the end of the cooling phase after regeneration, the cooling gas valve is closed, and the temperature rebound amplitude of the outlet temperature after the cooling gas valve is closed is obtained; The residual moisture content inside the bed is determined based on the temperature rebound amplitude. No supplemental cooling is performed; tower switching operation is allowed directly. The residual moisture content is used as a feedforward correction amount to adjust the hot blowing parameters during the next regeneration of the adsorption tower, so that the residual moisture from this regeneration can be removed together with the moisture that needs to be removed normally in the next regeneration.

2. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 1, characterized in that, Obtaining the temperature rebound amplitude includes: The outlet temperature at the moment of shutdown is recorded as the reference temperature. The outlet temperature within a preset time window after shutdown is continuously monitored. The difference between the highest temperature within the preset time window and the reference temperature is taken as the temperature rebound amplitude.

3. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 1, characterized in that, The residual moisture content is used as a feedforward correction factor to adjust the hot blowing parameters during the next regeneration of the adsorption tower, specifically including: Establish a mapping relationship between the rebound amplitude and the hot blowing time of the adsorption tower, where different rebound amplitude ranges correspond to different hot blowing time increments; When the temperature rebound falls within a certain range at the end of this cooling cycle, the corresponding hot blowing time increment is obtained according to the mapping relationship, and the hot blowing time for the next regeneration is set as the base hot blowing time plus the hot blowing time increment.

4. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 3, characterized in that, The mapping relationship is dynamically updated during the operation of the adsorption tower: Record the cooling rebound amplitude after each regeneration and the hot blowing time used in this regeneration; When the cooling rebound amplitude corresponding to a certain rebound amplitude range is greater than the maximum value of the historical rebound amplitude sequence of that range in three or more consecutive switching cycles, the hot blowing time increment corresponding to that range is automatically increased; when the cooling rebound amplitude is less than the minimum value of the historical rebound amplitude sequence of that range in three or more consecutive switching cycles, the hot blowing time increment corresponding to that range is automatically decreased.

5. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 1, characterized in that, The method is applied to a rotating dehydration device comprising at least a first adsorption tower and a second adsorption tower: The first adsorption tower and the second adsorption tower independently maintain their respective historical rebound data and hot blowing parameter adjustment mapping relationship; When the first adsorption tower performs a switch, only the historical data of the first adsorption tower itself is used to adjust the hot blowing parameters for its next regeneration; when the second adsorption tower performs a switch, only the historical data of the second adsorption tower itself is used.

6. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 1, characterized in that, Also includes: Obtain the outlet dew point curve of the current adsorption tower in the next adsorption cycle, and compare the outlet dew point curve with the temperature rebound amplitude obtained this time; When the outlet dew point curve rises at the beginning of the adsorption cycle and the temperature rebound amplitude is not greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the molecular sieve is aging and a molecular sieve replacement warning signal is generated. When the outlet dew point curve shows a sudden increase in the second half of the adsorption cycle, and the temperature rebound amplitude is greater than the maximum value of the most recent ten cycles in the historical rebound amplitude sequence of the adsorption tower, it is determined that the residual moisture release is caused by incomplete regeneration, and the hot blowing time increment for the next regeneration is increased by a preset fixed step.

7. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 6, characterized in that, When molecular sieve aging is determined, it also includes: Obtain the historical switching cycle record of the current adsorption tower, the historical switching cycle record including the actual effective adsorption time of the next adsorption cycle after each switching; The correlation analysis was performed between the historical temperature rebound amplitude sequence and the historical sequence of the actual effective adsorption time; If, in three or more consecutive switching cycles, the actual effective adsorption time under the temperature rebound amplitude within the same preset rebound amplitude range is less than the minimum historical effective adsorption time corresponding to that rebound amplitude range, it is predicted that the current adsorption tower will soon fail, and a molecular sieve replacement prompt will be generated.

8. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 7, characterized in that, When impending failure is predicted, automatically perform at least one of the following actions: Reduce the processing gas volume of the dehydration unit where the current adsorption tower is located; Increase the regeneration temperature of the current adsorption tower or extend the regeneration time; The failure risk ranking is determined based on the order of the predicted failure times of each tower. The rotation sequence is then adjusted so that the adsorption towers with higher failure risk are scheduled to operate during periods with better gas and gas conditions, while the adsorption towers with lower failure risk are scheduled to operate during periods with worse gas and gas conditions.

9. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 8, characterized in that, The method is applied to a rotating dehydration device comprising at least a first adsorption tower and a second adsorption tower: Each adsorption tower independently maintains its historical temperature rebound amplitude sequence and actual effective adsorption duration sequence to independently determine the failure risk ranking of each tower; The rotation sequence is dynamically generated based on the failure risk ranking of each tower, so that the adsorption towers that are at the front of the failure risk ranking appear less often in the rotation sequence than the adsorption towers that are at the back of the failure risk ranking. If an adsorption tower fails to reduce the cooling rebound amplitude to below the maximum value of the most recent ten cycles in the tower's historical rebound amplitude sequence after applying its maximum allowed hot blowing time increment for more than three consecutive switching cycles, the tower will be marked as unusable, automatically removed from the rotation sequence, and a replacement instruction will be generated.

10. The method for switching control of a natural gas molecular sieve dehydration bed according to claim 9, characterized in that, When any adsorption tower is marked as unavailable and removed from the rotation sequence, the following also applies: Automatically identify the number of remaining adsorption towers in the current rotation sequence; The adsorption cycle duration of each remaining adsorption tower is dynamically adjusted based on the number of remaining adsorption towers to maintain the continuous operation capability of the device; wherein, the adjustment method for the adsorption cycle duration is that the adsorption cycle duration is inversely proportional to the number of remaining adsorption towers; Furthermore, when the number of remaining adsorption towers falls below the minimum number of continuous operations, the system automatically generates a load reduction operation prompt and recommends the maximum allowable gas throughput.

Images