Dynamic control purification system and method for multi-process source hydrogen

Abstract

The invention belongs to the technical field of gas separation, and particularly relates to a dynamic control purification system and method for multi-process source hydrogen, and the system comprises a mixing and pretreatment unit, a multi-tower PSA separation unit, a product and waste gas collection unit and a control and execution unit; the mixing and pretreatment unit comprises a gas mixer, a preheater and a deaerator which are connected in sequence; the multi-tower PSA separation unit is connected between the mixing and pretreatment unit and the product and waste gas collection unit through a pipeline, and the multi-tower PSA separation unit at least comprises four adsorption towers which are connected in parallel; the control and execution unit comprises a control module, a first online analyzer and a second online analyzer. Compared with the prior art, the problem that in the prior art, operation of a fixed time sequence program cannot effectively cope with impurity fluctuation existing in hydrogen of a mixed source is solved. According to the scheme, on the basis of dynamic regulation and control of the impurity concentration, tiny fluctuation of the product purity is maintained under large feeding concentration fluctuation, and qualified output of the product is guaranteed.

Description

A dynamic control purification system and method for hydrogen from multiple process sources Technical Field

[0001] This invention belongs to the field of gas separation technology, specifically relating to a dynamic control purification system and method for hydrogen from multiple process sources. Background Technology

[0002] Coke oven gas hydrogen production, as a representative of industrial by-product hydrogen, has a complex gas composition, typically containing impurities such as hydrogen (H2), carbon monoxide (CO), methane (CH4), carbon dioxide (CO2), and trace amounts of sulfides. On the other hand, water electrolysis can produce high-purity "green hydrogen," but the production process consumes a lot of energy.

[0003] Currently, industrial applications of these two hydrogen sources mostly employ independent purification routes. For example, coke oven gas is used to extract hydrogen through conventional PSA units, while hydrogen from water electrolysis is used directly after simple deoxygenation and drying. This separate processing model leads to problems such as redundant equipment investment, large footprint, and low operating efficiency. Especially when it is necessary to mix and purify hydrogen from different sources at hydrogen refueling stations or chemical industrial parks, existing technologies lack efficient, economical, and flexible solutions for such specific mixed gases; for example, the main components of hydrogen produced by water electrolysis are: H2: 99.8%, O2: 0.2%, H2O: 4 g / m³. 3 The main components of hydrogen extracted from coke oven gas via PSA are: H2: 87%, CO: 3.5%, CH4: 3.0%, CO2: 5.5%, with the remainder being N2. The main difference between the two gas sources is that hydrogen produced by water electrolysis contains oxygen and moisture, resulting in higher hydrogen purity; while hydrogen extracted from coke oven gas mainly contains CO, CH4, CO2, and N2 as impurities. Existing separation processes typically select adsorption packing materials and equipment based on specific gas compositions, using fixed program control to remove impurities. Directly processing a mixture of gases from different sources using existing separation processes would lead to low equipment utilization, high equipment investment, and low hydrogen yield.

[0004] Traditional PSA process control systems typically operate based on preset, fixed timing sequences. All valves open and close at fixed times according to each step within a cycle (such as adsorption, equalization, forward release, reverse release, rinsing, and pressurization). Examples include a cryogenic liquid carbon dioxide pressure swing adsorption system for hydrogen purification disclosed in CN112239197A, a two-stage concentration-separation PSA method for recovering C2+ from refinery dry gas disclosed in CN107433107B, and a pressure swing adsorption gas separation and purification system and method for hydrogen purification disclosed in CN111204712A. However, when the feed composition deviates from the design values ​​due to fluctuations in the mixed gas content, the fixed timing control program reveals significant shortcomings: if the impurity concentration increases, the adsorption bed will prematurely break through, resulting in substandard product purity; if the impurity concentration decreases, the adsorption bed will not be fully utilized before switching, leading to a decrease in product recovery. To address these fluctuations, the conventional approach is conservative design, increasing the adsorbent loading or shortening the adsorption time, but this undoubtedly increases investment and operating costs.

[0005] The prior art CN121426055A discloses a high-purity hydrogen recovery process based on a PSA device to improve denitrification efficiency by treating feed gas from different sources with PSA. However, this scheme mainly separates and removes different impurity components in the feed gas in sequence, without taking into account the problem of concentration fluctuations, and still belongs to the operation of a fixed time sequence procedure.

[0006] Therefore, developing a PSA dynamic control system that can sense changes in feed, predict bed behavior, and optimize operating procedures in real time is of great significance for improving the adaptability, economy, and reliability of the process. Summary of the Invention

[0007] To address at least one of the aforementioned problems, this invention provides a dynamic control purification system and method for hydrogen from multiple process sources. This addresses the issue that existing technologies with fixed-sequence operations cannot effectively handle impurity fluctuations in mixed-source hydrogen. Based on dynamic control of impurity concentration, this solution maintains minimal fluctuations in product purity even under large feed concentration variations, ensuring qualified product output and enhancing the robustness and economy of the PSA process.

[0008] The objective of this invention is achieved through the following technical solution: The first aspect of this invention discloses a dynamic control purification system for hydrogen from multiple process sources, comprising a mixing and pretreatment unit, a multi-tower PSA separation unit, a product and waste gas collection unit, and a control and execution unit. The mixing and pretreatment unit includes a gas mixer, a preheater, and a deaerator connected in sequence. The gas mixer's inlet receives a coke oven gas hydrogen production stream and an electrolytic water hydrogen production stream, respectively. The multi-tower PSA separation unit is connected between the mixing and pretreatment unit and the product and waste gas collection unit via pipelines, and includes at least four parallel adsorption towers. The control and execution unit includes a control module, a first online analyzer, and a second online analyzer. The first online analyzer is used to obtain the concentration of at least one key impurity component in the mixed gas entering the adsorption tower. The second online analyzer is used to obtain the concentration of at least one key impurity component in the product hydrogen stream exiting the top of the adsorption tower. The control module is communicatively connected to the first and second online analyzers.

[0009] Preferably, the mixing and pretreatment unit further includes proportional regulating valves for adjusting the flow rates of the coke oven gas hydrogen production stream and the electrolytic water hydrogen production stream, respectively.

[0010] Preferably, the multi-tower PSA separation unit further includes a control valve group, wherein each control valve of the control valve group is installed on the pipeline connected to the adsorption tower.

[0011] Preferably, the product and waste gas collection unit includes a product buffer tank and a waste gas processor; the product buffer tank is connected to the top of the adsorption tower and is used to collect the product hydrogen gas stream obtained after adsorption separation; the waste gas processor is connected to the bottom of the adsorption tower and is used to recover the desorption waste gas stream obtained after desorption separation.

[0012] Preferably, the control and execution unit further includes a gas concentration sensor; multiple gas concentration sensors are arranged at intervals along the height of the adsorption tower to obtain the concentration of at least one key impurity component at different locations in different sections of the adsorption tower; the control module is communicatively connected to the gas concentration sensor.

[0013] Preferably, the control and execution unit further includes a host computer; the control module is a PLC controller; and the host computer and the control module are communicatively connected.

[0014] Preferably, the adsorption tower includes a tower body, an inlet distributor and a support grid plate disposed at the bottom of the tower body, and an outlet collector disposed at the top of the tower body; a composite adsorbent is filled on the support grid plate inside the tower body for absorbing and removing water, carbon dioxide, methane, carbon monoxide and nitrogen respectively.

[0015] The second aspect of this invention discloses a dynamic control purification method for hydrogen from multiple process sources, used for dynamic control of any of the systems described above; the method includes the following steps: S1: setting a concentration change threshold ΔC set , i and maximum safe duration T max Based on the standard feed component concentration C0, i Determine the baseline duration T0 of the adsorption step, where i represents the critical impurity component (the number of monitored components can be appropriately increased or decreased according to the control objective; considering the high coupling and computational complexity of simultaneously monitoring multiple components and implementing feedback control, usually only one or two difficult-to-adsorb components can be selected as representative critical impurities for control); S2: Obtain the concentration C of at least one critical impurity component in the mixed gas entering the adsorption tower. t , i S3: Calculate concentration C using a feedforward compensation model. t , i Predicted adsorption time T p The feedforward compensation model is a function T of the concentration change of the key impurity component and the baseline time of the adsorption step. p = f(C t , i , C0, i S0: Obtain the gas concentration C at the outlet of the adsorption tower. out , i If the adsorption time t reaches T p C has already been satisfied. out , i - C t , i ≥ ΔC set , i Then the adsorption time t is set to the actual adsorption step time T. actual If the adsorption time t reaches T p C was not satisfied at that time out , i - C t , i ≥ ΔC set , i Then the adsorption time t will be extended by a set time increment Δt until C is satisfied. out , i - C t , i ≥ ΔC set , i Or reach the maximum safe duration T max The adsorption time t is set to the actual adsorption step time T. actual S5: Based on the actual adsorption step duration T actualThe duration of the flushing step or the flushing air flow rate, as well as the duration of the pressure equalization step, can be dynamically adjusted.

[0016] Preferably, the feedforward compensation model is: T p = T0× Π[(C0, i / C t , i )^k i ]; where k i This is a correction factor, with a value ranging from 0.5 to 1.0.

[0017] Preferably, the dynamic adjustment satisfies: total flushing air consumption V purge = K × T actual Where K is the proportionality coefficient.

[0018] The working principle of this invention is as follows: two hydrogen sources (from different sources) are mixed and preheated for deoxygenation, and then adsorption and purification are carried out through a PSA tower packed with composite adsorbent. The process parameters are dynamically adjusted according to the changes in the composition of the feed and outlet materials, and finally high-purity hydrogen is output and the waste gas is purified.

[0019] Compared with the prior art, the present invention has the following beneficial effects: The present invention addresses the problem that the fixed timing control of the existing PSA process cannot adapt to the fluctuation of feed components, resulting in unstable product purity or decreased recovery rate. It constructs a "feedforward-feedback" dual-loop intelligent control system, realizing a fundamental transformation from "time-based control" to "state-based control". Within the range of feed concentration fluctuation of ±30%, it can maintain product purity fluctuation of less than ±0.2%, while increasing the average recovery rate by 5-15%, significantly enhancing the robustness and economy of the PSA process.

[0020] This invention is designed to address the impurities present in the mixed gas of hydrogen produced from coke oven gas and hydrogen produced from water electrolysis. The adsorbent combination adopted can efficiently remove key impurities such as CO, CH4, and CO2, thus solving the selectivity conflict problem of single adsorbent solutions.

[0021] This invention constructs a dual-loop intelligent control architecture with "feedforward-feedback" coordination: 1) Feedforward loop: A live analyzer on the inlet pipeline captures major disturbances in the feed components in real time. Based on the intrinsic relationship between adsorption capacity and concentration, a feedforward model immediately predicts and calculates the reasonable duration of the adsorption step under the current operating conditions, achieving rapid compensation for disturbances. 2) Feedback loop: Gas concentration sensors are deployed at key locations at the adsorption bed outlet to directly monitor the real-time concentration at the adsorption front. The arrival signal of the concentration front serves as the final and most reliable basis for switching adsorption steps, ensuring safe switching before impurities penetrate.

[0022] This invention uses dynamic control to allow one system to replace multiple devices, reducing initial equipment investment and floor space requirements. At the same time, the high recovery rate reduces raw material gas loss, significantly lowering operating costs and demonstrating excellent economic efficiency. Attached Figure Description

[0023] Figure 1 is a schematic diagram of the system structure.

[0024] Figure 2 is a schematic diagram of the adsorption tower.

[0025] Figure 3 is a flowchart of the method.

[0026] In the diagram: 1- Hydrogen production stream from coke oven gas; 2- Hydrogen production stream from water electrolysis; 3- Proportional regulating valve; 4- Gas mixer; 5- Preheater; 6- Adsorption tower; 7- Product hydrogen stream; 8- Product buffer tank; 9- Control valve group; 10- Desorption waste gas stream; 11- Waste gas processor; 12- Control module; 13- Deaerator; 14- First online analyzer; 15- Second online analyzer; 16- Host computer; 17- Gas concentration sensor; 201- Tower body; 202- Inlet distributor; 203- Activated alumina layer; 204- Activated carbon layer; 205- Zeolite molecular sieve layer; 206- Support grid plate; 207- Outlet gas collector. Detailed Implementation

[0027] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0028] Unless otherwise specified, the reagents used in the following description are conventional commercial products, the methods used are well-known in the art, and any other matters not covered herein can be handled using existing technology.

[0029] Example 1: A multi-process hydrogen dynamic control purification system, as shown in Figure 1, includes: a mixing and pretreatment unit for receiving and proportionally mixing coke oven gas hydrogen production stream 1 and water electrolysis hydrogen production stream 2, and preheating and deoxygenating the mixed gas; a multi-tower PSA separation unit connected to the mixing and pretreatment unit for adsorbing and separating impurities in the mixed gas to obtain high-purity hydrogen; a control and execution unit connected to the mixing and pretreatment unit and the multi-tower PSA separation unit for real-time monitoring of changes in the composition and concentration of the mixed gas and dynamically adjusting the PSA process parameters according to the real-time composition of the mixed gas; and a product and waste gas treatment unit connected to the multi-tower PSA separation unit for storing high-purity hydrogen and treating desorption waste gas, respectively.

[0030] The mixing and pretreatment unit includes a gas mixer 4, a proportional control valve 3, a preheater 5, and a deaerator 13 (deoxygenation tower). Oxygen from different sources (coke oven gas hydrogen production stream 1 from coke oven gas to hydrogen production, and electrolytic water hydrogen production stream 2 from electrolytic water hydrogen production) enter the gas mixer 4 independently through their respective inlet pipes for mixing. The proportional control valves 3 are arranged on the inlet pipes to receive and mix the coke oven gas hydrogen production stream 1 and the electrolytic water hydrogen production stream 2 in the gas mixer 4 according to a set ratio. The preheater 5 utilizes waste heat generated by the upstream electrolytic water hydrogen production system to preheat the mixed gas. The deaerator 13 then deoxygenates the preheated mixed gas.

[0031] The multi-tower PSA separation unit consists of at least four adsorption towers 6 connected in parallel. Each adsorption tower 6 is filled with a layered composite adsorbent, as shown in Figure 2, which, from the inlet to the outlet, includes: an activated alumina layer 203 for deep removal of moisture; an activated carbon layer 204 for adsorbing carbon dioxide, methane, and trace sulfides; and a zeolite molecular sieve layer 205 for selectively adsorbing carbon monoxide and nitrogen. The tower structure of the adsorption tower 6 specifically includes a tower body 201, an inlet distributor 202 located at the bottom of the tower body 201 as the inlet, a support grid plate 206 located inside the tower body 201 near the bottom (for supporting the composite adsorbent), and an outlet collector 207 located at the top of the tower body 201 as the outlet.

[0032] Control and execution unit: A first online analyzer 14 (online gas analyzer) upstream of the inlet pipe of adsorption tower 6, multi-point gas concentration sensors 17 distributed axially along the adsorption bed of each adsorption tower 6, a second online analyzer 15 (outlet gas concentration detector) on the outlet pipe of adsorption tower 6, a control module 12 (high-speed PLC), and a control valve group 9 (programmable valve). The first online analyzer 14 and the second online analyzer 15 are specifically tunable diode laser absorption spectrometers (TDLAS) or non-dispersive infrared gas analyzers (NDIR), with a system response time (T) of less than 10 seconds. The first online analyzer 14 can be installed on the connecting pipe between the mixing and pretreatment unit and the multi-tower PSA separation unit to measure the concentration of at least one key impurity component in the feed mixed gas in real time. The gas concentration sensors 17 are installed at least three gas concentration detection points in each adsorption tower 6: upper (near the outlet), middle (located in the middle of the bed), and lower (near the inlet), to monitor the gas concentration changes during adsorption / desorption. The second online analyzer 15 is installed on the pipeline connected to the outlet of each adsorption tower 6 to obtain the concentration change results of at least one key impurity component in the discharged gas. The control valve group 9 is arranged according to the setup required for multi-tower PSA adsorption, as shown in Figure 1. In the above, the signal input terminal of the control module 12 is communicatively connected to the first online analyzer 14, each gas concentration sensor 17 and the second online analyzer 15 to obtain real-time monitored component concentration data; the signal output terminal of the control module 12 is connected to the control valve group 9 to realize valve opening adjustment, forming a closed-loop control system. The control module 12 is further connected to the host computer 16, which integrates a "feedforward-feedback" module for implementing local "feedforward-feedback" control of the PSA separation unit. The feedforward compensation model establishes a functional relationship between the concentration change of key impurity components and the baseline duration of the adsorption step. The functional relationship can be obtained through regression analysis based on adsorption isotherm data or historical operating data. The feedback control logic uses the concentration change value of the gas concentration detection point at the outlet of adsorption tower 6 (monitored by the second online analyzer 15) as the trigger signal for step switching. When the change value exceeds the preset threshold, the current adsorption step is immediately terminated and the next step (adsorption-desorption cycle) is started.

[0033] Product and exhaust gas treatment unit: including product buffer tank 8 (connected to the top of adsorption tower 6) for collecting and storing high-purity hydrogen stream; and exhaust gas processor 11 (connected to the bottom of adsorption tower 6) for catalytic oxidation treatment of desorbed exhaust gas stream 10 to make it meet emission standards.

[0034] A dynamic control purification method for hydrogen from multiple process sources, using the above system, includes the following steps: 1. Mixing, preheating, and deoxygenation: The coke oven gas hydrogen production stream 1 and the water electrolysis hydrogen production stream 2 are mixed in a preset ratio, and the mixed gas is preheated to the target temperature range (determined according to process requirements) using the waste heat from the water electrolysis hydrogen production process, and the oxygen in the mixed gas is removed by a deoxygenation tower.

[0035] 2. Dynamic adsorption purification: The preheated and deoxygenated mixed gas is passed into one or more adsorption towers 6 of the multi-tower PSA separation unit under the target adsorption pressure (determined according to process requirements). The impurity components in the mixed gas are selectively adsorbed by the composite adsorbent, thereby obtaining high-purity hydrogen product at the outlet of adsorption tower 6.

[0036] 3. Pressure regulation and energy recovery: The adsorption tower 6, after completing adsorption, undergoes multiple pressure equalization steps to recover the high-pressure hydrogen inside to other adsorption towers 6 that have completed regeneration, thereby achieving efficient energy utilization.

[0037] 4. Adsorbent regeneration: The adsorbent is regenerated by countercurrent depressurization and flushing with hydrogen byproducts or a small amount of product hydrogen, so that the adsorbed impurities are desorbed from the adsorbent and the adsorption capacity of the adsorbent is restored.

[0038] 5. Product output and exhaust gas treatment: The high-purity hydrogen obtained from the top of the tower in step 2 is transported as product hydrogen stream 7 to product buffer tank 8; at the same time, the exhaust gas desorbed from the bottom of the tower in step 4 is collected as desorbed exhaust gas stream 10 and transported to exhaust gas processor 11 for harmless treatment.

[0039] Step 2, the dynamic adsorption purification process, includes the following steps, as shown in Figure 3: S1. Under initial stable operating conditions, determine the standard feed component concentration C0, i and the corresponding adsorption step baseline time T0 (theoretical optimal value), where i represents the key impurity component; S2. Real-time acquisition of the concentration C of at least one key impurity component in the current feed gas mixture. t , i (Acquired by the first online analyzer 14); S3. Transfer C t , i Input the feedforward compensation model to calculate the predicted duration T of the current cyclic adsorption step. p = f(C t , i , C0, i , T0); S4. With T p The adsorption process is initiated based on the baseline, and the gas concentration C at the outlet of adsorption tower 6 is monitored simultaneously in real time. out , i (Acquired by the second online analyzer 15); S5. Judgment condition: If the duration reaches T...p Previously, C was detected. out , i - C t , i ≥ ΔC set , i (ΔC) set , i If the concentration change of the key impurity component is not lower than the preset threshold (meaning the adsorption target has been achieved), then the adsorption step is immediately terminated, and the process proceeds to the next step. Simultaneously, the adsorption time t is defined as the actual adsorption step time T that satisfies the adsorption requirements. actual If the duration reaches T p The concentration change threshold ΔC has not yet been triggered (reached). set , i If so, the adsorption time t of the adsorption step is extended by a set time increment Δt, and monitoring continues until the concentration change threshold ΔC is triggered (reached). set , i Or reach the maximum safe duration T max (Preset value), and set the adsorption time t to the actual adsorption step time T. actual (T) p +n·Δt or T max Where n is the number of time increments, and T p +n·Δt is not greater than T max S6. Based on the actual adsorption step duration T actual The duration of subsequent flushing steps or flushing air flow rate, as well as the duration of pressure equalization steps, are dynamically adjusted according to preset rules.

[0040] S7: Record and store the C for each loop. t T p T actual And product gas purity data, used to periodically optimize feedforward compensation model parameters and feedback control thresholds.

[0041] Based on the above description, in step S3, the calculation formula for the feedforward compensation model is: T p = T0× Π[(C0, i / C t , i )^k i ]; where k i This is a correction factor, with a value ranging from 0.5 to 1.0, determined through experiments or simulations.

[0042] In step S6, the rule for dynamically adjusting the flushing steps is: total flushing air consumption V purge With T actual Proportional, i.e., Vpurge = K × T actual , where K is a proportionality coefficient, which is determined based on the characteristics of the adsorbate and the adsorbent.

[0043] Example 2 uses a comprehensive chemical industrial park integrating hydrogen production from coke oven gas in a steel plant and hydrogen production from photovoltaic power generation as an example. The coke oven gas hydrogen production unit provides coke oven gas hydrogen production stream 1 (crude hydrogen, flow rate 150 Nm³). 3 / h, composition (volume fraction): H2 87%, CO 3.5%, CH4 3.0%, CO2 5.5%, remainder N2) and the electrolytic hydrogen production stream 2 (flow rate 100 Nm) provided by the photovoltaic water electrolysis hydrogen production unit. 3 / h, the main components (volume fraction): H2>99.6%, O20.2%) need to be mixed and purified to the hydrogen standard for fuel cells (≥99.99%).

[0044] The process is as follows: Two streams of air enter the mixer at a volume ratio of 3:2 through the proportional control valve 3 and are fully mixed.

[0045] The mixed gas is preheated by the preheater 5 using 70°C hot water from the electrolytic cell cooling system, and the temperature is raised to 50°C before entering the deaerator 13 to remove oxygen.

[0046] The preheated and deoxygenated mixed gas enters the PSA unit, which consists of four adsorption towers 6. The control module 12 automatically adjusts the control program through dynamic control based on the acquired component concentration change data.

[0047] During the adsorption stage, impurities are effectively captured by the composite adsorbent layered within the tower. The high-purity hydrogen obtained from the top of the tower enters the product buffer tank 8 as the product hydrogen stream 7. Testing shows that the purity of the product hydrogen is consistently above 99.9993%.

[0048] During the adsorbent regeneration stage, the adsorbent is desorbed by rapid counter-current depressurization and brief product gas flushing through the control valve group 9. The desorbed waste gas is collected and sent to the waste gas processor 11 as the desorbed waste gas stream 10. The CO and CH4 in the gas are catalytically oxidized into CO2 and H2O and then discharged.

[0049] The dynamic control process is as follows: First online analyzer 14: installed in the inlet pipeline of adsorption tower 6, using a multi-channel NDIR analyzer to measure CO and CO2 concentrations in real time, with a data update cycle of 2 seconds.

[0050] Second online analyzer 15: A TDLAS analyzer is installed before the product gas outlet valve of each adsorption tower 6 to specifically monitor the concentration of CO (the most dangerous impurity due to its strong adsorption properties) in the mixed gas. The sampling system is a high-temperature (~150℃) rapid sampler (used to shorten sampling time, reduce delays, avoid system reaction delays, and ensure representative sampling), ensuring no condensation or adsorption retention in the sample, and a system response time T < 3 seconds. Detection range: 0-50 ppmv, accuracy: ±0.5 ppmv.

[0051] Controller: Configured with a high-performance industrial PC as the host computer 16, running a real-time operating system, and having built-in control algorithms as given in Example 1.

[0052] Adsorbent: The adsorption tower 6 is filled with a special adsorbent combination, with activated alumina (for water removal) at the bottom, activated carbon (for CO2 and CH4 removal) in the middle, and zeolite molecular sieve (for deep removal of CO and N2) at the top.

[0053] Control method implementation (refer to logic principle in Figure 3): Initialization: Under standard feed conditions (CO 3.5%, CO2 5.5%), the optimal adsorption step duration T0 = 480 seconds was determined through debugging to ensure 99.999% purity. The outlet CO concentration change threshold ΔC was set. set , CO = 0.18ppmv, maximum safe duration T max = 550 seconds.

[0054] Establishing a feedforward model: Through fitting experimental data, the bivariate feedforward model was determined to be: T p = 480 × [(3.5 / C t,CO )^0.8 × (5.5 / C t,CO2 )^0.5].

[0055] Dynamic control of a cycle (taking the adsorption step of the A1 tower as an example): Real-time feedforward: The feed C is read by the first online analyzer 14. t,CO = 4.0%, C t,CO2 =5.0%. Calculated T p =480 × [(3.5 / 4.0)^0.8 × (5.5 / 5.0)^0.5] ≈ 452 seconds.

[0056] Start-up and Monitoring: The adsorption of column A1 in the system shown in Figure 1 was started with an expected start time of 452 seconds. Simultaneously, the CO concentration C was collected by the second online analyzer 15 at the outlet of column A1. t,CO,out Initial C t,CO,out 0 ppbv (background value).

[0057] Feedback decision: At 430 seconds, Ct,CO,out The system enters a high alert state when the concentration jumps from 0 ppbv to 0.16 ppmv.

[0058] At 465 seconds, C t,CO,out Reaching 0.19 ppmv, exceeding ΔC set,CO The controller immediately (within milliseconds) issues a command to close the inlet valve of column A1, terminating the adsorption step. Actual adsorption time T actual = 465 seconds. Therefore, the adsorption time of adsorption tower 6 in A2 is adjusted to 465 seconds for adsorption operation.

[0059] Linkage adjustment: Calculate the adsorption load factor for this cycle. T actual (465 seconds) is less than the baseline T0 (480 seconds), but greater than the predicted T. p (452 seconds), according to rule V purge = K × T actual The load factor (proportional coefficient K) is introduced for fine-tuning, and the final calculation shows that the required flushing air volume for this cycle is about 8% higher than the standard value (the standard value is the theoretical calculation value without dynamic adjustment). The controller executes this precisely through the regulating valve.

[0060] Effect comparison (simulated operation for 72 hours, with feed CO concentration fluctuating between 3.0% and 4.5%): Product purity: Using the dynamic control of this invention, the product hydrogen purity curve is a straight line closely following the 99.999% line, with a fluctuation range of ±0.0002%.

[0061] Hydrogen recovery rate: The average recovery rate under dynamic control of this invention is 88.5%.

[0062] In summary, this invention increases the recovery rate by 7.3 percentage points, demonstrating significant benefits.

[0063] Compared to Example 2, Comparative Example 1 did not employ a dynamic control system. The feed status and other system configurations were identical. All valves opened and closed at fixed time points set for each step within a cycle (e.g., adsorption, equalization, forward discharge, reverse discharge, rinsing, pressurization). Under standard feed conditions (CO 3.5%, CO 25.5%), the optimal adsorption step duration T0 = 480 seconds was determined through debugging to ensure 99.999% purity. During simulation operation, the adsorption time T0 = 480 seconds did not adjust automatically based on the feed parameters (i.e., the adsorption time was fixed at 480 seconds for each cycle).

[0064] The simulation runs for 72 hours, and the inlet CO concentration fluctuates between 3.0% and 4.5%. When the CO concentration > 4%, the product hydrogen purity has repeatedly shown a "pin-like" decline and is lower than the target value of 99.99%, reaching a minimum of 99.98%. The impurity CO concentration exceeds the limit value, resulting in unqualified products. When the CO concentration is less than 3.5%, due to the decrease in impurity concentration, and the system sets the adsorption time T0 = 480 seconds to remain unchanged, and the flushing time is not dynamically adjusted either. Although the products are qualified, in fact, due to the reduction in the impurity gas content, only shorter adsorption time and flushing time are required to obtain qualified products. This leads to an average hydrogen recovery rate of only 81.2% during operation.

[0065] In summary, in view of the problem that the existing PSA process using fixed timing control cannot adapt to the fluctuations in feed components, resulting in unstable product purity or decreased recovery rate, the present invention constructs a "feedforward-feedback" dual-loop intelligent control system, achieving a fundamental transformation from "time-based control" to "state-based control". Within the range of ±30% fluctuation in feed concentration, it can maintain the product purity fluctuation less than ±0.2%, while increasing the average recovery rate by 5 - 15%, significantly enhancing the robustness and economy of the PSA process.

[0066] The above description of the embodiments is to enable those of ordinary skill in the art to understand and use the invention. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative efforts. Therefore, the present invention is not limited to the above embodiments, and all improvements and modifications made by those skilled in the art without departing from the scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A dynamic control purification system for hydrogen from multiple process sources, characterized in that, The system includes a mixing and pretreatment unit, a multi-tower PSA separation unit, a product and waste gas collection unit, and a control and execution unit. The mixing and pretreatment unit includes a gas mixer (4), a preheater (5), and a deaerator (13) connected in sequence. The gas mixer (4) receives a coke oven gas hydrogen production stream (1) and an electrolytic water hydrogen production stream (2) at its inlet end. The multi-tower PSA separation unit is connected to the mixing and pretreatment unit and the product and waste gas collection unit via pipelines. The multi-tower PSA separation unit includes at least four parallel adsorption towers (6). The control and execution unit includes a control module (12), a first online analyzer (14), and a second online analyzer (15). The first online analyzer (14) is used to obtain the concentration of at least one key impurity component in the mixed gas entering the adsorption tower (6). The second online analyzer (15) is used to obtain the concentration of at least one key impurity component in the product hydrogen stream (7) discharged from the top of the adsorption tower (6). The control module (12) is communicatively connected to the first online analyzer (14) and the second online analyzer (15).

2. The dynamic control purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The mixing and pretreatment unit also includes proportional regulating valves (3) for adjusting the flow rates of the coke oven gas hydrogen production stream (1) and the electrolytic water hydrogen production stream (2), respectively.

3. The dynamic control and purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The multi-tower PSA separation unit also includes a control valve group (9), and each control valve of the control valve group (9) is installed on the pipeline connected to the adsorption tower (6).

4. The dynamic control and purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The product and waste gas collection unit includes a product buffer tank (8) and a waste gas processor (11); the product buffer tank (8) is connected to the top of the adsorption tower (6) and is used to collect the product hydrogen gas stream (7) obtained after adsorption separation; the waste gas processor (11) is connected to the bottom of the adsorption tower (6) and is used to recover the desorption waste gas stream (10) obtained after desorption separation.

5. The dynamic control purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The control and execution unit also includes a gas concentration sensor (17); multiple gas concentration sensors (17) are arranged at intervals along the height of the adsorption tower (6) to obtain the concentration of at least one key impurity component at different locations in different sections of the adsorption tower (6); the control module (12) is communicatively connected to the gas concentration sensor (17).

6. The dynamic control and purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The control and execution unit also includes a host computer (16); the control module (12) is a PLC controller; the host computer (16) and the control module (12) are connected in communication.

7. The dynamic control and purification system for hydrogen from multiple process sources according to claim 1, characterized in that, The adsorption tower (6) includes a tower body (201), an inlet distributor (202) and a support grid plate (206) disposed at the bottom of the tower body (201), and an outlet collector (207) disposed at the top of the tower body (201); a composite adsorbent is filled on the support grid plate (206) inside the tower body (201) for absorbing and removing water, carbon dioxide, methane, carbon monoxide and nitrogen respectively.

8. A method for dynamically controlled purification of hydrogen from multiple process sources, characterized in that, For dynamic control of the system as described in any one of claims 1 to 7; the method includes the following steps: S1: setting a concentration change threshold ΔC set , i and maximum safe duration T max Based on the standard feed component concentration C0, i Determine the baseline duration T0 for the adsorption step, where i represents the critical impurity component; S2: Obtain the concentration C of at least one critical impurity component in the mixed gas entering the adsorption tower (6). t , i S3: Calculate concentration C using a feedforward compensation model. t , i Predicted adsorption time T p The feedforward compensation model is a function T of the concentration change of the key impurity component and the baseline time of the adsorption step. p = f(C t , i , C0, i , T0); S4: Obtain the outlet gas concentration C of the adsorption tower (6) out , i If the adsorption time t reaches T p C has already been satisfied. out , i - C t , i ≥ ΔC set , i Then the adsorption time t is set to the actual adsorption step time T. actual If the adsorption time t reaches T p C was not satisfied at that time out , i - C t , i ≥ ΔC set , i Then the adsorption time t will be extended by a set time increment Δt until C is satisfied. out , i - C t , i ≥ ΔC set , i Or reach the maximum safe duration T max The adsorption time t is set to the actual adsorption step time T. actual S5: Based on the actual adsorption step duration T actual The duration of the flushing step or the flushing air flow rate, as well as the duration of the pressure equalization step, can be dynamically adjusted.

9. The method for dynamic control and purification of hydrogen from multiple process sources according to claim 8, characterized in that, The feedforward compensation model is: T p = T0× Π[(C0, i / C t , i )^k i ]; where k i This is a correction factor, with a value ranging from 0.5 to 1.

0.

10. The method for dynamic control and purification of hydrogen from multiple process sources according to claim 8, characterized in that, The dynamic adjustment satisfies: total flushing air consumption V purge = K × T actual Where K is the proportionality coefficient.

Images