A hydrogen adaptive separation purification control method and system for a hydrogen production system

Abstract

The application provides a hydrogen adaptive separation and purification control method and system for a hydrogen production system, relates to the technical field of hydrogen separation and purification control, and dynamically adjusts the supply of an alkaline solution to a main reaction container and waste liquid recovery of a storage module of a reaction control subsystem; a collaborative reaction of a silicon-based material and the alkaline solution is carried out in the main reaction container, the alkaline reaction liquid is pumped to a catalytic reaction subsystem; a sensing subsystem monitors the temperature state of the main reaction container and the pressure state of a hydrogen purification and output subsystem in real time, and feeds back data to a controller; the controller dynamically adjusts the catalyst filling rate of a catalyst filling unit and the pumping rate of the alkaline reaction liquid of the main reaction container; the catalyst filling unit adds catalyst to a hydrogen production module, ensures the ratio of the catalyst and the alkaline reaction liquid, the hydrogen production module realizes accelerated reaction of hydrogen by using the filled catalyst and the alkaline reaction liquid, and realizes purification and output of hydrogen by a hydrogen purification and output subsystem.

Description

Technical Field

[0001] This invention proposes an adaptive separation and purification control method and system for hydrogen in hydrogen production systems, relating to the field of hydrogen separation and purification control technology. Background Technology

[0002] Currently, alkaline water hydrogen production technology (ALK) is the most mature industrialized hydrogen production method. Based on KOH / NaOH, it has advantages such as low cost and large single-tank capacity, and is widely used in the large-scale production of green hydrogen. Figure 1 As shown, in existing hydrogen production systems, hydrogen storage tank 1 stores the prepared hydrogen. Multiple hydrogen storage tanks ensure continuous gas supply. Valves 2 and pipelines 3 jointly control the flow direction and flow rate of hydrogen, ensuring safe and efficient operation of the system. Valves 2 are used to open or close the gas flow, and pipelines 3 are used to transport hydrogen. Valves and pipelines are distributed throughout the hydrogen system, connecting various components. Compressor 5 is connected to hydrogen storage tank 1 and is used to compress the prepared hydrogen to the required high pressure for storage and transportation. Control cabinet 4 contains electrical and control systems for monitoring and controlling the operating status of the entire hydrogen production system. It typically includes sensors, controllers, and display devices. Instruments 6 are distributed on pipelines 3 and equipment to monitor parameters such as pressure, temperature, and flow rate in the system, ensuring safe and stable operation. Filters and dryers 7 are installed between compressor 5 and hydrogen storage tank 1 to remove impurities and moisture from the hydrogen, ensuring the purity and quality of the hydrogen. However, existing hydrogen production systems suffer from low operating current density and slow dynamic response in their alkaline tanks, making it difficult to adapt to the volatility of renewable energy sources. The catalyst addition method is usually a fixed ratio, making it difficult to adapt to changes in reaction conditions. The cost of electricity in hydrogen production accounts for a large proportion of the total cost; abnormal temperature or pressure may lead to runaway reaction or equipment damage. Summary of the Invention

[0003] Unlike existing hydrogen production systems, this system is an adaptive hydrogen separation and purification control method and system for hydrogen production. Through the deep integration of mechanical innovation and intelligent control, it achieves efficient, safe, and low-energy-consumption green hydrogen production. The hydrogen production system adopts a modular design, integrating dynamic pressure regulation, waste liquid recycling, hydrogen purification, and intelligent feedback control functions, making it suitable for distributed energy supply, fuel cell hydrogen supply, and other scenarios. It provides a new path for the large-scale production of green hydrogen, aligning with the clean energy transition needs under the global carbon neutrality goal.

[0004] This invention proposes an adaptive separation and purification control method for hydrogen in a hydrogen production system, comprising: The storage module of the reaction control subsystem dynamically regulates the recovery of waste liquid and the supply of alkaline solution to the main reaction vessel; the main reaction vessel carries out a synergistic reaction between silicon-based materials and alkaline solution, and pumps the alkaline reaction liquid after the synergistic reaction to the catalytic reaction subsystem; The sensing subsystem monitors the temperature status of the main reaction vessel and the pressure status of the hydrogen purification and output subsystem in real time, and feeds the data back to the controller; the controller dynamically adjusts the catalyst injection rate of the catalytic reaction subsystem and the pumping rate of the alkaline reaction liquid in the main reaction vessel. Within the catalytic reaction subsystem, the catalyst feeding unit adds catalyst to the hydrogen production module according to the adjusted catalyst feeding rate. The main reaction vessel pumps the alkaline reaction liquid after synergistic reaction to the catalytic reaction subsystem according to the adjusted pumping rate of the alkaline reaction liquid. The hydrogen production module uses the added catalyst and alkaline reaction liquid to accelerate the hydrogen reaction and achieves purified hydrogen output through the hydrogen purification and output subsystem.

[0005] Furthermore, the controller calculates the pressure change rate and temperature change rate, and through a machine learning model, outputs the rate at which the alkaline solution is pumped from the alkaline solution storage to the main reaction vessel and the rate at which the catalyst is added by the catalyst addition unit.

[0006] Furthermore, the sensing subsystem collects the temperature data T(t) of the main reaction vessel and the pressure data P(t) of the hydrogen purification and output subsystem in real time, forming a data time series and calculating the interval time. internal temperature change rate and pressure change rate ; Calculate the average time of the sliding window over the past M seconds and average pressure And calculate the standard deviation of pressure, which measures volatility. and time standard deviation ; Input feature X into the machine learning model: The machine learning model predicts the alkaline solution pumping rate and the catalyst loading rate.

[0007] Furthermore, the hydrogen purity target f1: ; System energy consumption target f2: ; in: Rated pressure, 1.6 MPa: Rated current, 1000A; This is the compression efficiency coefficient.

[0008] Operational stability target f3: ; For pressure standard deviation, This represents the temperature standard deviation.

[0009] Furthermore, the multi-objective problem is transformed into a single-objective problem by introducing dynamic weight coefficients, and a weighted total objective function F is constructed and optimized: ;

[0010] in, These are reference values, i=1,2,3, used for normalization; weighting coefficients. Adjust dynamically according to the operating mode.

[0011] The present invention also proposes a hydrogen adaptive separation and purification control system for implementing the above-mentioned hydrogen adaptive separation and purification control method for hydrogen production systems, comprising: a reaction control subsystem, a catalytic reaction subsystem, a sensing subsystem, and a hydrogen purification and output subsystem; The reaction control subsystem includes: a storage module and a main reaction vessel; the storage module realizes waste liquid recovery and supply of alkaline solution to the main reaction vessel through dynamic balance adjustment of the solution storage tank and the waste liquid storage tank; the main reaction vessel carries out the synergistic reaction of silicon-based materials and alkaline solution, and pumps the alkaline reaction liquid after the synergistic reaction to the catalytic reaction subsystem; The sensor subsystem monitors the temperature status of the main reaction vessel and the pressure status of the hydrogen purification and output subsystem in real time, and feeds the data back to the controller; the controller dynamically adjusts the catalyst injection rate of the catalytic reaction subsystem and the pumping rate of the alkaline reaction liquid in the main reaction vessel. The catalytic reaction subsystem includes a catalyst feeding unit, a controller, and a hydrogen production module. The catalyst feeding unit adds catalyst to the hydrogen production module according to an adjusted catalyst feeding rate. The main reaction vessel pumps the alkaline reaction liquid after synergistic reaction to the catalytic reaction subsystem according to an adjusted alkaline reaction liquid pumping rate. The hydrogen production module uses the added catalyst and alkaline reaction liquid to accelerate the hydrogen reaction and achieves purified hydrogen output through a hydrogen purification and output subsystem.

[0012] Preferably, when the alkaline solution is pumped to the main reaction vessel, the solution storage container is compressed due to the reduction in volume, and the waste liquid storage container expands due to the inflow of waste liquid. The solution storage container and the waste liquid storage container of the alkaline solution achieve dynamic balance adjustment through the linkage of the piston plate and the spring. After the waste liquid storage container removes the alkaline components of the waste liquid using ion exchange resin technology, part of it is returned to the main reaction vessel for recycling.

[0013] Preferably, the hydrogen production module includes: a ball valve, a check valve, a pipeline, a normally open throttle valve, and a gas-liquid separator; the inlet of the hydrogen production module is connected to the catalyst injection unit through the ball valve and the check valve, the liquid channel of the gas-liquid separator is connected to the waste liquid storage tank of the reaction control subsystem, and the gas channel of the gas-liquid separator is connected to the hydrogen purification and output subsystem by controlling the opening degree through the normally open throttle valve to regulate the hydrogen output flow rate.

[0014] Preferably, the hydrogen purification and output subsystem includes: a gas drying device and a multi-stage purification device. The hydrogen initially separated from the gas channel is guided to the gas drying device, which uses a combination of molecular sieve and hydrophobic membrane technology to remove moisture and alkaline mist from the hydrogen. The multi-stage purification device includes a condenser and a precision filter to improve the purity of the hydrogen.

[0015] Preferably, the main reaction vessel is made of corrosion-resistant stainless steel and has a removable filter basket inside for loading silicon-based materials. The co-reaction chamber of the main reaction vessel enables the co-reaction of silicon-based materials with alkaline solution. The liquid inlet valve of the main reaction vessel and the piston plate of the storage module are mechanically coupled through an axial connecting rod to achieve synchronous control of the flow rate of alkaline reaction liquid and the piston movement frequency.

[0016] Compared with the prior art, the present invention has the following beneficial technical effects: 1. Dynamically adjust the supply of alkaline solution and waste liquid recovery. Through real-time monitoring and feedback control, ensure the dynamic balance of alkaline solution concentration in the main reaction vessel, avoid the decrease in reaction rate or the generation of by-products due to excessively high or low concentration, thereby improving the overall reaction efficiency.

[0017] 2. The catalyst dosing rate is adjusted in real time by the controller to ensure the optimal ratio of catalyst to reaction liquid, avoid the waste of excessive catalyst or the decrease in reaction efficiency caused by insufficient catalyst, and extend the catalyst life.

[0018] 3. The sensing subsystem monitors temperature and pressure in real time, and combines temperature and pressure sensing data to dynamically adjust the pumping rate and catalyst dosing rate to adapt to fluctuations in reaction conditions, thereby improving the robustness and flexibility of the system. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the structure of a hydrogen production system in the prior art; Figure 2 Here is a flowchart of the reaction control subsystem. Figure 3 This is a schematic diagram of an axial connecting rod. Figure 4 This is a flowchart of the catalytic reaction subsystem. Figure 5 This is a flowchart of the sensing subsystem's workflow. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the accompanying drawings of specific embodiments of the present invention, in order to better and more clearly describe the working principle of each component in the system and show the connection relationship of each part in the device, only the relative positional relationship between each component is clearly distinguished. It does not constitute a limitation on the signal transmission direction, connection sequence, or size and shape of each part within the component or structure.

[0023] Example 1 The adaptive separation and purification control system of the present invention includes: a reaction control subsystem, a catalytic reaction subsystem, a sensing subsystem, and a hydrogen purification and output subsystem. The reaction control subsystem and the sensing subsystem are the core of dynamic balance and intelligent regulation. The catalytic reaction subsystem is used for efficient hydrogen production and pressure management. The hydrogen purification and output subsystem performs multi-stage purification and safe output of hydrogen.

[0024] like Figure 2 As shown, the reaction control subsystem includes: a storage module and a main reaction vessel. The storage module includes: a solution storage tank for the alkaline solution and a waste liquid storage tank.

[0025] When the alkaline solution is pumped to the main reaction vessel, the solution reservoir compresses due to volume reduction, while the waste liquid reservoir expands due to waste liquid inflow, creating a dynamic balance. The solution and waste liquid reservoirs are dynamically balanced through the linkage of a piston plate and a spring, and the spring's elasticity compensation ensures stable system operation under different loads.

[0026] The main reaction vessel includes: a filter basket and a co-reaction chamber.

[0027] The main reaction vessel is made of corrosion-resistant stainless steel and has a removable filter basket inside to hold silicon-based materials. The co-reaction chamber works with the catalyst dosing unit to achieve the synergistic reaction between silicon-based materials and alkaline solution, and pumps alkaline reaction liquid into the catalytic reaction subsystem.

[0028] The inlet valve of the main reaction vessel is connected to the piston plate of the storage module via, for example... Figure 3The axial linkage mechanical coupling is shown; when the piston plate is compressed, it pushes the linkage to open the inlet valve, and when it expands, the linkage resets and closes the valve, achieving synchronous control of the alkaline reaction liquid flow rate and the piston movement frequency, thus regulating the supply rate of the synergistic reaction liquid and the return path of the waste liquid. When the piston plate compresses the solution storage tank, the alkaline solution enters the main reaction vessel through the pipeline; simultaneously, the expansion of the waste liquid storage tank reserves space for subsequent waste liquid recovery, forming a dynamic balance.

[0029] In a preferred embodiment, after the alkaline components are removed using ion exchange resin technology in the waste liquid storage tank, a portion can be returned to the main reaction vessel for recycling, which not only reduces raw material consumption but also lowers environmental pollution. Simultaneously, the counterweight device optimizes the energy loss of the waste liquid return path.

[0030] like Figure 4 As shown, the catalytic reaction subsystem includes a catalyst dosing unit, a controller, and a hydrogen production module.

[0031] The sensing subsystem monitors the temperature status of the main reaction vessel and the pressure status of the hydrogen purification and output subsystem in real time, and feeds the data back to the controller; the controller dynamically adjusts the catalyst injection rate of the catalytic reaction subsystem and the pumping rate of the alkaline reaction liquid in the main reaction vessel.

[0032] The hydrogen production module controls the catalyst dosing rate via a controller. The catalyst dosing unit consists of a catalyst storage tank, a precision metering pump, and a piping system, enabling precise catalyst addition with a control accuracy of ±0.5 ml / min. This ensures that the ratio of the alkaline reaction solution to the catalyst in the hydrogen production module conforms to the stoichiometric ratio, avoiding localized over-concentration or under-concentration.

[0033] The hydrogen production module includes: ball valve, check valve, pipeline, normally open throttle valve and gas-liquid separator.

[0034] The hydrogen production module adopts a tubular spiral structure and is directly connected to the main reaction vessel of the reaction control subsystem. The alkaline reaction solution generated in the main reaction vessel is transported to the tubular spiral structure of the hydrogen production module through pipelines. The highly active nickel-based catalyst filled inside the hydrogen production module further accelerates the reaction and improves the hydrogen generation efficiency. Preferably, the highly active nickel-based catalyst is foamed nickel or nickel mesh sprayed catalyst, which can significantly improve the hydrolysis reaction efficiency.

[0035] The hydrogen production module inlet is connected to the catalyst feeding unit via a ball valve and a check valve to ensure that the catalyst is supplied as needed. The liquid channel of the gas-liquid separator at the outlet of the hydrogen production module is connected to the waste liquid storage tank of the reaction control subsystem, and the separated hydrogen enters the subsequent purification process. The gas channel of the liquid separator is controlled by a normally open throttle valve and is connected to the hydrogen purification and output subsystem to regulate the hydrogen output flow rate.

[0036] The hydrogen purification and output subsystem includes: a gas drying device and a multi-stage purification device.

[0037] After the hydrogen production module of the catalytic reaction subsystem completes the hydrolysis reaction, the generated hydrogen first enters the gas-liquid separator. The initially separated hydrogen is then guided to the gas drying unit. The gas drying unit uses a combination of molecular sieve and hydrophobic membrane technology to remove moisture and trace amounts of alkaline mist from the hydrogen. The multi-stage purification device includes a condenser and a precision filter, increasing the hydrogen purity to over 99.9%.

[0038] Example 2 Hydrogen adaptive separation and purification control method I. System Initialization and Dynamic Balancing Establishment Check the level of alkaline solution in the solution storage tank and the empty status of the waste liquid storage tank. Confirm that the catalyst storage tank is filled with highly active nickel-based catalyst. The alkaline solution is preferably 30% KOH, and the nickel-based catalyst is foamed nickel or nickel mesh sprayed type.

[0039] The pump in the storage module is turned on, and the alkaline solution is delivered to the main reaction vessel through the pipeline. The volume of the solution storage tank shrinks, driving the piston plate to move axially and compressing the internal spring to store energy; at the same time, the waste liquid storage tank expands synchronously to reserve space for waste liquid, and the spring force compensates for load fluctuations in real time, ensuring that the flow rate of alkaline solution is stable at 5–10 L / min.

[0040] The piston plate, through an axial connecting rod, links the inlet valve of the main reaction vessel to adjust the supply rate of the alkaline reaction solution and the waste liquid return path, thus forming an initial dynamic equilibrium.

[0041] II. Catalytic hydrogen production reaction main reaction vessel coordinated reaction An alkaline solution enters the stainless steel main reaction vessel and flows through a removable filter basket loaded with silicon-based material. The system uses the alkaline solution to catalyze the hydrolysis of the silicon-based material to produce hydrogen. Figure 5 As shown, the sensor subsystem monitors the temperature of the main reaction vessel and the pressure of the hydrogen purification and output subsystem in real time. The sensor subsystem feeds the data back to the controller for data link linkage. Based on the data from the sensor subsystem, the controller dynamically adjusts the rate at which the alkaline solution in the storage module is pumped into the main reaction vessel from the alkaline solution storage unit using a machine learning model, while simultaneously adjusting the rate at which the catalyst is added by the catalyst addition unit.

[0042] Specifically, the controller calculates the pressure change rate and temperature change rate, and through a machine learning model, obtains the predicted alkaline solution pumping rate as the rate at which the alkaline solution is pumped from the alkaline solution storage to the main reaction vessel, and the predicted catalyst charging rate as the rate at which the catalyst charging unit charges the catalyst, and outputs pumping and charging control parameters.

[0043] Specifically, the method is as follows: (1) Data collection The sensing subsystem collects the temperature T(t) of the main reaction vessel and the pressure P(t) of the normally open throttle valve in real time, with a sampling frequency of Δt = 1 second, forming a data time series of n seconds: [T1, T2, ..., T n [P1,P2,...,P] n ].

[0044] (2) Extraction interval Rate of change of key features within Rate of temperature change: ; Pressure change rate: ; Time series feature extraction Calculate the average time of the sliding window over the past M seconds and average pressure : ; ; Calculate the standard deviation of pressure, which measures volatility. and time standard deviation : ; ; Where i represents the i-th second in the past.

[0045] Interval time The value range is 0.5s to 2s, with Δt=1s being preferred. This range balances data real-time performance with computational load. In particular, the 1s interval can accurately capture the dynamic changes in temperature and pressure during the reaction process without overloading the controller's computing power due to excessive sampling.

[0046] The sliding window duration M can range from 30s to 120s, with M=60s being the preferred value. 30s or more can cover the dynamic response cycle of the reaction system, while 60s is the optimal value, reflecting parameter trends while avoiding data redundancy; if the reaction system fluctuates significantly, it can be adjusted to 120s.

[0047] The temperature T acquisition accuracy is ±0.1℃, and the pressure P acquisition accuracy is ±0.001MPa; the sampling frequency of the sensing subsystem is... Consistency ensures the continuity of the data time series.

[0048] (3) Training the machine learning model The machine learning model uses a multilayer perceptron neural network, the structure of which includes: Input layer: 6 nodes, corresponding to feature vectors .

[0049] Hidden layer: 10 nodes, using the ReLU activation function.

[0050] Output layer: 2 nodes, outputting the predicted alkaline solution pumping rate. and the predicted catalyst loading rate All units are mL / min.

[0051] Data is collected from historical system operation data or simulation data, including temperature and pressure sequences, as well as corresponding optimal pumping and filling rates. The data set comprises at least 1000 samples, consisting of measured data from the hydrogen production system under different operating conditions: covering the entire operating range from 60℃ to 95℃ and system pressure from 0.8MPa to 1.6MPa; data is continuously collected for 24 hours under each operating condition, resulting in at least 1000 sets of training data to ensure the model's generalization ability.

[0052] The network weights were optimized using backpropagation and stochastic gradient descent, with a learning rate of 0.01 and 1000 training iterations.

[0053] The improved loss function L is: ; in: This is the weighting coefficient for the prediction error of the alkaline solution pumping rate. These are weighting coefficients for the catalyst dosing rate prediction error. These two weighting coefficients are used to balance the importance of the two different output variables and can be adjusted according to actual control requirements. If the catalyst cost is high, they can be appropriately increased. To improve the accuracy of catalyst loading rate prediction, To represent the actual pumping rate of alkaline solutions, The actual catalyst loading rate is given, where N is the total optimization time and t is the time number. The predicted alkaline solution pumping rate is obtained by optimizing the machine learning model. and the predicted catalyst loading rate , This is the regularization coefficient, which controls the penalty strength for model complexity; For model parameters The L2 norm represents the sum of squares of all neural network weight parameters.

[0054] After training is complete, the model weights are fixed and deployed in the controller.

[0055] Online inference: The controller calculates the feature vector X in real time, inputs it into the trained model, and obtains the predicted pumping rate of the alkaline solution. and the predicted catalyst loading rate .

[0056] Based on the above-predicted alkaline solution pumping rate and the predicted catalyst loading rate The catalyst in the main reaction vessel accelerates the reaction, achieving gas-liquid separation and waste liquid circulation control. After the waste liquid is dealkalized by ion exchange resin, 50-70% is recycled. The hydrogen produced by the reaction enters the gas-liquid separator and then enters the hydrogen purification and output subsystem.

[0057] III. Hydrogen Purification and Safe Output The initially separated hydrogen gas will be guided to a gas drying unit. The gas drying unit uses a combination of molecular sieve and hydrophobic membrane technology to remove 99% of the moisture and alkaline mist. Afterward, the gas enters a condenser to condense trace amounts of water vapor before passing through a precision filter, increasing the hydrogen purity to over 99.6%. The purified hydrogen is then metered by a flow meter and delivered to a hydrogen storage tank. Table 1 shows a comparison of the three water-to-hydrogen technologies.

[0058] Table 1 Comparison of three hydrogen production technologies

[0059] Example 3 This embodiment, based on Embodiments 1 and 2, proposes a dynamic impurity accumulation model. Existing hydrogen production systems all employ steady-state gas purity models, which can only predict the hydrogen purity at the condenser outlet. Therefore, a dynamic impurity accumulation model is needed to predict the gas purity at the condenser outlet of the alkaline water hydrogen production unit. The dynamic impurity accumulation model proposed in this embodiment aims to broaden the load range of the alkaline water hydrogen production system by reducing system pressure during low-load periods.

[0060] Since cross-contamination of gaseous impurities can easily form flammable gas mixtures, and the minimum load of alkaline water hydrogen production units is usually limited to 10% to 40% of the rated current, which hinders the operation of the unit during low-load periods, a dynamic impurity accumulation model was established to improve the flexibility of the alkaline water hydrogen production system.

[0061] As shown in the alkaline hydrogen production process, oxygen is produced at the anode and hydrogen at the cathode. The gas output from the cathode contains trace amounts of oxygen, which is not completely removed during the hydrogen production reaction and enters the condenser along with the gas. This also means that the change in the volume fraction of hydrogen is a dynamic process that accumulates over time; as the current density increases, the steady-state value of the oxygen volume fraction decreases.

[0062] In the reaction control subsystem, a dynamic impurity accumulation model is used to predict the gas purity at the condenser outlet, considering the accumulation effect of impurities over time. The dynamic impurity accumulation model is constructed using differential equations: ; in: The oxygen concentration in the hydrogen gas at the condenser outlet Rate of change over time, let Q impurity for , representing the impurity generation term; V condenser The volume of the condenser is in L. The number of moles of oxygen produced per unit current. Real-time load current (A).

[0063] Represents pressure dissolution and removal items. The dissolution and removal coefficient (1 / s) is the dissolution and removal coefficient. The system pressure is measured in MPa. System temperature (K); For physical separation and removal items, is the physical separation rate constant (1 / s).

[0064] When the dynamic impurity accumulation model reaches steady state The predicted steady-state value of oxygen concentration is... for: ; This steady-state solution clearly shows that the impurity concentration is directly proportional to the load current, inversely proportional to the system pressure, and directly proportional to the temperature.

[0065] Impurities mainly originate from cross-diffusion of gases on both sides, and their rate is related to the pressure difference, temperature, and diaphragm characteristics, affecting the impurity input rate Q. impurity Perform the calculation: ; Where: D is the diffusion coefficient (m 2 / s); A is the diaphragm area (m²) 2 ); ΔP is the pressure difference between the hydrogen and oxygen sides (Pa); δ is the membrane thickness (m).

[0066] This embodiment reduces system pressure during low-load periods, minimizes cross-contamination of impurities, and broadens the load range.

[0067] Specifically, a multi-objective function system is constructed.

[0068] Hydrogen purity target f1: ; System energy consumption target f2: ; in: Rated pressure, 1.6 MPa: Rated current, 1000A; This is the compression efficiency coefficient.

[0069] Operational stability target f3: ; For pressure standard deviation, This represents the temperature standard deviation.

[0070] The multi-objective problem is transformed into a single-objective problem, and dynamic weight coefficients are introduced to construct and optimize the weighted total objective function F: ; in, These are reference values, i=1,2,3, used for normalization; weighting coefficients. Dynamically adjust according to the operating mode: High purity mode: ; Energy-saving mode: ; Stable operating mode: .

[0071] This embodiment optimizes three objectives simultaneously: hydrogen purity, system energy consumption, and operational stability. This avoids performance imbalances that may result from optimizing a single objective, achieving optimal coordination of overall system performance. By adjusting dynamic weighting coefficients, the system can automatically adjust its control strategy according to different high-purity, energy-saving, and stable operation modes, adapting to the needs of various application scenarios. Multi-objective optimization considers the system's dynamic characteristics, namely pressure and temperature standard deviations, which helps improve the system's stable operation under external disturbances.

[0072] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered 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 adaptive separation and purification control of hydrogen in a hydrogen production system, characterized in that, include: The storage module of the reaction control subsystem dynamically regulates the recovery of waste liquid and the supply of alkaline solution to the main reaction vessel; the main reaction vessel carries out a synergistic reaction between silicon-based materials and alkaline solution, and pumps the alkaline reaction liquid after the synergistic reaction to the catalytic reaction subsystem; The sensing subsystem monitors the temperature status of the main reaction vessel and the pressure status of the hydrogen purification and output subsystem in real time, and feeds the data back to the controller; the controller dynamically adjusts the catalyst injection rate of the catalytic reaction subsystem and the pumping rate of the alkaline reaction liquid in the main reaction vessel. The controller calculates the pressure change rate and temperature change rate, and through a machine learning model, outputs the rate at which the alkaline solution is pumped from the alkaline solution storage to the main reaction vessel and the rate at which the catalyst is added by the catalyst addition unit. The controller calculates the feature vector X in real time, uses a multilayer perceptron neural network to construct an improved loss function L, optimizes the machine learning model, and obtains the predicted alkaline solution pumping rate. and the predicted catalyst loading rate ; The improved loss function L is: ； in: This is the weighting coefficient for the prediction error of the alkaline solution pumping rate. The weighting coefficients for the prediction error of the catalyst loading rate. To represent the actual pumping rate of alkaline solutions, The actual catalyst loading rate is given, N represents the total optimization time, and t represents the time number. The regularization coefficient is . For model parameters The L2 norm; Within the catalytic reaction subsystem, the catalyst feeding unit adds catalyst to the hydrogen production module according to the adjusted catalyst feeding rate. The main reaction vessel pumps the alkaline reaction liquid after synergistic reaction to the catalytic reaction subsystem according to the adjusted alkaline reaction liquid pumping rate. The hydrogen production module uses the added catalyst and alkaline reaction liquid to accelerate the hydrogen reaction and achieves purified hydrogen output through the hydrogen purification and output subsystem.

2. The adaptive separation and purification control method for hydrogen in a hydrogen production system according to claim 1, characterized in that, The sensing subsystem collects real-time temperature data T(t) from the main reaction vessel and pressure data P(t) from the hydrogen purification and output subsystem, forming a data time series and calculating the interval time. internal temperature change rate and pressure change rate : ； ； Calculate the average time of the sliding window over the past M seconds and average pressure : ; ; Calculate the standard deviation of pressure, which measures volatility. and time standard deviation : ; ; Where i represents the i-th second in the past.

3. The adaptive separation and purification control method for hydrogen in a hydrogen production system according to claim 2, characterized in that, The machine learning model uses a multilayer perceptron neural network: Input layer: 6 nodes, corresponding to feature vectors ; Hidden layer: 10 nodes, using the ReLU activation function; Output layer: 2 nodes, outputting the alkaline solution pumping rate and catalyst dosing rate respectively.

4. The adaptive separation and purification control method for hydrogen in a hydrogen production system according to claim 2, characterized in that, The interval time Δt ranges from 0.5s to 2s, and the sliding window duration M ranges from 30s to 120s.

5. The adaptive separation and purification control method for hydrogen in a hydrogen production system according to claim 2, characterized in that, The machine learning model is pre-trained as follows: a dataset containing temperature sequences, pressure sequences, and corresponding optimal pumping and filling rates is collected from historical system operation data or simulation data. The dataset contains at least 1000 samples, covering the operating conditions of reaction temperature from 60℃ to 95℃ and system pressure from 0.8 MPa to 1.6 MPa. Data is continuously collected for 24 hours under each operating condition. The network weights are optimized using the backpropagation algorithm and stochastic gradient descent method, with a learning rate set to 0.01 and 1000 training iterations.

6. A hydrogen adaptive separation and purification control system, characterized in that, The method for adaptive separation and purification control of hydrogen in a hydrogen production system as described in any one of claims 1-5 includes: a reaction control subsystem, a catalytic reaction subsystem, a sensing subsystem, and a hydrogen purification and output subsystem. The reaction control subsystem includes: a storage module and a main reaction vessel; the storage module realizes waste liquid recovery and supply of alkaline solution to the main reaction vessel through dynamic balance adjustment of the solution storage tank and the waste liquid storage tank; the main reaction vessel carries out the synergistic reaction of silicon-based materials and alkaline solution, and pumps the alkaline reaction liquid after the synergistic reaction to the catalytic reaction subsystem; The sensor subsystem monitors the temperature status of the main reaction vessel and the pressure status of the hydrogen purification and output subsystem in real time, and feeds the data back to the controller; the controller dynamically adjusts the catalyst injection rate of the catalytic reaction subsystem and the pumping rate of the alkaline reaction liquid in the main reaction vessel. The catalytic reaction subsystem includes a catalyst feeding unit, a controller, and a hydrogen production module. The catalyst feeding unit adds catalyst to the hydrogen production module according to an adjusted catalyst feeding rate. The main reaction vessel pumps the alkaline reaction liquid after synergistic reaction to the catalytic reaction subsystem according to an adjusted alkaline reaction liquid pumping rate. The hydrogen production module uses the added catalyst and alkaline reaction liquid to accelerate the hydrogen reaction and achieves purified hydrogen output through a hydrogen purification and output subsystem.

7. The hydrogen adaptive separation and purification control system according to claim 6, characterized in that, When the alkaline solution is pumped to the main reaction vessel, the solution reservoir is compressed due to the decrease in volume, and the waste liquid reservoir expands due to the inflow of waste liquid. The solution reservoir and the waste liquid reservoir of the alkaline solution are dynamically balanced through the linkage of the piston plate and the spring. After the waste liquid reservoir uses ion exchange resin technology to remove the alkaline components of the waste liquid, part of it is returned to the main reaction vessel for recycling.

8. The hydrogen adaptive separation and purification control system according to claim 6, characterized in that, The hydrogen production module includes: a ball valve, a check valve, pipelines, a normally open throttle valve, and a gas-liquid separator; the inlet of the hydrogen production module is connected to the catalyst injection unit through the ball valve and the check valve, the liquid channel of the gas-liquid separator is connected to the waste liquid storage tank of the reaction control subsystem, and the gas channel of the gas-liquid separator is connected to the hydrogen purification and output subsystem by controlling the opening degree through the normally open throttle valve to regulate the hydrogen output flow rate.

9. The hydrogen adaptive separation and purification control system according to claim 8, characterized in that, The hydrogen purification and output subsystem includes a gas drying device and a multi-stage purification device. The hydrogen initially separated from the gas channel is guided to the gas drying device, which uses a combination of molecular sieve and hydrophobic membrane technology to remove moisture and alkaline mist from the hydrogen. The multi-stage purification device includes a condenser and a precision filter to improve the purity of the hydrogen.

10. The hydrogen adaptive separation and purification control system according to claim 6, characterized in that, The main reaction vessel is made of corrosion-resistant stainless steel and has a removable filter basket inside to hold silicon-based materials. The co-reaction chamber of the main reaction vessel enables the co-reaction of silicon-based materials and alkaline solution. The liquid inlet valve of the main reaction vessel and the piston plate of the storage module are mechanically coupled through an axial connecting rod to achieve synchronous control of the flow rate of alkaline reaction liquid and the piston movement frequency.

Images