A photovoltaic-driven continuous air water extraction and hydrogen production integrated system and method based on intelligent adaptive control

Abstract

A photovoltaic-driven continuous air water taking and hydrogen production integrated system and method based on intelligent adaptive control, the system comprising a photovoltaic module, a continuous multi-adsorption bed, a thermoelectric temperature control module, a built-in air pump, a thermal insulation layer, a condensation water collection module and an electrolytic hydrogen production module; when working, after the adsorption bed adsorbs atmospheric moisture, the top cover is pressed down to form a sealed desorption chamber, the thermoelectric temperature control module adopts the PWMEKFMPC strategy for precise heating, the built-in air pump forcibly transports water vapor, the water vapor enters the condensation module through the heat preservation pipeline to produce water, and the condensed water is supplied to the PEM electrolytic cell for hydrogen production; the intelligent control module is based on Bayesian optimization and adaptively adjusts the adsorption-desorption cycle and heating target according to environmental parameters; compared with the prior art, the sealing performance is improved by replacing the double springs with a shaft sleeve, the mass transfer efficiency of water vapor is improved by about 6400 times by the air pump and the thermal insulation layer, the temperature control response time is shortened by more than 60%, the annual water production is about 46.3 L, the annual hydrogen production is about 58 Nm3, and efficient self-sustaining acquisition of water resources and clean energy in arid areas is realized.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of comprehensive utilization of renewable energy and intelligent control technology. Specifically, they relate to a photovoltaic-driven continuous air intake and water production hydrogen production integrated system and method based on intelligent adaptive control, which is particularly suitable for water resource acquisition and clean energy production in areas with abundant solar energy resources and in scenarios without grid coverage. Background Technology

[0002] Photovoltaic power generation technology converts sunlight into electricity using solar cells, offering advantages such as being clean, pollution-free, and suitable for distributed deployment, especially in remote areas without grid coverage. Electrolysis hydrogen production technology converts electricity into hydrogen through water electrolysis, with alkaline electrolysis technology being widely used due to its mature process and low cost. Adsorption-based air-to-water extraction technology enriches atmospheric moisture and generates high-temperature water vapor through an adsorption-desorption cycle using adsorbent materials, which is then condensed into liquid water. Integrating these three technologies enables self-sustaining clean energy production powered by solar energy and using air as a water source in arid regions and scenarios without grid coverage, demonstrating significant application value.

[0003] In the prior art, patent application number PCN25003224 proposes an integrated system combining photovoltaic power generation, continuous air-to-water extraction, and electrolytic hydrogen production. This system includes a photovoltaic power generation module, a continuous multi-adsorption bed module, a thermoelectric temperature control module, a condensation and water collection module, and an electrolytic hydrogen production module. The continuous multi-adsorption bed module employs multiple rotating adsorption beds that alternately perform adsorption and desorption. A first spring is installed on the spline shaft between the top cover and the rotating adsorption bed, and a second spring is installed between the rotating adsorption bed and the base plate. A thrust linear motor pushes the top cover down, causing the adsorbent material to come into contact with the thermoelectric cooler, completing the heating and desorption process. The water vapor generated during desorption diffuses naturally into the condensation and water collection module to condense and produce water. The resulting liquid water is then electrolyzed in a PEM electrolyzer to produce hydrogen.

[0004] However, the above-mentioned device still has the following shortcomings: Firstly, springs are installed on both the top and bottom of the spline shaft between the top cover and the rotary adsorption bed. The stiffness coefficients of the two springs must be precisely matched. Otherwise, the rotary adsorption bed will be subjected to uneven force and deform, leading to the failure of the desorption chamber seal and affecting the reliability of the system.

[0005] Secondly, the water vapor generated in the desorption chamber enters the condensation and water collection module only through natural diffusion, resulting in an extremely low mass transfer rate. Furthermore, the outer wall of the desorption chamber and the connecting pipes lack insulation design, causing a large amount of water to condense and precipitate before reaching the condensation and water collection module, significantly reducing the water production.

[0006] Third, the device uses a thermoelectric cooler to heat the adsorption bed at a constant power. Since the system involves a hydrogel phase change process, the heating parameters change dynamically with the operating conditions. Constant power heating results in a slow temperature rise. There is almost no effective water production in the first 10 minutes after heating begins, and the heating power is difficult to determine precisely, which poses a risk of damaging the adsorption material due to excessively high temperature.

[0007] Fourth, the device uses a fixed time ratio (desorption:adsorption = 1:3) to control the switching between adsorption and desorption, which cannot adapt to the dynamic changes in ambient temperature and humidity: in low humidity environments, the adsorption material reaches adsorption equilibrium quickly, and after adsorption is completed, it still needs to wait for the fixed cycle to end before switching can be done, resulting in wasted time; in high humidity environments, the adsorption material is forced to switch to the desorption stage before it is fully saturated, resulting in insufficient water production and low overall water production efficiency. Summary of the Invention

[0008] The purpose of this invention is to provide an integrated system and method for photovoltaic-driven continuous air intake and water production based on intelligent adaptive control. By replacing the double springs with bushings to improve sealing, and by using an air pump and insulation layer to improve water vapor mass transfer efficiency, this invention overcomes to some extent the problems caused by limitations and defects in related technologies, such as low water vapor mass transfer efficiency, insufficient pipeline insulation, poor sealing, low temperature control accuracy, and inability to adaptively adjust the adsorption-desorption cycle. This results in a significant improvement in the system's water production and hydrogen production efficiency.

[0009] A photovoltaic-driven continuous air-to-water hydrogen production integrated system based on intelligent adaptive control includes: The system comprises: a photovoltaic power generation module A, a continuous multi-adsorption bed module B, a thermoelectric temperature control module C, an air pump D, an insulation layer E, a condensate collection module F, an electrolytic hydrogen production module G, and an intelligent control module H; characterized in that: The photovoltaic power generation module A is used to provide electrical energy to the system and is electrically connected to the thermoelectric temperature control module C, the drive component of the continuous multi-adsorption bed module B, the air pump D, the condensate collection module F, the electrolytic hydrogen production module G, and the intelligent control module H. The continuous multi-adsorption bed module B is used for the adsorption and desorption of atmospheric moisture. In the desorption state, it is connected to the air pump D and the condensate collection module F through the air path. The thermoelectric temperature control module C is used to heat the continuous multi-adsorption bed module (B) in the desorption state to generate water vapor; The air pump D is installed inside the desorption chamber 22 of the continuous multi-adsorption bed module B, and is used to force the water vapor generated by desorption to the condensation and water collection module F. The insulation layer E covers the outer wall of the desorption chamber 22 and the outer wall of the connecting pipe between the desorption chamber and the condensate collection module F. The condensation and water collection module F is used to condense the received water vapor into liquid water and collect it, and its outlet is connected to the water inlet of the electrolysis hydrogen production module G. The electrolysis hydrogen production module G is used to electrolyze the received liquid water to produce hydrogen. The intelligent control module H is connected to the continuous multi-adsorption bed module B, the thermoelectric temperature control module C, the gas pump D, the condensate collection module F, and the electrolysis hydrogen production module G, respectively, and adaptively adjusts the adsorption-desorption switching timing and the target heating temperature according to environmental parameters and system operating status.

[0010] The photovoltaic power generation module A includes a photovoltaic panel 14, a maximum power point tracking device 15, a battery 16, and a DC-DC converter 17; The output terminal of the photovoltaic panel 14 is electrically connected to the input terminal of the maximum power point tracking device 15. The output terminal of the maximum power point tracking device 15 is divided into three paths: The first circuit is bidirectionally connected to battery 16 to achieve energy storage and deficit compensation. The second path is electrically connected to the power supply terminal of the thermoelectric temperature control module C via DC-DC converter 17; The third path is electrically connected to the power supply terminal of the electrolysis hydrogen production module G via the controlled switch tube 20; The maximum power point tracking device 15 is also electrically connected to the power supply terminals of the drive motor, air pump D, and intelligent control module H of the continuous multi-adsorption bed module B, respectively.

[0011] The continuous multi-adsorption bed module B includes a base plate 9, a rotary adsorption bed 3, a top cover 2, a spline shaft 6, a thrust linear motor 7, a rotary servo motor 11, a bushing 5, and a spring 10. The base plate 9 is set horizontally, and the rotary servo motor 11 is fixed below the base plate 9 and its output shaft is coaxially and fixedly connected to the lower end of the spline shaft 6 that runs vertically through the base plate 9. The rotary adsorption bed 3 consists of multiple sector-shaped beds evenly distributed around the spline shaft 6 as the center. Each sector-shaped bed is separated from the others and is sleeved on the spline shaft 6 to form a spline fit with the spline shaft 6. The rotary servo motor 11 drives the rotary adsorption bed 3 to rotate as a whole through the spline shaft 6. The top cover 2 is located above the rotary adsorption bed 3 and is slidably sleeved on the spline shaft 6. The thrust linear motor 7 is fixed above the top cover 2 and its push rod is connected to the upper surface of the top cover 2. A fixed-height bushing 5 is fitted on the spline shaft 6 between the top cover 2 and the rotary adsorption bed 3. The height of the bushing 5 is equal to the required sealing distance of the desorption chamber 22, which is used to limit the downward position when the top cover 2 is pressed down. A spring 10 is fitted on the spline shaft 6 between the rotary adsorption bed 3 and the base plate 9, which is used to reset the rotary adsorption bed 3 and keep it in contact with the base plate 9 after the top cover 2 is raised. When the top cover 2 is pressed down to the limit position of the bushing 5, the sealing sidewall extending downward from the lower surface of the top cover 2 fits against the upper surface and circumferential edge of the corresponding single fan-shaped rotating adsorption bed 3, together forming a sealed desorption chamber 22.

[0012] Each sector-shaped surface of the rotary adsorption bed 3 is fixed with a high thermal conductivity copper sheet 4, and the high thermal conductivity copper sheet 4 is loaded with water-absorbing adsorption material. The water-absorbing adsorbent material is PHEA-LiCl hydrogel; The desorption chamber 22 is equipped with a gas pump D. The outlet of the gas pump D is connected to the air inlet of the condensate collection module F through a water vapor outlet 8 opened on the top cover 2 and an insulation pipe.

[0013] The thermoelectric temperature control module C includes an electric heating element 12, a K-type thermocouple 13, a PWM control unit, and a HIP4081 power amplifier module; The electric heating element 12 is horizontally positioned below the rotating adsorption bed 3 and fixed to the base plate 9 for heating the water-absorbing adsorption material above it. K-type thermocouple 13 is attached to the lower surface of the high thermal conductivity copper sheet 4 at the bottom of the rotating adsorption bed 3 in the desorption state, for real-time acquisition of the temperature of the adsorption material. The signal output terminal of the K-type thermocouple 13 is electrically connected to the temperature signal input terminal of the intelligent control module H. The PWM control signal output terminal of the intelligent control module H is connected to the power supply terminal of the electric heating element 12 via the PWM control unit and the HIP4081 power amplifier module in sequence, forming a temperature closed-loop control circuit.

[0014] The thermoelectric temperature control module C includes an electric heating element 12 horizontally positioned below the rotating adsorption bed 3, used to heat the water-absorbing adsorbent material 4 on the surface of the high thermal conductivity copper sheet 4; a K-type thermocouple is provided between the electric heating element 12 and the rotating adsorption bed 3, and its temperature measuring end is in direct contact with the bottom of the rotating adsorption bed 3 in the desorption state, used to collect the bottom temperature of the adsorbent material 4 in real time.

[0015] The thermoelectric temperature control module C models the controlled object, consisting of the electric heating element 12, the K-type thermocouple 13, the high thermal conductivity copper sheet 4, the water-absorbing adsorption material, and the desorption chamber 22, as a transfer function of a first-order inertial plus pure time-delay element:

[0016] in s For the Laplace operator, K is the gain coefficient, T is the time constant, and τ is the pure time delay; The intelligent control module H uses an extended Kalman filter algorithm to identify time-varying parameters K, T, and τ online. It takes the real-time heating power uk of the electric heating element (12) as the input, the temperature yk collected by the K-type thermocouple as the observation, and the time-varying parameters K, T, and τ as the state variable θk, updated according to the formula: ; Recursive update, where Kk is the Kalman gain and h(·) is the observation function. Adaptive weighting coefficients bound to the real-time excitation intensity of the system: in steady state Approaching 0, during dynamic transition Approaching 1; Based on the identified parameters, a model predictive control algorithm is used to solve the optimization problem in each control cycle:

[0017]

[0018] in, To predict the time domain, Set the temperature value. This is the maximum temperature that the desorption chamber can withstand. The rated power of the electric heating element is given by Q and R, which are weighting coefficients. Solving the optimization problem yields the optimal heating power sequence that minimizes the objective function within the prediction time domain. Following the rolling time-domain control principle, the thermoelectric temperature control module (C) only takes the optimal power value for the current control cycle, converts it into a PWM signal with the corresponding duty cycle, and outputs it to the HIP4081 power amplifier module. In the next control cycle, the problem is resolved based on updated identification parameters. The temperature setpoint in the optimization problem is the heating target temperature adaptively output by the intelligent control module (H).

[0019] The condensate collection module F includes a condensate chamber and a water collection tank; The inner wall of the condensing chamber is provided with several parallel triangular pyramidal fins, and the spaces between adjacent fins are filled with hydrophilic material. The air inlet of the condensing chamber is connected to the heat-insulating pipe from the desorption chamber 22, and the water outlet at the bottom of the condensing chamber is connected to the water collection tank. An electronic scale 24 is installed below the water collection tank to collect the cumulative water production in real time and feed it back to the intelligent control module H; The outlet of the water collection tank is connected to the inlet of the electrolytic hydrogen production module G via a gravity pipeline.

[0020] The electrolytic hydrogen production module G uses a PEM electrolyzer, its membrane electrode uses a Nafion proton exchange membrane, the working voltage range is 0 to 3.5V, the constant current is 7A, the effective area is 6.5cm², and the rated hydrogen production rate is 50mL / min. The hydrogen outlet of the PEM electrolyzer is connected to a sealed constant-pressure hydrogen collection container, while the oxygen outlet is either vented or collected. The intelligent control module H is used to control the switch tube 20 to turn on to supply power to start the electrolytic hydrogen production module G when the cumulative water production fed back by the electronic scale 24 reaches the minimum start-up water volume threshold of the PEM electrolyzer.

[0021] The intelligent control module H includes a data perception layer and a machine learning optimization decision layer; The data sensing layer includes a temperature and humidity sensor, a light sensor, an electronic scale 24, and a K-type thermocouple 13, which are used to collect ambient temperature, ambient relative humidity, light intensity, cumulative water production, and desorption chamber temperature, respectively, and input them into the machine learning optimization decision layer. The machine learning optimization decision layer uses the Bayesian optimization method to construct a Gaussian process regression probabilistic surrogate model with the optimization objective of maximizing water production per unit of energy consumption. The model takes ambient temperature, relative humidity, light intensity, and cumulative water production as inputs and outputs the adsorption-desorption switching cycle duration and the target heating temperature. In the model initialization phase, virtual samples are generated based on the system's thermal-mass coupling numerical simulation, and the initial model is trained by combining a small amount of measured data. In the online operation phase, measured data is used first to update the model, and simulation data is used as auxiliary samples.

[0022] A method for producing hydrogen from air using water in a photovoltaic-driven continuous air-water extraction hydrogen production integrated system based on intelligent adaptive control includes the following steps: S1: Photovoltaic power generation module A supplies power to the system. Rotary servo motor 11 drives the rotary adsorption bed 3 to rotate, exposing the water-absorbing adsorption material to the atmosphere to adsorb water vapor. S2: After the single-bed adsorption is completed, the rotary servo motor 11 positions the saturated bed at the desorption station, and the thrust linear motor 7 pushes the top cover 2 down. The top cover 2 is limited by the bushing 5, so that the sealing side wall extending from the lower surface of the top cover 2 fits with the rotary adsorption bed 3 to form a closed desorption chamber 22. At the same time, the bottom of the rotary adsorption bed 3 fits tightly with the electric heating plate 12. S3: The thermoelectric temperature control module C adopts the PWM-EKF-MPC control strategy to quickly raise the temperature of the desorption chamber 22 to the set value, so that the adsorption material desorbs and generates water vapor; the air pump D forces the hot and humid air in the desorption chamber 22 to be sent through the heat-insulated pipe into the condensation and water collection module F to condense into liquid water and collect it into the water collection tank. S4: The intelligent control module H uses a Gaussian process regression Bayesian optimization model to adaptively output the optimal adsorption-desorption switching timing and heating target temperature based on real-time ambient temperature, humidity, light intensity, and water production, and dynamically updates the operating parameters. S5: When the cumulative water production in the water collection tank reaches the start-up threshold of the electrolysis hydrogen production module G, the intelligent control module H controls the switch tube 20 to conduct, and the electrolysis hydrogen production module G electrolyzes liquid water to produce hydrogen; at other times, the electrolysis hydrogen production module G is powered off and goes into hibernation.

[0023] The beneficial effect of the technical solution provided in the embodiments of the present invention is that: by setting an air pump inside the desorption chamber, the water vapor flow rate is increased from 4.91 × 10⁻⁶ under natural diffusion conditions. -6 The efficiency was significantly increased to 0.0314 g / s, an improvement of approximately 6400 times, resulting in a substantial increase in the actual water production of the condensation and collection module. By covering the desorption chamber and the outer walls of the connecting pipes with insulation, premature condensation loss of water vapor before reaching the condensation chamber was effectively suppressed. Replacing the spring above the spline shaft with a fixed-height bushing eliminated the problem of uneven force distribution on the adsorption disk caused by the mismatch of the spring coefficients of the two springs, improving the sealing reliability and structural stability of the desorption chamber. A temperature control scheme combining PWM with EKF online parameter identification and MPC optimization control enabled rapid response and precise temperature control of the desorption chamber, solving the problems of slow heating and precise temperature control under constant power heating. By introducing a Bayesian optimization intelligent decision-making method based on Gaussian process regression, and adaptively adjusting the adsorption-desorption switching timing based on real-time environmental perception, the shortcomings of fixed-time-ratio cycles in adapting to dynamic changes in temperature and humidity were overcome, achieving a systematic improvement in all-weather water production and hydrogen production efficiency. The system's annual water production is approximately 46.3 L, and its annual hydrogen production is approximately 58 Nm³. Attached Figure Description

[0024] Figure 1 This is a diagram showing the relationships between the modules of an integrated system that combines photovoltaic power generation, continuous air-water extraction, and electrolytic hydrogen production.

[0025] Figure 2 This is a schematic diagram of the integrated system device.

[0026] Figure 3 Numerical simulation diagram of water vapor concentration distribution in the desorption chamber, pipeline and condensation chamber without the addition of an air pump.

[0027] Figure 4 Numerical simulation diagram of water vapor concentration distribution in the desorption chamber, pipeline and condensation chamber after adding air pump. Figure 5 The absolute humidity distribution map is obtained from the numerical simulation of the thermo-humidity coupling in the desorption chamber. Figure 6 The relative humidity distribution diagram is obtained from the numerical simulation of the thermo-humidity coupling in the desorption chamber.

[0028] Figure 7 Photographs showing water droplets precipitating in the desorption chamber and pipes before the installation of the insulation layer. Figure 8 This is a picture of the device after the insulation layer has been installed.

[0029] Figure 9 This is a main view of the continuous multi-adsorption bed module and temperature control module, showing the top cover (2) in its initial position.

[0030] Figure 10 This is a main view of the continuous multi-adsorption bed module and temperature control module, showing the state where the top cover (2) is pressed down to the limit position of the bushing (5). Figure 11 This is a main view of the continuous multi-adsorption bed module and temperature control module structure, showing the state in which the rotating adsorption bed (3) is reset under the action of the spring (10) and keeps in contact with the bottom plate (9) after the top cover (2) rises.

[0031] Figure 12 This is a top view of the continuous multi-adsorption bed module and temperature control module.

[0032] Figure 13 A schematic diagram of the top cover (2) shows the sealing sidewalls extending downward along the fan-shaped edge on its lower surface.

[0033] Figure 14 This is a schematic diagram of the electrical system of the photovoltaic power generation module, the continuous multi-adsorption bed module, and the electrolysis hydrogen production module.

[0034] Figure 15 This is a schematic diagram of the hardware connections for the temperature control system.

[0035] Figure 16 The figure shows a comparison of the temperature and weight response curves of constant power heating and PWM power regulation temperature control strategies, where (a) represents constant power heating and (b) represents PWM power regulation temperature control strategy.

[0036] Figure 17 Flowchart of the extended Kalman filter parameter identification algorithm.

[0037] Figure 18 This is a block diagram illustrating the principle of model predictive control strategy.

[0038] Figure 19 This is a schematic diagram of a machine learning adaptive optimization control system architecture, showing the information flow between the data perception layer, the machine learning optimization decision layer, and the control execution layer.

[0039] Figure 20 This is a schematic diagram of the simulation-assisted data augmentation process, including the process of generating virtual samples based on numerical simulation, fusing them with measured samples, and initializing and iteratively updating the machine learning optimization model.

[0040] In the diagram, 1-spring damper, 2-top cover, 3-rotary adsorption bed, 4-hygroscopic adsorption material (PHEA-LiCl hydrogel), 5-shaft sleeve, 6-spline shaft, 7-thrust linear motor, 8-water vapor outlet, 9-base plate, 10-spring, 11-rotary servo motor, 12-electric heating element, 13-K-type thermocouple, 14-photovoltaic panel, 15-maximum power point tracking device, 16-battery, 17-DC-DC converter, 18-thermoelectric temperature control module, 19-microcontroller control system, 20-switching tube, 21-PEM electrolytic cell, 22-analytical auxiliary chamber, 24-electronic scale. Detailed Implementation

[0041] The technical features of the embodiments of the present invention will be clearly and completely described below in conjunction with the technical solutions of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0042] See Figure 1 This invention provides a photovoltaic-driven continuous air-water extraction and hydrogen production integrated system based on intelligent adaptive control, which solves the problems of poor sealing reliability, low water vapor mass transfer efficiency, lag in temperature control, and rigid adsorption-desorption cycle in the prior art.

[0043] 1. Overall System Structure and Connections The system mainly consists of a photovoltaic power generation module A, a continuous multi-adsorption bed module B, a thermoelectric temperature control module C, a gas pump D, an insulation layer E, a condensate collection module F, an electrolytic hydrogen production module G, and an intelligent control module H.

[0044] Photovoltaic power generation module A serves as the system's energy source, with its output electrically connected to the input of a maximum power point tracking device (MPPT). The MPPT's output is divided into three parallel paths: the first connects to a battery to store excess energy and provide power compensation during periods of insufficient sunlight; the second connects to the power supply of thermoelectric temperature control module C via a DC-DC converter; and the third connects to the power supply of electrolytic hydrogen production module G via a controlled switch. Furthermore, this output also provides operating voltage to the drive motor, gas pump D, and intelligent control module H of the continuous multi-adsorption bed module B. See the electrical system documentation for details. Figure 14 , The continuous multi-adsorption bed module B is used to perform the adsorption and desorption cycle of atmospheric moisture. The thermoelectric temperature control module C is used to precisely heat the continuous multi-adsorption bed module B during the desorption stage. The air pump D is used to force the high-temperature water vapor generated during desorption to the condensation and collection module F. An insulation layer E covers the outer wall of the desorption chamber and connecting pipes to prevent premature condensation of water vapor. The condensation and collection module F condenses the received water vapor into liquid water and collects it; its outlet is connected to the water inlet of the electrolysis hydrogen production module G via a gravity pipeline. The intelligent control module H is connected to the drive and signal terminals of each of the above modules to realize closed-loop control and adaptive decision-making of the system.

[0045] 2. Specific structure of continuous multi-adsorption bed module The continuous multi-adsorption bed module B adopts a rotary-pressing multi-station structure, specifically including a base plate, rotary adsorption bed, top cover, spline shaft, thrust linear motor, bushing and spring.

[0046] The base plate is horizontally fixed to the frame, and a rotary servo motor is fixedly installed below it. A splined shaft vertically passes through the center hole of the base plate and the rotary adsorption bed, and its lower end is coaxially fixedly connected to the output shaft of the rotary servo motor via a coupling. The rotary adsorption bed consists of three or more sector-shaped bed bodies evenly distributed circumferentially around the splined shaft. Each sector-shaped bed body is separated from each other and does not contact each other. A high thermal conductivity copper sheet is fixed on the upper surface of each sector-shaped bed body, and the surface of the copper sheet is loaded with PHEA-LiCl composite hydrogel as a water-absorbing adsorbent material.

[0047] The top cover is located directly above the rotary adsorption bed and is slidably fitted onto the splined shaft along its axial direction. A linear thrust motor is fixed above the top cover, with its push rod end rigidly connected to the upper surface of the top cover, used to drive the top cover downwards. A metal bushing of fixed height is fitted on the splined shaft between the top cover and the rotary adsorption bed. The height tolerance of this bushing is precisely machined to equal the required sealing distance of the desorption chamber, used to limit the downward limit position of the top cover. A return spring is fitted on the splined shaft between the rotary adsorption bed and the base plate.

[0048] During the adsorption stage, a rotary servo motor drives the rotary adsorption bed to rotate intermittently, exposing each sector-shaped bed to ambient air in sequence. Once the target sector-shaped bed is saturated with adsorption, the rotary servo motor positions it directly below the top cover, and a thrust linear motor pushes the top cover downwards until it abuts against the upper end of the bushing. At this point, the sealing sidewall extending from the lower surface of the top cover fits against the corresponding individual sector-shaped bed, forming a sealed desorption chamber. Simultaneously, the bushing applies uniform pressure to the sector-shaped bed through the top cover, ensuring its bottom is tightly fitted against the electric heating element below. (See also...) Figures 9 to 11 . 3. Gas transmission and condensate collection structure The air pump is fixedly installed inside the desorption chamber, and its outlet is connected to the air inlet of the condensate collection module F via a water vapor outlet on the top cover and a subsequent insulated pipe. A water vapor mass transfer diffusion model was established using COMSOL multiphysics simulation software, and simulation verification was performed. The parameter settings are shown in Table 1. Simulation results show that under the conditions of an air pump flow rate of 24 L / min and an adsorbent material surface temperature of 100℃, the water vapor mass transfer flow rate can be increased from 4.91 × 10⁻⁶ under natural diffusion conditions. -6 The water vapor flow rate increased to 0.0314 g / s, an increase of approximately 6400 times. The maximum water vapor flow rate here refers to the water vapor mass flow rate when the hydrogel absorbs water to saturation and the temperature has reached the target temperature. For its concentration distribution, please refer to [reference needed]. Figure 3 , Figure 4 .

[0049] Figure 3 100℃ 25％ 0 20℃ Figure 4 100℃ 25％ 24L / min 20℃ In an exemplary embodiment of the present invention, a thermo-humidity coupling simulation model is also established for the desorption chamber to verify the rationality of the gas pump setting. Based on Figure 4 The parameters were calculated, and the results show that the absolute humidity at the desorption chamber outlet is slightly lower than the surface humidity of the hydrogel. The encounter of hot and cold air does not cause premature liquefaction of water vapor within the desorption chamber, verifying the rationality of the air pump flow rate setting. See also... Figure 5 , Figure 6 .

[0050] The insulation layer, made of aerogel felt, is continuously wrapped around the outer walls of the desorption chamber and the connecting pipes, effectively suppressing heat loss and premature liquefaction of water vapor within the pipe walls. Experimental results show that the actual water production of the condensate collection module is significantly increased after adding the insulation layer. See also Figure 7 , Figure 8 .

[0051] The condensation and water collection module F includes a condensation chamber and a water collection tank. The inner wall of the condensation chamber has several parallel triangular pyramidal fins, with hydrophilic material filling the spaces between adjacent fins to enhance gas-liquid heat exchange efficiency. A hydrophobic hole is located at the bottom of the condensation chamber, allowing the liquid water produced during condensation to flow into the water collection tank below under gravity. A high-precision electronic scale is installed at the bottom of the water collection tank to collect the cumulative water production in real time and convert the weight signal into an electrical signal, which is then fed back to the intelligent control module. Key gas path parameters are shown in the table below, verified through multiphysics simulation and field measurements.

[0052] 4. Hardware and control algorithm of thermoelectric temperature control module The thermoelectric temperature control module C consists of a temperature sensing unit, a power regulation unit, and a heating execution unit, forming a closed-loop control circuit. The heating execution unit is a high thermal conductivity silicone rubber electric heating element, horizontally mounted on the base plate, facing the desorption station. The temperature sensing unit uses a thin-film K-type thermocouple, with its sensing surface in close contact with the bottom surface of a high thermal conductivity copper sheet to ensure real-time and accurate temperature response. The power regulation unit consists of a PWM control unit and a HIP4081 power amplifier module. See the temperature control hardware connection diagram. Figure 15 .

[0053] For heated objects involving hydrogel phase transition processes, this invention models them as a transfer function model of a first-order inertial element plus a pure time-delay element: ,in, s For the Laplace operator, K This is the gain coefficient. T It is a time constant. τ This represents the pure time lag. Because the above parameters dynamically change with the water content of the adsorbent material and the ambient temperature, the system exhibits strong nonlinear characteristics.

[0054] Because the above parameters change dynamically with the water content of the adsorbent material and the ambient temperature, the system exhibits strong nonlinear characteristics.

[0055] To achieve rapid temperature tracking and interference resistance, this invention employs a composite control strategy of PWM-EKF-MPC (Pulse Width Modulation-Extended Kalman Filter-Model Predictive Control). See the flowchart of the EKF parameter identification algorithm. Figure 17 See the block diagram of the MPC control strategy. Figure 18 The specific control parameter configurations are shown in the table below:

[0056] 5. Intelligent control and adaptive operation decision-making The intelligent control module H is integrated into the microcontroller control system and includes a data sensing layer and a machine learning optimization decision layer. The data sensing layer consists of temperature and humidity sensors, light sensors, electronic scales, and thermocouples distributed throughout the system, used to collect ambient temperature, relative humidity, light intensity, cumulative water production, and desorption chamber temperature in real time.

[0057] The machine learning optimization decision layer employs a Bayesian optimization framework to construct a Gaussian process regression probabilistic surrogate model. The model uses ambient temperature, relative humidity, light intensity, and the current water content of the adsorption bed as input features, and the adsorption-desorption switching cycle duration and the target heating temperature for the next desorption cycle as output control variables. The optimization objective is to maximize water production per unit of energy consumption. (See...) Figure 19The intelligent control module (H) sends out the two output control quantities for execution respectively: the heating target temperature is used as the temperature setpoint of the model predictive control optimization problem in the thermoelectric temperature control module (C). The thermoelectric temperature control module (C) uses the PWM-EKF-MPC composite control strategy to precisely adjust the temperature of the desorption chamber (22) to the target temperature, thereby reflecting the adaptive adjustment of the heating target temperature; the adsorption-desorption switching cycle duration is used to determine the adsorption-desorption switching timing. The intelligent control module (H) controls the rotary servo motor (11) to drive the rotary adsorption bed (3) to rotate at the corresponding time, so that each sector bed switches between the adsorption station and the desorption station, thereby reflecting the adaptive adjustment of the adsorption-desorption switching timing.

[0058] To address the scarcity of measured data during the initial operation of the device, a two-stage strategy of "simulation-assisted data augmentation combined with online iteration using real-world data" was adopted for model training. In the initialization phase, a large amount of virtual sample data was generated based on the system's thermo-mass coupling numerical simulation model, and this data, along with a small amount of measured data, was used for the initial fitting of the Gaussian process regression model. In the online iteration phase, real data collected during the actual operation of the device was prioritized for updating and correcting the model parameters, while simulation data was used as auxiliary samples in the training process, thereby improving the model's generalization and adaptability to different regional climatic characteristics.

[0059] 6. Complete operation process After the system is powered on, photovoltaic power generation module A supplies power to the entire unit. A rotary servo motor drives the rotary adsorption bed to rotate, and the PHEA-LiCl hydrogel adsorbs moisture from the atmosphere. When the intelligent control module H determines that the current bed has reached adsorption saturation based on the adsorption kinetic model, the rotary servo motor moves the saturated bed into the desorption station. A thrust linear motor pushes the top cover down, and after being limited by the bushing, a sealed desorption chamber is formed.

[0060] Subsequently, the thermoelectric temperature control module C starts, rapidly raising the temperature of the desorption chamber to the set value (typically 100°C) using a PWM-EKF-MPC strategy. The hydrogel releases high-temperature water vapor upon heating. Simultaneously, the air pump starts, forcibly blowing the hot, humid air from the desorption chamber into the condensation and water collection module F. The water vapor condenses into liquid water on the fin surface and collects in the water collection tank. To verify the water production performance of the device under actual operating conditions, a practical operation verification experiment was conducted in the laboratory. In the experiment, the target heating temperature of the desorption chamber was set to 100°C, and the device ran continuously for 20 minutes, evaporating most of the water from the hydrogel. A high-precision electronic balance was used to record the mass change of the condensation and water collection module in real time. The experimental results are as follows: Figure 16As shown, due to the thermal stress in the pipe, the weight data on the weighing platform experienced brief fluctuations. After these fluctuations, the weight data began to steadily increase, increasing by approximately 10g within 20 minutes. This measured value differs from the simulated upper limit of 0.0314 g / s for water vapor flow rate. This difference is because blowing air into the analytical chamber lowers the temperature and humidity, and the weighing data still contains errors. These two differences are not physically contradictory. Figure 16 (a) Compared with constant power heating data, under the condition of similar water yield, the water collection time is greatly reduced after adopting the PWM-EKF-MPC temperature control scheme. The experimental results show that the improvements proposed in this invention, such as forced convection by the air pump, insulation layer to suppress condensation, and PWM-EKF-MPC temperature control, can stably achieve efficient collection of water vapor and condensation water production in the actual device.

[0061] The intelligent control module H monitors the water production data fed back by the electronic scale in real time. When the cumulative water production reaches the minimum start-up threshold of the PEM electrolyzer (approximately 50 mL), the microcontroller control system closes the switch tube, and the electrolytic hydrogen production module G starts. The electrolytic hydrogen production module G uses a PEM electrolyzer, and the membrane electrode uses a Nafion proton exchange membrane. The electrolysis reaction is carried out under a constant current of 7A. The generated hydrogen is dried and pressure-stabilized before being collected in a sealed constant-pressure container, while the oxygen is either vented or collected separately.

[0062] To verify the actual performance of this system, continuous operation tests were conducted under standard laboratory conditions (temperature 25℃, relative humidity 50%). The simulation-aided data augmentation flowchart can be found here. Figure 20 The test results are shown in the table below: Duration of a single desorption cycle min 20 60 Shortened by 66.7% Single water production g 10.0 1.95 Increased by 412% Temperature control response time min <5 >10 Shortened by more than 50% Annual water production L 46.3 ~15 Increased by 208% Annual hydrogen production Nm³ 58 ~20 Increase by 190% Experimental data show that, compared with existing technologies, this invention effectively solves the bottleneck of water vapor mass transfer and the problem of premature condensation through forced convection by an air pump and the design of an insulation layer; it significantly improves the sealing reliability of the desorption chamber by replacing the double spring structure with a bushing; it shortens the heating response time by more than 60% and avoids the risk of overshoot through the PWM-EKF-MPC temperature control strategy; and it maximizes the water production per unit of energy consumption through Bayesian optimization-based intelligent decision-making, enabling the system to maintain maximum water production per unit of energy consumption under different temperature and humidity environments. This system is particularly suitable for long-term self-sustaining operation in arid areas without power grids.

[0063] The above description is only of the preferred embodiment of the present invention and should not be construed as limiting the scope of the claims. The present invention is not limited to the above embodiments, and variations in its specific structure are permitted. All variations made within the scope of the independent claims of the present invention are also within the scope of protection of the present invention.

[0064] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used in this invention includes any and all combinations of one or more of the associated listed items.

Claims

1. A photovoltaic-driven continuous air-water extraction and hydrogen production integrated system based on intelligent adaptive control, comprising: Photovoltaic power generation module (A) is used to power the system; A continuous multi-adsorption bed module (B) is configured to alternately perform atmospheric moisture adsorption and desorption, including a desorption chamber (22). Thermoelectric temperature control module (C) is used to heat the continuous multi-adsorption bed module (B) in the desorption state to generate water vapor; An air pump (D) is installed inside the desorption chamber (22) to force the water vapor generated by desorption to the condensation and water collection module (F). The condensate collection module (F) is used to condense the received water vapor into liquid water and collect it. Electrolysis hydrogen production module (G) is used to produce hydrogen by electrolysis of collected liquid water; The intelligent control module (H) is configured to adaptively adjust the adsorption-desorption switching timing and the target heating temperature based on environmental parameters and system operating status.

2. The integrated system according to claim 1, characterized in that: The photovoltaic power generation module (A) includes a photovoltaic panel (14), a maximum power point tracking device (15), a battery (16), and a DC-DC converter (17). The output terminal of the photovoltaic panel (14) is electrically connected to the input terminal of the maximum power point tracking device (15), and the output terminal of the maximum power point tracking device (15) is divided into three paths: The first circuit is bidirectionally connected to the battery (16) to achieve energy storage and deficit compensation; The second path is electrically connected to the power supply terminal of the thermoelectric temperature control module (C) via the DC-DC converter (17); The third path is electrically connected to the power supply terminal of the electrolysis hydrogen production module (G) via the controlled switch tube (20); The maximum power point tracking device (15) is also electrically connected to the power supply terminals of the drive motor, air pump (D), and intelligent control module (H) of the continuous multi-adsorption bed module (B).

3. The integrated system according to claim 1, characterized in that: The continuous multi-adsorption bed module (B) includes a base plate (9), a rotary adsorption bed (3), a top cover (2), a spline shaft (6), a thrust linear motor (7), a rotary servo motor (11), a bushing (5), and a spring (10). The base plate (9) is a horizontally set plate structure. The rotary servo motor (11) is fixed below the base plate (9) and its output shaft is coaxially fixedly connected to the lower end of the spline shaft (6) that runs vertically through the base plate (9). The rotary adsorption bed (3) consists of multiple fan-shaped beds evenly distributed around the spline shaft (6) as the center. Each fan-shaped bed is separated from the others and is fitted onto the spline shaft (6) to form a spline fit with the spline shaft (6). The rotary servo motor (11) drives the rotary adsorption bed (3) to rotate as a whole through the spline shaft (6). The top cover (2) is located above the rotary adsorption bed (3) and is slidably sleeved on the spline shaft (6). The lower surface of the top cover (2) is vertically provided with a sealing sidewall surrounding the circumferential edge of the fan-shaped bed area. The thrust linear motor (7) is fixed above the top cover (2) and its push rod is connected to the upper surface of the top cover (2). A fixed-height bushing (5) is fitted on the spline shaft (6) between the top cover (2) and the rotary adsorption bed (3). The height of the bushing (5) is equal to the required sealing distance of the desorption chamber (22) and is used to limit the downward position when the top cover (2) is pressed down. A spring (10) is sleeved on the spline shaft (6) between the rotating adsorption bed (3) and the bottom plate (9) to reset the rotating adsorption bed (3) and keep it in contact with the bottom plate (9) after the top cover (2) rises; When the top cover (2) is pressed down to the limit position of the bushing (5), the sealing sidewall extending downward from the lower surface of the top cover (2) fits against the upper surface and circumferential edge of each fan-shaped rotating adsorption bed (3) to form a sealed desorption chamber (22); the heat insulation layer (E) covers the desorption chamber (22) and the outer wall of the connecting pipe.

4. The integrated system according to claim 3, characterized in that: Each sector of the rotating adsorption bed (3) is fixed with a high thermal conductivity copper sheet (4), and the high thermal conductivity copper sheet (4) is loaded with a water-absorbing adsorption material. The water-absorbing adsorbent material is PHEA-LiCl hydrogel; The desorption chamber (22) is equipped with a gas pump (D). The outlet of the gas pump (D) is connected to the air inlet of the condensate collection module (F) through a water vapor outlet (8) opened on the top cover (2) and an insulation pipe.

5. The integrated system according to claim 1, characterized in that: The thermoelectric temperature control module (C) includes an electric heating element (12), a K-type thermocouple (13), a PWM control unit, and a HIP4081 power amplifier module; The electric heating element (12) is horizontally positioned below the rotating adsorption bed (3) and fixed on the base plate (9) for heating the water-absorbing adsorption material above it; A K-type thermocouple (13) is attached to the lower surface of a high thermal conductivity copper sheet (4) at the bottom of the rotating adsorption bed (3) in the desorption state, for real-time acquisition of the temperature of the adsorption material. The signal output terminal of the K-type thermocouple (13) is electrically connected to the temperature signal input terminal of the intelligent control module (H). The PWM control signal output terminal of the intelligent control module (H) is connected to the power supply terminal of the electric heating element (12) in sequence through the PWM control unit and the HIP4081 power amplifier module, forming a temperature closed-loop control circuit. The thermoelectric temperature control module (C) includes an electric heating element (12) horizontally positioned below the rotating adsorption bed (3) for heating the water-absorbing adsorption material (4) on the surface of the high thermal conductivity copper sheet; a K-type thermocouple is provided between the electric heating element (12) and the rotating adsorption bed (3), and its temperature measuring end is in direct contact with the bottom of the rotating adsorption bed (3) in the desorption state for real-time acquisition of the bottom temperature of the adsorption material (4).

6. The integrated system according to claim 5, characterized in that: The thermoelectric temperature control module (C) models the controlled object, consisting of the electric heating element (12), the K-type thermocouple (13), the high thermal conductivity copper sheet (4), the water-absorbing adsorption material, and the desorption chamber (22), as a transfer function of a first-order inertial plus pure time-delay element: Where K is the gain coefficient, T is the time constant, and τ is the pure time delay; s For the Laplace operator; The intelligent control module (H) uses an extended Kalman filter algorithm to identify time-varying parameters K, T, and τ online. The real-time heating power uk of the electric heating element (12) is used as the input, and the temperature yk collected by the K-type thermocouple is used as the observation. The time-varying parameters K, T, and τ are used as state variables θk, and the update formula is as follows: ； Recursive update, where, Let h(·) be the Kalman gain, and h(·) be the observation function. The adaptive weighting coefficients are bound to the real-time excitation intensity of the system. They are adjusted online according to the intensity of the innovation sequence: when the system approaches steady state and the excitation weakens, they approach 0 to suppress noise and maintain stable parameter estimation; when the operating conditions change abruptly and the excitation strengthens, they approach 1 to accelerate the tracking of time-varying parameters K, T, and τ. Based on the identification of time-varying parameters, a model predictive control algorithm is used to solve the optimization problem in each control cycle: in, To predict the time domain, Set the temperature value. This is the maximum temperature that the desorption chamber can withstand. The rated power of the electric heating element is Q and R are weighting coefficients. Solving the optimization problem yields the optimal heating power sequence that minimizes the objective function in the prediction time domain. According to the rolling time domain control principle, the thermoelectric temperature control module (C) only takes the optimal power value of the current control cycle, converts it into a PWM signal with the corresponding duty cycle, outputs it to the HIP4081 power amplifier module, and re-solves the problem based on the updated identification parameters in the next control cycle. The temperature setpoint in the optimization problem is the heating target temperature adaptively output by the intelligent control module (H). The adsorption-desorption switching timing is executed by the switching cycle duration output by the intelligent control module (H) via the rotary servo motor (11).

7. The integrated system according to claim 1, characterized in that: The condensate collection module (F) includes a condensate chamber and a water collection tank; The inner wall of the condensing chamber is provided with several parallel triangular cone-shaped fins, and the space between adjacent fins is filled with hydrophilic material. The air inlet of the condensing chamber is connected to the heat-insulating pipe from the desorption chamber (22), and the water outlet at the bottom of the condensing chamber is connected to the water collection tank. An electronic scale (24) is installed below the water collection tank to collect the cumulative water production in real time and feed it back to the intelligent control module (H); The outlet of the water collection tank is connected to the inlet of the electrolytic hydrogen production module (G) via a gravity pipeline.

8. The integrated system according to claim 1, characterized in that: The electrolytic hydrogen production module (G) adopts a PEM electrolyzer, its membrane electrode adopts Nafion proton exchange membrane, the working voltage range is 0 to 3.5V, the constant current is 7A, the effective area is 6.5cm², and the rated hydrogen production rate is 50mL / min. The hydrogen outlet of the PEM electrolyzer is connected to a sealed constant-pressure hydrogen collection container, while the oxygen outlet is either vented or collected. The intelligent control module (H) is used to control the switch tube (20) to turn on to power the electrolytic hydrogen production module (G) when the cumulative water production fed back by the electronic scale (24) reaches the minimum start-up water volume threshold of the PEM electrolyzer.

9. The integrated system according to claim 1, characterized in that: The intelligent control module (H) includes a data perception layer and a machine learning optimization decision layer; The data sensing layer includes a temperature and humidity sensor, a light sensor, an electronic scale (24) and a K-type thermocouple (13), which are used to collect ambient temperature, ambient relative humidity, light intensity, cumulative water production, and desorption chamber temperature, respectively, and input them into the machine learning optimization decision layer. The machine learning optimization decision layer uses the Bayesian optimization method to construct a Gaussian process regression probabilistic surrogate model with the optimization objective of maximizing water production per unit of energy consumption. The model takes ambient temperature, relative humidity, light intensity, and cumulative water production as inputs and outputs the adsorption-desorption switching cycle duration and the target heating temperature. In the model initialization phase, virtual samples are generated based on the system's thermal-mass coupling numerical simulation, and the initial model is trained by combining a small amount of measured data. In the online operation phase, measured data is used first to update the model, and simulation data is used as auxiliary samples.

10. A method for producing hydrogen from water by air extraction based on the integrated system according to any one of claims 1-9, characterized in that: Includes the following steps: S1: The photovoltaic power generation module (A) supplies power to the system, and the rotary servo motor (11) drives the rotary adsorption bed (3) to rotate, so that the water-absorbing adsorption material is exposed to the atmosphere to adsorb water vapor; S2: After the single-bed adsorption is completed, the rotary servo motor (11) positions the saturated bed at the desorption station, and the thrust linear motor (7) pushes the top cover (2) down. The top cover (2) is limited by the bushing (5), so that the sealing side wall extending from the lower surface of the top cover (2) fits with the rotary adsorption bed (3) to form a closed desorption chamber (22). At the same time, the bottom of the rotary adsorption bed (3) fits tightly with the electric heating plate (12). S3: The thermoelectric temperature control module (C) adopts the PWM-EKF-MPC control strategy to quickly raise the temperature of the desorption chamber (22) to the set value, so that the adsorption material desorbs and generates water vapor; the air pump (D) forces the hot and humid air in the desorption chamber (22) to be sent into the condensation and water collection module (F) through the heat-insulated pipe to condense into liquid water and collect it into the water collection tank. S4: The intelligent control module (H) uses a Gaussian process regression Bayesian optimization model to adaptively output the optimal adsorption-desorption switching timing and heating target temperature based on real-time ambient temperature, humidity, light intensity, and water production, and dynamically updates the operating parameters. S5: When the cumulative water production of the water collection tank reaches the start-up threshold of the electrolysis hydrogen production module (G), the intelligent control module (H) controls the switch tube (20) to turn on, and the electrolysis hydrogen production module (G) electrolyzes the liquid water to produce hydrogen; during other periods, the electrolysis hydrogen production module (G) is powered off and goes into hibernation.

Images