A photocatalysis-high temperature thermochemical coupling fuel production system and a method for operating and regulating the same

Abstract

The application discloses a photocatalysis-high-temperature thermochemical coupling fuel production system and a regulation and control method. The system comprises a solar spectrum frequency divider, a parabolic concentrator, a circular tube reactor, a high-temperature solar energy driven metal oxide thermochemical cycle fuel production device, a waste heat recovery device, a hydrogenation station, a Fischer-Tropsch synthesis device, a product separation device, a gas turbine, a power grid system and a residential and industrial park. The system couples and complements photocatalysis-high-temperature thermochemical fuel production technology, and widens the solar spectrum utilization range. Meanwhile, for the photocatalytic hydrolysis hydrogen production device, according to the solar radiation energy and the user load, the parabolic concentrator structure and the reactor operation parameters are regulated and controlled under full light conditions; for the high-temperature thermochemical fuel production device, the reactor operation parameters are regulated and controlled, so that the radiation energy input into the reactor matches the energy required by the chemical reaction, and the solar energy-fuel energy conversion efficiency of the system is improved.

Description

Technical Field

[0001] This invention belongs to the fields of photocatalytic hydrogen production, solar thermochemical hydrogen production, and carbon monoxide production, specifically relating to a photocatalytic-high-temperature thermochemical coupled fuel production system and its operation control method. Background Technology

[0002] Utilizing solar energy to decompose water and carbon dioxide to produce hydrogen and carbon monoxide, and then synthesizing fuels, can effectively alleviate energy shortages. In suspended photocatalytic water splitting for hydrogen production, the photocatalyst's response range to solar energy is primarily ultraviolet and visible light. Solar thermochemical-driven metal oxide decomposition of water and carbon dioxide to produce hydrogen and carbon monoxide mainly utilizes the near-infrared spectrum of solar energy. Currently, the main bottleneck hindering the industrial application of photocatalysis and thermochemical fuel production is their low solar-fuel conversion efficiency. The main reasons are: first, photocatalytic water splitting for hydrogen production and high-temperature solar thermochemical fuel production technologies are selective in their use of the solar spectrum, failing to utilize the entire spectral range and reducing energy utilization; second, the energy provided by the solar concentrator system does not match the energy required for the chemical reactions in the reactor, reducing energy conversion efficiency; and third, the intermittent and uncertain nature of solar energy is detrimental to stable system operation.

[0003] Current research on solar-powered hydrogen production technology mainly focuses on the development of photocatalysts and the design of high-temperature thermochemical reactors, neglecting the matching of the concentrating system and the reaction system, as well as the dynamic response characteristics of the system under full-sunlight conditions to meet user loads. Furthermore, research on the impact of solar uncertainties on system stability is also limited. Photocatalytic water splitting for hydrogen production and high-temperature solar thermochemical-driven metal oxide fuel production technologies have different solar spectral response ranges. Coupled with photocatalytic water splitting for hydrogen production and thermochemical fuel production technologies, the solar spectral response range can be improved. Parabolic concentrators can absorb scattered light without requiring solar tracking; therefore, a circular tubular reactor with a parabolic concentrator is selected for photocatalytic water splitting for hydrogen production. For high-temperature solar thermochemical fuel production technology, an oxidation reactor and a reduction reactor are connected in series to ensure continuous fuel production under full-sunlight conditions. When the system is running, based on the measured fuel production and calculated energy conversion efficiency, the parabolic concentrator structure and reactor operating parameters in the photocatalytic water electrolysis hydrogen production device, as well as the operating parameters of the oxidation reactor and reduction reactor in the high-temperature solar-driven metal oxide thermochemical cycle fuel production device, are adjusted to match the surface radiation distribution of the reactor with the energy required for the reaction, thereby ultimately improving the solar-fuel conversion efficiency. Summary of the Invention

[0004] To overcome the bottleneck of low energy conversion efficiency in existing technologies, this invention provides a photocatalytic-high-temperature thermochemical coupled fuel production system and its operation control method. This system employs a coupled photocatalytic water splitting hydrogen production and a solar-driven metal oxide thermochemical cycle fuel production technology. Based on solar radiation and user load, for the photocatalytic water splitting hydrogen production device, the structure of the parabolic concentrator and the reactor operating parameters are adjusted to improve the catalyst's utilization of solar radiation energy, promote charge transfer and energy conversion processes, and enhance energy transfer and energy matching characteristics. For the high-temperature solar thermochemical fuel production device, reactor operating parameters are adjusted. These operations improve the solar-fuel conversion efficiency of the energy system under full sunlight conditions.

[0005] Therefore, the technical solution adopted by the present invention is as follows:

[0006] A photocatalytic-thermochemical coupled fuel production system includes a solar spectral divider, a photocatalytic water splitting hydrogen production device, a high-temperature solar-driven metal oxide thermochemical cycle fuel production device, a Fischer-Tropsch synthesis device, a product separation device, a gas turbine, a hydrogen refueling station, a power grid system, and a residential and industrial park.

[0007] The solar spectrum frequency divider is used after solar irradiation and before the photocatalytic water electrolysis hydrogen production device and the high-temperature thermochemical fuel production device. The frequency-divided solar radiation energy is then injected into the parabolic concentrator and the heliostat, respectively.

[0008] The photocatalytic water splitting hydrogen production device includes: a parabolic concentrator, a circular tube reactor, a first control system, a vacuum pump, and a control atomizer; the circular tube reactor is located near the focal point of the parabolic concentrator; the control system adjusts the structure of the parabolic concentrator and the operating parameters of the photoreactor based on feedback from the system, including the opening direction of the parabolic concentrator, the distance from the tip to the bottom of the reactor, the inlet flow rate of the suspension, the concentration of the photocatalyst, and the diameter of the catalyst particles; the vacuum pump is located before the inlet of the circular tube reactor; and the control atomizer is located between the circular tube reactor and the vacuum pump.

[0009] The aforementioned photocatalytic water splitting hydrogen production device employs a cylindrical photocatalytic reactor with a parabolic concentrator—a cylindrical reactor for hydrogen production. Ultraviolet and visible light are input into the parabolic concentrator within the device via a solar spectral divider. Based on the solar incident angle and radiation intensity, the optimal original receiving half-angle and truncated edge ray angle of the parabolic concentrator are used. The distance from the concentrator tip to the bottom of the reactor, the opening direction, the inlet flow rate of the suspension, the photocatalyst concentration, and the catalyst particle diameter are adjusted in real time to match the solar energy distribution on the reactor surface with the energy required for the chemical reaction. For the high-temperature thermochemical reaction device, the mass flow rates of metal oxides, carbon dioxide, and water vapor in the cylindrical reactor are adjusted to match the incident radiation with the energy required for the chemical reaction, reducing energy loss. Combining photocatalytic hydrogen production with high-temperature thermochemical fuel production technology broadens the solar spectral response range, improves the system's solar-to-fuel energy conversion efficiency, and mitigates the impact of fluctuations in renewable energy and user energy loads on the system's stable operation.

[0010] The solar-driven thermochemical cycle fuel production device utilizes solar energy to drive a CeO2 thermochemical cycle to produce hydrogen and carbon monoxide. Both the reduction and oxidation reactors in the high-temperature solar thermochemical fuel production process are chamber reactors, effectively capturing the input solar radiation. A parabolic concentrator is installed at the aperture of both the reduction and oxidation reactors to increase reflectivity. A transparent quartz window is installed outside the parabolic concentrator, sealing the reaction chamber and helping to maintain the temperature inside the chamber while preventing reactants and products from overflowing.

[0011] The heliostat reflects and concentrates solar radiation onto the quartz window surfaces at the inlets of the reduction and oxidation reactors. The reduction reactor is maintained at 1500℃, and the oxidation reactor at 900℃. A two-step fuel production process using a high-temperature solar-driven thermochemical cycle of CeO2 metal oxides to produce hydrogen and carbon monoxide is employed. The reduction and oxidation reactors are located at the top of the solar tower. The inlet of the reduction reactor is connected to the outlet of the oxidation reactor via a pipeline; the outlet of the reduction reactor is connected to the inlet of the oxidation reactor via a pipeline, enabling continuous fuel production. The reduction and oxidation reactions occurring in the reduction and oxidation reactors are as follows:

[0012] Reduction reaction:

[0013] Oxidation reaction: CeO 2-δ +δCO2→CeO2+δCO

[0014] CeO 2-δ +δH2O→CeO2+δH2

[0015] Wherein, the non-stoichiometric coefficient δ represents the degree of reduction, and its specific expression is as follows:

[0016]

[0017] Where T represents the reduction reaction temperature; P0 represents the oxygen bias and the pressure of the reduction and oxidation reactions.

[0018] The waste heat recovery device is connected to the reduction reactor and the oxidation reactor. After absorbing the heat from the oxidation reactor to heat CeO2, it is introduced into the reduction reactor to improve the energy utilization efficiency of the reactor.

[0019] The hydrogen produced by the photocatalytic water electrolysis hydrogen production device, along with the hydrogen produced by the high-temperature solar-driven metal oxide thermochemical cycle, is transported to hydrogen refueling stations via gas pipelines and long-tube trailers for use by fuel cell vehicles.

[0020] The Fischer-Tropsch synthesis apparatus synthesizes methane from a mixture of carbon monoxide and hydrogen under the conditions of commercial low-temperature methanation catalyst BC-H-10 and temperature of 150-200°C, and is connected to a product separation device via a pipeline.

[0021] The product separation device includes a separation tank, a desorption tower, a demethanizer, a deethaner, a depropanizer, a debutanizer, a distillation tower, etc. The inlet of the primary separation tank is connected to the Fischer-Tropsch synthesis unit via a pipeline, and the outlets of the demethanizer and stripping tower are connected to the gas turbine via pipelines.

[0022] This invention also provides a method for operating and controlling a photocatalytic-high-temperature thermochemical coupled fuel production system, comprising the following steps:

[0023] Step 1: Establish a photocatalytic-high-temperature thermochemical coupled fuel production system model, including establishing a three-dimensional multiphysics model of fluid dynamics-radiation transfer-chemical reaction dynamics coupling for the circular tube reactor, oxidation reactor, and reduction reactor, and establishing input-output nonlinear models for the waste heat recovery device, Fischer-Tropsch synthesis device, product separation device, and gas turbine based on experimental data.

[0024] Step 2: For the circular tube reactor, the hydrogen production was measured by changing the solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the tip to the bottom of the circular tube reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter, and multiple experimental samples were obtained. For the oxidation and reduction reactors, the hydrogen production and carbon monoxide production were measured by changing the solar radiation intensity, solar incidence angle, carbon dioxide and water vapor flow ratio, and CeO2 mass flow rate, and multiple experimental samples were obtained.

[0025] Step 3: Propose the Extreme Learning Machine;

[0026] Step 4: Based on experimental and computational data, the limit learning machine from Step 3 is used to approximate the nonlinear relationships between solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the apex to the bottom of the reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter and hydrogen production in the circular tube reactor. Simultaneously, for the oxidation and reduction reactors, the limit learning machine from Step 3 is used to approximate the nonlinear relationships between solar radiation intensity, solar incidence angle, carbon dioxide to water vapor flow ratio, CeO2 mass flow rate and hydrogen production, and carbon monoxide content in the oxidation and reduction reactors. A surrogate model is obtained, reducing computational and experimental time and costs.

[0027] Step 5: Establish a robust optimization model for the energy system. Use a non-dominated sorting genetic algorithm to solve the proposed robust optimization model. Adjust the opening direction of the parabolic concentrator, the distance from the tip to the bottom of the reactor, the inlet flow rate of the suspension, the concentration of the photocatalyst, and the diameter of the catalyst particles in the photocatalytic water electrolysis hydrogen production device in real time under all-day sunshine conditions. This ensures that the energy provided by the parabolic concentrator matches the solar radiation distribution required by the reactor. For a high-temperature solar-driven metal oxide thermochemical cycle fuel production device, control the atomizer to automatically adjust the flow ratio of carbon dioxide and water vapor and the mass flow rate of CeO2. This maximizes the solar-to-fuel energy conversion efficiency of the energy system and mitigates the impact of uncertainties and intermittencies in solar energy and user loads on the stable operation of the system.

[0028] Furthermore, step three, proposing the extreme learning machine, includes:

[0029] Given a dataset Where s i For the i-th input sample, z i Let N be the number of output samples, and let N be the number of samples. The mathematical model of the Extreme Learning Machine is established as follows:

[0030] In the first model, the input variables are the operating parameters of the circular tube reactor, including solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the tip to the bottom of the reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter. The output parameter is the hydrogen production.

[0031] In the second model, the input variables are the operating parameters of the oxidation reactor and the reduction reactor, including solar radiation intensity, solar incidence angle, carbon dioxide and water vapor flow rate ratio, and CeO2 mass flow rate. The output parameters are hydrogen production and carbon monoxide production.

[0032] Hg=z (1)

[0033] Where H represents the model matrix; g represents the output weight vector; z represents the output value; the specific expression is:

[0034]

[0035]

[0036] Where φ(·) is the activation function; a i and v j Define the i-th input weight and the i-th bias; g and h define the output weight and the input matrix; a1 and a n Represents the first and nth elements of the input weight vector a; v1 and v n Represents the first and nth elements of the bias vector v; g1 and g n Let g represent the first and nth elements of vector g; n represents the number of vectors.

[0037] Equation (1) is solved using regularization.

[0038] The two-layer optimization training model achieves adaptive selection of regularization parameters and training of model parameters, expressed as the following mathematical model:

[0039]

[0040] In the formula, ||Hg-z||1 is the data fidelity term, which reduces the adverse effects of noise and bias; λ is the regularization parameter; ||z||1-||z||2 is the L1-2 norm, used to ensure the sparsity of numerical solutions; H T This is the model matrix corresponding to the validation set; z T This corresponds to the output data of the validation set; This indicates the search for the minimum value of the objective function with * as the variable; It is the objective function of the higher-level optimization problem, which attempts to find the optimal regularization parameter λ; is the objective function of the lower-level optimization problem, which attempts to compute the output weight g under given regularization parameters and training data; st represents the constraint condition; the upper-level optimization problem and the lower-level optimization problem constitute a bi-level optimization problem.

[0041] Furthermore, a nested algorithm is used to solve the two-level optimization problem of equation (2), specifically including:

[0042] Lower-level optimization problem In the case where g is not a zero vector, the lower-level optimization problem is relaxed as follows:

[0043]

[0044] In the formula, v k As an auxiliary variable, it is defined as:

[0045] v k =-λgk / ||g k ||2 (4)

[0046] In the formula, k represents the number of iterations; ||g k ||2 is g k L2 norm;

[0047] Using a semi-quadratic splitting algorithm, by introducing two auxiliary variables d1 and d2, equation (3) is transformed into the following equality-constrained optimization problem:

[0048]

[0049] In the formula, z T Let z be the transpose of vector z.

[0050] According to the semi-quadratic splitting algorithm, equation (6) is further transformed into an unconstrained problem:

[0051]

[0052] In the formula, ψ(g,d1,d2) is the objective function, defined as:

[0053]

[0054] In the formula, g T μ1 and μ2 are the transpose of vector g; μ1 and μ2 are penalty parameters.

[0055] Equation (7) is solved using the separation optimization method:

[0056]

[0057]

[0058]

[0059] According to equation (7), equations (8)-(10) are specifically as follows:

[0060]

[0061]

[0062]

[0063] In the formula, ||d1||1 represents the L1 norm of the auxiliary variable d1.

[0064] Equations (11) and (12) are solved using the soft thresholding algorithm, i.e.:

[0065]

[0066]

[0067] In the formula, shrink(·,·) represents the soft threshold operator.

[0068] Equation (13) is differentiable, and its solution is:

[0069]

[0070] In the formula, H T H is the transpose of vector H; I is the identity matrix.

[0071] Equation (3) is solved iteratively using equations (14)-(16);

[0072] Genetic algorithms are used to solve the upper-level optimization problem.

[0073] Compared with the prior art, the present invention has the following advantages:

[0074] This invention discloses a novel photothermal coupled fuel production system that integrates photocatalytic water splitting for hydrogen production and solar-driven metal oxide thermochemical cycle fuel production technologies, broadening the spectral utilization range. To mitigate the uncertainty and fluctuations of solar radiation and user load, a robust optimization model for the energy system is established. To reduce experimental and computational costs, surrogate models are established for the circular tube reactor, reduction reactor, and oxidation reactor using a proposed improved limit learning machine. By solving the established models, under full illumination, the opening direction of the parabolic concentrator, the distance from the tip to the bottom of the reactor, the inlet flow rate of the suspension in the reactor, the concentration of the photocatalyst, and the diameter of the catalyst particles are adjusted in real time to match the radiation distribution on the surface of the photoreactor with the energy required for the chemical reaction, thereby improving the solar-to-fuel conversion efficiency. For the reduction and oxidation reactors, the carbon dioxide and water vapor flow ratio, as well as the CeO2 flow rate, are adjusted in real time. By combining real-time measurement data and operational data, the operation of the photothermal coupled fuel production system is controlled to achieve the maximum solar-to-fuel energy conversion efficiency, mitigating the impact of solar uncertainty and intermittency on the stable operation of the system and ensuring its safe and efficient operation. This invention uses a waste heat recovery device to recover the heat released from the thermochemical oxidation reactor, thereby improving energy utilization efficiency.

[0075] In summary, the photocatalytic-thermochemical coupled fuel production system and its operation control method disclosed in this invention have good application prospects. Attached Figure Description

[0076] Figure 1 This is a schematic diagram of a photocatalytic-high-temperature thermochemical coupling fuel production system proposed in an embodiment of the present invention;

[0077] In the diagram: A - Solar Spectrum Analyzer, B - Photocatalytic Hydrogen Electrolysis Unit, C - High-Temperature Solar-Driven Metal Oxide Thermochemical Cycle Fuel Production Unit, D - Hydrogen Refueling Station, E - Fischer-Tropsch Synthesis Unit, F - Product Separation Unit, G - Gas Turbine, H - Power Grid System, I - Residential and Industrial Park; 1 - Vacuum Pump, 2 - First Control System, 3 - Control Atomizer, 4 - Parabolic Concentrator, 5 - Circular Tube Reactor, 6 - Heliostat, 7 - Solar Tower, 8 - Oxidation Reactor, 9 - Reduction Reactor, 10 - Pump, 11 - Second Control System, 12 - Waste Heat Recovery Unit. Detailed Implementation

[0078] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0079] like Figure 1 As shown, a photocatalytic-high-temperature thermochemical coupled fuel production system of the present invention includes a solar spectrum divider A. Based on the spectral response characteristics of the photocatalytic water splitting hydrogen production catalyst and the high-temperature solar thermochemical fuel production metal oxide catalyst, the solar spectrum is divided into two bands: a short-wavelength band of solar radiation from 290nm to 1200nm and a long-wavelength band of the remaining band. The short-wavelength light enters the parabolic concentrator 4 of the photocatalytic water splitting hydrogen production device B to provide energy to the photocatalytic circular reactor 5, exciting the photocatalytic water splitting hydrogen production reaction in the tube. The long-wavelength light enters the heliostat 6 in the high-temperature solar-driven metal oxide thermochemical cycle fuel production device C, reflecting the light to the inlets of the oxidation reactor 8 and the reduction reactor 9 of the thermochemical reactor; driving the CeO2 thermochemical cycle to decompose water and carbon dioxide to produce hydrogen and carbon monoxide.

[0080] The photocatalytic water splitting hydrogen production device B uses a photocatalyst and a sacrificial agent to decompose water under sunlight to produce hydrogen and oxygen.

[0081] The high-temperature solar-driven metal oxide thermochemical cycle fuel production device C uses a heliostat 6 to concentrate solar radiation energy at the inlet of the thermochemical reactor at the top of the solar tower 7 to produce methane fed into the gas turbine G and hydrogen fed into the hydrogen refueling station D; and adds a waste heat recovery device 12 to improve energy utilization efficiency.

[0082] Also includes:

[0083] Hydrogen refueling station D is used to supply hydrogen to hydrogen fuel cell vehicles;

[0084] Fischer-Tropsch synthesis unit E is used to synthesize methane from a mixture of carbon monoxide and hydrogen gas.

[0085] Product separation unit F is used to separate methane from the products of the Fischer-Tropsch synthesis reaction.

[0086] Gas turbine G is used to convert the chemical energy of gas into mechanical energy and then generate electricity to provide users with electrical energy.

[0087] The power grid system H is used to provide electrical energy to users;

[0088] Residential and Industrial Park I: Users consume energy to meet their residential and industrial needs.

[0089] The photocatalytic water splitting hydrogen production device B includes a vacuum pump 1, a first control system 2, a control atomizer 3, a parabolic concentrator 4, and a circular tube reactor 5. By measuring the hydrogen flow rate at the outlet of the circular tube reactor 5 and obtaining the surface radiation distribution of the circular tube reactor 5 using the Monte Carlo ray tracing method, and based on the calculated solar-to-hydrogen energy conversion efficiency, the first control system 2 adjusts in real time the opening direction of the parabolic concentrator 4, the distance from the tip of the parabolic concentrator to the bottom of the circular tube reactor 5, the inlet flow rate of the suspension, the concentration of the photocatalyst, and the diameter of the catalyst particles, so that the solar energy distribution on the surface of the circular tube reactor 5 matches the energy required for the chemical reaction process, thereby maximizing the system's solar-to-fuel conversion efficiency.

[0090] The high-temperature solar-driven thermochemical cycle fuel production device C includes a heliostat 6, a solar tower 7, an oxidation reactor 8, a reduction reactor 9, a pump 10, a second control system 11, and a waste heat recovery device 12. The solar tower 7 and heliostat 6 are equipped with a solar tracking transmission system, which adjusts the reflector in real time according to the solar incidence angle, concentrating more than 95% of the solar radiation energy at the inlet of the thermochemical reduction reactor 9 and the oxidation reactor 8, thereby improving the system's solar-fuel conversion efficiency.

[0091] The temperature of the reduction reactor 9 is maintained at 1500℃ and the O2 pressure is 0.1 mbar, while the temperature of the oxidation reactor 8 is maintained at 900℃ and the carbon dioxide pressure is 1 bar.

[0092] The waste heat recovery device 12 is connected to the reduction reactor 9 and the oxidation reactor 8. It recovers the heat released from the oxidation reaction to heat CeO2, and then introduces the heated CeO2 into the reduction reactor 9 to improve the energy utilization efficiency of the high-temperature solar-driven metal oxide thermochemical cycle fuel production device C.

[0093] The hydrogen refueling station D stores the hydrogen produced and transported by the photocatalytic water electrolysis hydrogen production device B and the high-temperature solar-driven metal oxide thermochemical cycle fuel production device C, and supplies it to hydrogen fuel cell vehicles according to user demand.

[0094] The solar-fuel conversion efficiency of the system is obtained from the following formula:

[0095]

[0096] Among them, H co The higher heating value of the carbon monoxide in the product is [value missing]. Q is the higher heating value of the product hydrogen gas; solar The solar radiation energy incident on the surface of the photocatalytic hydrolysis hydrogen production reactor and the inlet of the high-temperature thermochemical cycle fuel production reactor is referred to as solar radiation energy.

[0097] The high-temperature solar-driven metal oxide thermochemical cycle fuel production device C utilizes metal oxides to decompose water and carbon dioxide to produce hydrogen and carbon monoxide under high-temperature solar drive. The syngas is then transported to the Fischer-Tropsch synthesis unit E, where methane is synthesized under a catalyst and low-temperature environment. The synthesis product is then passed through a product separation unit F to separate the methane. The separated methane is then piped sequentially into the compressor chamber, combustion chamber, and gas turbine of the gas turbine G, ultimately driving a generator to produce electricity.

[0098] The Fischer-Tropsch synthesis unit E synthesizes liquid hydrocarbons from a mixture of carbon monoxide and hydrogen under low-temperature methanation catalyst BC-H-10 at a temperature of 150-200℃, and then connects to a product separation unit F to separate methane.

[0099] The product separation device F includes a demethanizer to separate methane from the Fischer-Tropsch synthesis reaction products, and is connected to a gas turbine G via a pipeline.

[0100] The gas turbine G burns the injected methane to generate high-temperature, high-pressure gas, which then enters the turbine to expand and do work, outputting electrical energy.

[0101] The power grid system H provides electricity to users.

[0102] The users in the residential and industrial park I consume electricity and fuel.

[0103] This invention uses a solar spectral divider A to direct ultraviolet and visible light into a parabolic concentrator 4 in a photocatalytic water splitting hydrogen production device B, while directing long-wavelength light into a heliostat 6 in a thermochemical fuel production device. Using commercially available TiO2 as a carrier, a Pt / TiO2 photocatalyst is designed and synthesized, focusing on the light absorption capacity of the semiconductor photocatalyst, the separation and migration of photogenerated carriers, and surface reaction kinetics. The photocatalytic water splitting hydrogen production device B employs a cylindrical photocatalytic reactor with a parabolic concentrator to produce hydrogen. Based on the solar incidence angle and radiation intensity, and according to optimization calculations, the original receiving half-angle of the parabolic concentrator, which maximizes the average energy-mass conversion efficiency under full sunlight conditions, is used, along with the truncated edge ray angle of the parabolic concentrator.

[0104] Before the photocatalytic water splitting hydrogen production reaction, a vacuum pump is used to remove air and other contaminants from the solar hydrogen production system, followed by the introduction of an inert gas and then a second vacuum. To reduce the energy consumption of the solution circulation pump, the photocatalytic hydrogen production system operates in natural circulation mode, and a high-pressure gas (i.e., hydrogen or nitrogen) is introduced every half hour to interfere with the deposited catalyst and maintain the solution in suspension.

[0105] The circular tube reactor 5 consists of heat-resistant glass tubes connected by flanges and sealed with PTFE gaskets. Hydrogen produced by solar photocatalysis flows along the tubes, collects at the gas outlet, and is then extracted and introduced into gas pipelines and a long-tube trailer. After long-term operation, low-activity photocatalysts and sacrificial agents are discharged from the reactor for post-treatment, while agglomerated photocatalysts and impurities are recovered.

[0106] The high-temperature solar-driven thermochemical cycle fuel production device C utilizes solar energy to drive a thermochemical cycle of CeO2 metal oxides to produce hydrogen and carbon monoxide. The reduction reactor 9 and oxidation reactor 8 employ cavity-type receivers to effectively capture the input solar radiation. A parabolic concentrator is installed at the aperture of the cavity reactor to increase reflectivity. A transparent quartz window is installed outside the parabolic concentrator to enclose the reaction chamber.

[0107] Heliostats reflect solar radiation and concentrate it onto the quartz window surfaces at the inlet of reduction reactor 9 and oxidation reactor 8. Reduction reactor 9 is maintained at 1500℃, and oxidation reactor 8 is maintained at 900℃. A two-step fuel production process using a high-temperature solar-driven thermochemical cycle of CeO2 (a metal oxide) to produce hydrogen and carbon monoxide is employed. First, CeO2 is introduced into reduction reactor 9. The products, along with carbon monoxide and water vapor, are then introduced into oxidation reactor 8, where CeO2, carbon monoxide, and hydrogen are generated. The CeO2 then enters waste heat recovery device 12 and returns to reduction reactor 9 for further reduction, achieving continuous fuel production. Carbon monoxide and hydrogen are then fed into Fischer-Tropsch synthesis unit E to produce methane.

[0108] The hydrogen produced by the photocatalytic water electrolysis hydrogen production device B and the hydrogen produced by the high-temperature solar-driven metal oxide thermochemical cycle fuel production device C are transported to hydrogen refueling stations for use in hydrogen fuel cell vehicles, and excess hydrogen is stored for use at night.

[0109] In the Fischer-Tropsch synthesis unit E, the domestically produced BC-H-10 low-temperature methanation catalyst is used, and the temperature is maintained at 150-200℃. The generated synthesis gas enters the product separation unit F, and the separated methane is fed into the gas turbine for power generation through the demethanizer tower, which is then connected to the grid to meet the electricity and fuel needs of users in residential and industrial parks.

[0110] The specific implementation steps of the operation control method of the present invention include:

[0111] Step 1: Establish a photocatalytic-high-temperature thermochemical coupled fuel production system model, including establishing a three-dimensional fluid dynamics-radiation transfer-chemical reaction dynamics coupled multiphysics field model for the circular tube reactor 5, oxidation reactor 8 and reduction reactor 9, and establishing an input-output nonlinear model based on experimental data for the waste heat recovery device 12, Fischer-Tropsch synthesis device E, product separation device F and gas turbine G.

[0112] Step 2: For the circular tube reactor 5, the hydrogen production was measured by changing the solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the tip to the bottom of the circular tube reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter, obtaining at least one thousand experimental samples; for the oxidation reactor 8 and reduction reactor 9, the hydrogen production and carbon monoxide production were measured by changing the solar radiation intensity, solar incidence angle, carbon dioxide and water vapor flow ratio, and CeO2 mass flow rate, obtaining at least one thousand experimental samples.

[0113] Step 3: Propose an improved Extreme Learning Machine, as detailed below;

[0114] Given a dataset Where s i For the i-th input sample, z i Let N be the number of output samples, and let N be the number of samples. The mathematical model of the Extreme Learning Machine is established as follows:

[0115] In the first model, the input variables are the operating parameters of the circular tube reactor, including solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the tip to the bottom of the reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter. The output parameter is the hydrogen production.

[0116] In the second model, the input variables are the operating parameters of the oxidation reactor and the reduction reactor, including solar radiation intensity, solar incidence angle, carbon dioxide and water vapor flow rate ratio, and CeO2 mass flow rate. The output parameters are hydrogen production and carbon monoxide production.

[0117] Hg=z (2)

[0118] Where H represents the model matrix; g represents the output weight vector; z represents the output value; the specific expression is:

[0119]

[0120]

[0121] Where φ(·) is the activation function; a i and v jDefine the i-th input weight and the i-th bias; g and h define the output weight and the input matrix; a1 and a n Represents the first and nth elements of the input weight vector a; v1 and v n Represents the first and nth elements of the bias vector v; g1 and g n Let g represent the first and nth elements of vector g; n represents the number of vectors.

[0122] ELM is a supervised learning method, and its training requires solving equation (2). Regularization is a preferred method for this task. However, the effectiveness of regularization depends on the reasonable selection of regularization parameters. Conventional experience-based selection methods are difficult to ensure the optimality of numerical solutions and increase the uncertainty of results. This invention proposes a new two-layer optimization training model to achieve adaptive selection of regularization parameters and training of model parameters, which can be expressed as the following mathematical model:

[0123]

[0124] In the formula, ||Hg-z||1 is the data fidelity term, which reduces the adverse effects of noise and bias; λ is the regularization parameter; ||z||1-||z||2 is the L1-2 norm, used to ensure the sparsity of numerical solutions; H T This is the model matrix corresponding to the validation set; z T This corresponds to the output data of the validation set; This indicates the search for the minimum value of the objective function with * as the variable; It is the objective function of the higher-level optimization problem, which attempts to find the optimal regularization parameter λ; is the objective function of the lower-level optimization problem, which attempts to compute the output weight g under given regularization parameters and training data; st represents the constraint condition; the upper-level optimization problem and the lower-level optimization problem constitute a bi-level optimization problem.

[0125] Equation (3) is a two-level optimization problem that requires solving two optimization problems. This invention proposes a novel nested algorithm to efficiently solve this problem.

[0126] First, we will discuss the solution of the lower-level optimization problem.

[0127] Lower-level optimization problem This is an optimization problem with a convex, differentiable function, which is extremely difficult to solve directly. When g is not a zero vector, the lower-level optimization problem is relaxed as follows:

[0128]

[0129] In the formula, v k As an auxiliary variable, it is defined as:

[0130] vk =-λg k / ||g k ||2(5)

[0131] In the formula, k represents the number of iterations; ||g k ||2 is g k L2 norm;

[0132] Note that equation (4) contains non-smooth terms, making it difficult to solve efficiently. To overcome this difficulty, a semi-quadratic splitting algorithm is used to alleviate the problem. By introducing two auxiliary variables, d1 and d2, equation (4) can be transformed into the following equality-constrained optimization problem:

[0133]

[0134] In the formula, z T Let z be the transpose of vector z.

[0135] According to the semi-quadratic splitting algorithm, equation (6) is further transformed into an unconstrained problem:

[0136]

[0137] In the formula, ψ(g,d1,d2) is the objective function, defined as:

[0138]

[0139] In the formula, g T μ1 and μ2 are the transpose of vector g; μ1 and μ2 are penalty parameters.

[0140] For ease of calculation, the separation optimization method is used to solve equation (8):

[0141]

[0142]

[0143]

[0144] According to equation (8), equations (9)-(11) are specifically as follows:

[0145]

[0146]

[0147]

[0148] In the formula, d 11 This represents the L1 norm of the auxiliary variable d1.

[0149] Equations (12) and (13) can be solved using the soft thresholding algorithm, i.e.:

[0150]

[0151]

[0152] In the formula, shrink(·,·) represents the soft threshold operator.

[0153] Equation (14) is differentiable, and its solution is:

[0154]

[0155] In the formula, H T H is the transpose of vector H; I is the identity matrix.

[0156] Equation (4) is solved iteratively according to equations (15)-(17). For ease of understanding, the entire calculation process is summarized in Algorithm 1 below. The significant feature of Algorithm 1 is that it uses relaxation methods and semi-quadratic splitting algorithms to reduce the difficulty of solving the problem and effectively handle non-smooth optimization problems. It only requires calculating the gradient vectors of subproblems, and the computational complexity is relatively low.

[0157]

[0158]

[0159] The decision variables in higher-level optimization problems are implicitly contained in the objective function, making it difficult to directly solve for the gradient vector or Hessian matrix. To overcome this difficulty, this invention uses a genetic algorithm to solve higher-level optimization problems.

[0160] To effectively address the hierarchical structure of bilevel optimization problems, and considering the unique characteristics of both the upper and lower levels, this invention proposes a novel nested algorithm. This algorithm uses a genetic algorithm to solve the upper-level optimization problem, nesting the lower-level optimization algorithm described above. The computational process of the algorithm is shown in the table below. This algorithm effectively utilizes the hierarchical structure of bilevel optimization problems and offers advantages such as speed and low computational complexity.

[0161]

[0162] Step 4: Based on experimental and computational data, an improved limit learning machine is used to approximate the nonlinear relationships between solar radiation intensity, solar incidence angle, parabolic concentrator opening direction, distance from the tip to the bottom of the reactor, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter and hydrogen production in the circular tube reactor 5. Simultaneously, for the oxidation reactor 8 and reduction reactor 9, the improved limit learning machine is used to approximate the nonlinear relationships between solar radiation intensity, solar incidence angle, carbon dioxide to water vapor flow ratio, CeO2 mass flow rate and hydrogen production, and carbon monoxide content in the oxidation reactor 8 and reduction reactor 9. A surrogate model is obtained, reducing computational and experimental time and costs.

[0163] Step 5: To mitigate the volatility and uncertainty of solar energy and user load, a robust optimization model for the energy system is established. A non-dominated sorting genetic algorithm is used to solve the proposed model. Under all-day illumination conditions, the opening direction of the parabolic concentrator, the distance from its tip to the bottom of the reactor, the inlet flow rate of the suspension, the concentration of the photocatalyst, and the diameter of the catalyst particles in the photocatalytic water electrolysis hydrogen production device B are adjusted in real time to match the energy provided by the parabolic concentrator with the solar radiation distribution required by the reactor. For the high-temperature solar-driven metal oxide thermochemical cycle fuel production device C, the carbon dioxide and water vapor flow ratio and the CeO2 mass flow rate are automatically adjusted by controlling the atomizer to maximize the solar-to-fuel energy conversion efficiency of the energy system and mitigate the impact of the uncertainty and intermittency of solar energy and user load on the stable operation of the system.

[0164] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A photocatalytic-high temperature thermochemical coupled fuel production system, characterized in that, It comprises: A solar spectrum splitter (A) which splits the solar spectrum into two bands according to the spectral response characteristics of the photocatalytic water-splitting catalyst and the high-temperature solar fuel production catalyst; The short-wave band of light is incident on the parabolic concentrator (4) of the photocatalytic water-splitting device (B), providing energy for the photocatalytic circular tube reactor (5) and triggering the photocatalytic water-splitting reaction; the long-wave band of light is incident on the heliostat (6) of the high-temperature solar fuel production device (C), which reflects the light to the inlet of the oxidation reactor (8) and the reduction reactor (9), driving the CeO2 thermochemical cycle to decompose water and carbon dioxide to produce hydrogen and carbon monoxide; The photocatalytic water-splitting device (B) uses a photocatalyst and a sacrificial agent to decompose water under sunlight to produce hydrogen and oxygen; The high-temperature solar fuel production device (C) uses a heliostat (6) to concentrate solar radiation energy to the inlet of the oxidation reactor (8) and the reduction reactor (9) at the top of the solar tower (7), producing methane for gas turbines (G) and hydrogen for hydrogenation stations (D); A waste heat recovery device (12) is provided to improve energy utilization efficiency; It also comprises: A hydrogenation station (D) for providing hydrogen to hydrogen fuel cell vehicles; A Fischer-Tropsch synthesis device (E) for synthesizing methane from a mixture of carbon monoxide and hydrogen; A product separation device (F) for separating methane from the products of the Fischer-Tropsch synthesis reaction; A gas turbine (G) for converting chemical energy into mechanical energy and then generating electricity to provide power to users; A power grid system (H) for providing power to users; A residential and industrial park (I) where users consume energy to meet the needs of life and industry.

2. The system according to claim 1, wherein the system is characterized by, The photocatalytic water-splitting device (B) comprises a vacuum pump (1), a first control system (2), a control atomizer (3), a parabolic concentrator (4), and a circular tube reactor (5). By measuring the hydrogen flow rate at the outlet of the circular tube reactor (5) and using the Monte Carlo ray tracing method to obtain the radiation distribution on the surface of the circular tube reactor (5), the first control system (2) adjusts the opening direction of the parabolic concentrator (4), the distance from the parabolic concentrator to the bottom of the circular tube reactor (5), the inlet flow rate of the suspension, the concentration of the photocatalyst, and the diameter of the catalyst particles in real time, so that the solar energy distribution on the surface of the circular tube reactor (5) matches the energy required by the chemical reaction process, maximizing the solar-to-fuel conversion efficiency.

3. The system according to claim 1, wherein the system is characterized by, The high-temperature solar-driven metal oxide thermochemical cycle fuel production device (C) comprises a heliostat (6), a solar tower (7), an oxidation reactor (8), a reduction reactor (9), a pump (10), a second control system (11) and a waste heat recovery device (12); the solar tower (7) and the heliostat (6) are installed with a solar tracking transmission system, the heliostat is adjusted in real time according to the solar incidence angle, and more than 95% of the solar radiation energy is gathered to the inlet of the reduction reactor (9) and the oxidation reactor (8), so that the solar-fuel conversion efficiency is improved.

4. The system according to claim 1, wherein the system is characterized by, The temperature of the reduction reactor (9) is kept at 1500 DEG C, the O2 pressure is 0.1 mbar, the temperature of the oxidation reactor (8) is kept at 900 DEG C, and the carbon dioxide pressure is 1 bar.

5. The system according to claim 1, wherein the system is characterized by: The waste heat recovery device (12) is connected with the reduction reactor (9) and the oxidation reactor (8), the heat release of the oxidation reaction is recovered to heat CeO2, the heated CeO2 is introduced into the reduction reactor (9), and the energy utilization efficiency of the solar-driven metal oxide thermochemical cycle fuel production device (C) is improved; the hydrogenation station (D) stores the hydrogen produced and transported by the photocatalytic hydrolysis hydrogen production device (B) and the solar-driven metal oxide thermochemical cycle fuel production device (C) and provides the hydrogen fuel cell vehicle according to the user demand.

6. The system according to claim 1, wherein the system is characterized by: The solar-fuel conversion efficiency of the system is obtained according to the following formula: （2） wherein, is the higher heating value of the product carbon monoxide; is the higher heating value of the product hydrogen; is the solar radiation energy incident to the surface of the photocatalytic round tube reactor and to the inlets of the oxidation reactor and the reduction reactor.

7. The system according to claim 1, wherein the system is characterized by: The Fischer-Tropsch synthesis device (E) synthesizes liquid hydrocarbons from the mixed gas of carbon monoxide and hydrogen under the condition that the low-temperature methanation catalyst BC-H-10 is at a temperature of 150-200 DEG C, and is connected with a product separation device (F) to separate methane; the product separation device (F) comprises a demethanizer to separate the methane in the Fischer-Tropsch synthesis reaction product, and is connected with a gas turbine (G) through a pipeline; the gas turbine (G) burns the injected methane to generate high-temperature and high-pressure gas, and then enters a turbine to expand and do work to output electric energy.

8. The method for operating and regulating a system for the production of fuel by photocatalytic-high temperature thermochemical coupling according to one of claims 1-7, characterized in that, The method comprises the following steps: Step one: a photocatalysis-high-temperature thermochemical coupling fuel production system model is established, a three-dimensional fluid dynamics-radiation transfer-chemical reaction kinetics coupled multi-physical field model is established for the circular tube reactor, the oxidation reactor and the reduction reactor, and an input-output nonlinear model is established for the waste heat recovery device, the Fischer-Tropsch synthesis device, the product separation device and the gas turbine according to experimental data; Step two: for the circular tube reactor, a plurality of experimental samples are obtained by measuring the hydrogen production amount by changing the solar radiation intensity, the solar incidence angle, the opening direction of the parabolic concentrator, the distance from the sharp top to the bottom of the circular tube reactor, the suspension inlet flow rate, the photocatalyst concentration and the catalyst particle diameter; for the oxidation reactor and the reduction reactor, a plurality of experimental samples are obtained by measuring the hydrogen production amount and the carbon monoxide amount by changing the solar radiation intensity, the solar incidence angle, the carbon dioxide and water vapor flow ratio and the CeO2 mass flow; Step three: an improved extreme learning machine algorithm is proposed; Step four: based on experimental data and calculation data, the nonlinear relationship between solar radiation intensity, solar incident angle, parabolic concentrator opening direction, distance from the tip to the reactor bottom, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter and hydrogen production in the pipe reactor is approximated using the extreme learning machine of step three; at the same time, for the oxidation reactor and the reduction reactor, the nonlinear relationship between solar radiation intensity, solar incident angle, carbon dioxide and water vapor flow ratio, CeO2 mass flow and hydrogen production, carbon monoxide content in the oxidation reactor and the reduction reactor is approximated using the extreme learning machine of step three; a proxy model is obtained to reduce calculation, experimental time and cost; Step five: a robust optimization model of the energy system is established, and a non-dominated sorting genetic algorithm is used to solve the proposed robust optimization model to adjust the parabolic concentrator opening direction, the distance from the tip to the reactor bottom, the suspension inlet flow rate, the photocatalyst concentration, and the catalyst particle diameter in the photocatalytic hydrolysis hydrogen production device in real time under all-day sunshine conditions, so that the energy provided by the parabolic concentrator matches the solar radiation distribution required by the reactor; for the high-temperature solar-driven metal oxide thermochemical cycle fuel production device, by controlling the atomizer, the carbon dioxide and water vapor flow ratio and the CeO2 mass flow are automatically adjusted to make the energy system obtain the maximum solar-fuel energy conversion efficiency and alleviate the impact of solar energy and user load uncertainty and intermittency on the stable operation of the system.

9. The operational regulation method according to claim 8, characterized in that, The extreme learning machine proposed in step three includes: Given a dataset where is the input sample, is the output sample, is the number of samples, the mathematical model of extreme learning machine is established by the following way: In the first model, the input variables are the operating parameters of the pipe reactor, including solar radiation intensity, solar incident angle, parabolic concentrator opening direction, distance from the tip to the reactor bottom, suspension inlet flow rate, photocatalyst concentration, and catalyst particle diameter, and the output parameter is hydrogen production; In the second model, the input variables are the operating parameters of the oxidation reactor and the reduction reactor, including solar radiation intensity, solar incident angle, carbon dioxide and water vapor flow ratio, and CeO2 mass flow, and the output parameters are hydrogen production and carbon monoxide content; （2） wherein represents a model matrix; represents an output weight vector; represents an output value; the specific expression is: and in, For activation functions; and Definition of the first The input weights and the first One bias; and The output weights and input matrix are defined; and Represents the input weight vector The first element and the nth element; and Represents the bias vector The first element and the nth element; and express The first and nth elements of the vector; Indicates the number of vectors; The regularization method is used to solve equation (2); The double-layer optimization training model realizes the adaptive selection of regularization parameters and the training of model parameters, and is expressed as the following mathematical model: （3） wherein is a data fidelity term, reducing the adverse effects of noise and bias; is a regularization parameter; is an L1-2 norm, used to ensure sparsity of the numerical solution; is a model matrix corresponding to the validation set; is output data corresponding to the validation set; denotes solving the minimization of the objective function with as variable; is the objective function of the upper optimization problem, trying to solve the optimal regularization parameter ; is the objective function of the lower optimization problem, trying to compute the output weights given the regularization parameter and the training data; s.t. denotes the constraint condition; the upper optimization problem and the lower optimization problem constitute a bi-level optimization problem.

10. The operational regulation method according to claim 9, wherein, The nested algorithm is used to solve the double-layer optimization problem of equation (3), which specifically includes: Lower level optimization problem When The lower level optimization problem is relaxed as follows when not a zero vector: （4） wherein are auxiliary variables, defined as: （5） wherein represents the number of iteration steps; is L2 norm; By introducing two auxiliary variables and , equation (4) is transformed into the following equality-constrained optimization problem: （6） wherein is the transpose of the vector ​ According to the semi-quadratic splitting algorithm, equation (6) is further transformed into an unconstrained problem: （7） In the formula, is the objective function, defined as: （8） wherein is the transpose of the vector and are penalty parameters;​ The separation optimization method is used to solve equation (8): （9） （10） （11） According to equation (8), equations (9)-(11) are specifically: （12） （13） （14） wherein denotes the L1 norm of the auxiliary variable denotes the L1 norm of the auxiliary variable Equations (12) and (13) are solved using the soft threshold algorithm, i.e.: （15） （16） In the formula, represents a soft threshold operator; Equation (14) is differentiable, and its solution is: （17） wherein is the identity matrix; Equation (4) is iteratively solved according to equations (15)-(17); The genetic algorithm is used to solve the upper-layer optimization problem.

Images