Flow control system and control method for hydrogen pressure reduction
By optimizing the flow channel structure and virtual state estimation model, differentiated opening compensation commands are generated, solving the problem of uneven flow distribution in the hydrogen depressurization system and realizing active balance of multi-valve flow and improved system stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2026-06-08
- Publication Date
- 2026-07-21
Smart Images
Figure CN122429323A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of fluid control technology, and more specifically, to a flow control system and control method for hydrogen decompression. Background Technology
[0002] Currently, hydrogen, as a clean and efficient energy carrier, is widely used in fuel cell vehicles and hydrogen internal combustion engines. Because hydrogen storage pressure is typically above 35 MPa, it cannot be directly supplied to the engine and must be reduced to a suitable operating range through a multi-stage pressure reduction system. In practical engineering applications, regardless of whether a series or parallel layout is used, the coordinated control of multiple secondary pressure reducing valves is crucial to ensuring stable system operation.
[0003] In existing technologies, multiple pressure-reducing valves arranged in series along the flow path exhibit inlet pressure differences between the distal and proximal valves. In parallel arrangements, multiple pressure-reducing valves are connected in parallel, making it difficult to achieve complete consistency in flow resistance and dynamic response characteristics across branches. These factors all contribute to variations in the actual opening degree of each pressure-reducing valve under the same control command, leading to uneven flow distribution. This uneven flow not only causes inconsistent wear on the valves, shortening component lifespan, but more seriously, it induces pressure fluctuations in the downstream common rail. These pressure fluctuations propagate backward through the fluid loop, further exacerbating the differences in inlet pressure among the upstream valves, creating a vicious cycle of "uneven load - differing opening responses - flow imbalance - pressure oscillation," ultimately resulting in decreased system control accuracy or even instability. Summary of the Invention
[0004] In view of this, the purpose of this application is to provide a flow control system and control method for hydrogen decompression, so as to overcome at least one of the above-mentioned defects.
[0005] In a first aspect, embodiments of this application provide a flow control system for hydrogen pressure reduction. The system includes: multiple pressure reducing valves connected in a hydrogen supply circuit; a flow channel structure for connecting the multiple pressure reducing valves to an upstream gas source and a downstream common rail; a boundary parameter acquisition unit for acquiring system boundary parameters, which include at least upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure; and a controller connected to the boundary parameter acquisition unit and each of the pressure reducing valves. The controller inputs the system boundary parameters into a virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve, and generates a differentiated opening compensation command for each pressure reducing valve based on the difference in the inlet virtual pressure of each pressure reducing valve, and outputs the differentiated opening compensation command to the corresponding pressure reducing valve to adjust its opening.
[0006] In one optional embodiment of this application, the flow channel structure includes a gas distribution chamber, wherein the gas distribution chamber has a preset length-to-diameter ratio, and the diameter of the pressure reducing valve inlet pipe at the distal end is larger than the diameter of the pressure reducing valve inlet pipe at the proximal end.
[0007] In one optional embodiment of this application, the flow channel structure includes a symmetrical Y-shaped flow branch cavity, wherein the symmetrical Y-shaped flow branch cavity has a preset bifurcation angle, and the length, curvature and surface roughness of each branch flow channel are consistent.
[0008] In one optional embodiment of this application, the controller is further configured to: compare the inlet virtual pressure of each pressure reducing valve, calculate the deviation between the inlet virtual pressure of each pressure reducing valve and a reference value, wherein the reference value is the average value of the inlet virtual pressure of all pressure reducing valves; and, based on the deviation, generate a compensation command to increase the opening degree for pressure reducing valves whose inlet virtual pressure is higher than the reference value, and generate a compensation command to decrease the opening degree for pressure reducing valves whose inlet virtual pressure is lower than the reference value.
[0009] In one optional embodiment of this application, the differentiated opening compensation instruction is a feedforward control instruction, wherein the controller is further configured to: generate a feedback control instruction, and superimpose the feedforward control instruction and the feedback control instruction and output them to each pressure reducing valve.
[0010] In one optional embodiment of this application, the boundary parameter acquisition unit includes a pressure sensor and a temperature sensor installed at the upstream storage tank, a pressure sensor installed at the outlet of the primary valve, and a pressure sensor installed at the downstream common rail.
[0011] In one optional embodiment of this application, the virtual state estimation model is pre-established by: acquiring experimental test data, which includes measured inlet pressures of each pressure reducing valve under different upstream tank pressures, upstream tank temperatures, and valve openings; acquiring computational fluid dynamics simulation data, which includes simulation results of flow field distribution under the operating conditions corresponding to the experimental test data; and fusing the experimental test data and the simulation data to fit the mapping relationship between system boundary parameters and inlet pressures of each pressure reducing valve, thereby generating an initial model.
[0012] In one optional embodiment of this application, the controller is further configured to: monitor the fluctuation amplitude of the downstream common rail pressure in real time; when the fluctuation amplitude exceeds a preset threshold, adjust the correction coefficient in the virtual state estimation model according to the magnitude of the fluctuation amplitude, so as to correct the virtual pressure at the inlet of each pressure reducing valve output by the model.
[0013] In one optional embodiment of this application, the virtual state estimation model is an embedded algorithm model obtained by dimensionality reduction and encapsulation of the initial model. The embedded algorithm model is deployed in the controller and is used to calculate the inlet virtual pressure of each pressure reducing valve in real time.
[0014] Secondly, embodiments of this application also provide a flow control method for hydrogen pressure reduction, the method comprising: acquiring system boundary parameters, the system boundary parameters including at least upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure; inputting the system boundary parameters into a virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve; generating a differentiated opening compensation command for each pressure reducing valve based on the difference in the inlet virtual pressure of each pressure reducing valve; and outputting the differentiated opening compensation command to the corresponding pressure reducing valve to adjust its opening.
[0015] This application provides a flow control system and method for hydrogen pressure reduction, comprising: multiple pressure reducing valves connected in a hydrogen supply circuit; a flow channel structure for connecting the multiple pressure reducing valves to an upstream gas source and a downstream common rail; a boundary parameter acquisition unit for acquiring system boundary parameters, which include at least the upstream storage tank pressure, upstream storage tank temperature, first-stage valve outlet pressure, and downstream common rail pressure; and a controller connected to both the boundary parameter acquisition unit and each pressure reducing valve. The controller inputs the system boundary parameters into a virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve, and generates differentiated opening compensation commands for each pressure reducing valve based on the differences in their inlet virtual pressures. These differentiated opening compensation commands are then output to the corresponding pressure reducing valves to adjust their openings. This application achieves flow balancing across multiple pressure reducing valves, improving system stability and control accuracy.
[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic diagram of the flow control system for hydrogen pressure reduction provided in an embodiment of this application; Figure 2 This is a front view of the first and second stage valves connected in series in the hydrogen supply circuit provided in the embodiments of this application; Figure 3 A top view of the first and second stage valves connected in parallel in the hydrogen supply circuit provided in the embodiments of this application; Figure 4 A schematic diagram of a parallel arrangement of multiple pressure reducing valves provided in an embodiment of this application; Figure 5 A schematic diagram of a series arrangement of multiple pressure reducing valves provided in an embodiment of this application; Figure 6 A flowchart for generating differentiated opening degree compensation instructions for each pressure reducing valve for the controller provided in the embodiments of this application; Figure 7 This is a flowchart of a flow control method for hydrogen decompression provided in an embodiment of this application. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. Based on the embodiments of this application, every other embodiment obtained by those skilled in the art without inventive effort falls within the scope of protection of this application.
[0020] First, the applicable scenarios for this application will be introduced. This application can be applied to the field of fluid control technology.
[0021] Further research revealed that in series or parallel configurations, when multiple secondary valves respond to the same control command simultaneously, differences in the actual inlet pressure of each valve due to factors such as flow channel structure and manufacturing tolerances will trigger the following chain reaction: First, uneven load leads to differences in opening response. That is, under the same control command, the valve with lower inlet pressure will have a larger actual opening, while the valve with higher inlet pressure will have a smaller opening. This is not due to different control commands, but rather the natural response characteristics of the actuator under non-uniform load. Second, differences in opening will directly amplify the instantaneous flow difference between branches, exacerbating the flow distribution imbalance. Finally, flow imbalance triggers system pressure oscillations. When the uneven flow from each branch converges at the outlet common rail, it interferes with each other and cannot be smoothly superimposed, resulting in low-frequency fluctuations in the downstream common rail pressure. These pressure fluctuations will then propagate in the reverse through the fluid loop, further interfering with the upstream flow field and the pressure before the valve, forming a vicious cycle of positive feedback: "uneven load - difference in opening response - flow imbalance - pressure oscillation - exacerbation of uneven load," ultimately damaging the overall stability and control accuracy of the system.
[0022] To address the aforementioned problem of uneven flow distribution, existing technical solutions often employ the method of adding physical sensors, i.e., installing pressure sensors or flow meters at each branch or valve inlet to monitor the status of each point in real time and implement feedback control. However, this solution has significant shortcomings: on the one hand, adding a large number of high-precision sensors significantly increases the system hardware cost and control complexity; on the other hand, the inherent hysteresis characteristic of feedback control makes it difficult to respond promptly under transient conditions, and it cannot fundamentally suppress pressure oscillations caused by uneven flow.
[0023] Based on this, this application provides a flow control system and method for hydrogen pressure reduction, which aims to solve the problem of uneven flow distribution among multiple pressure reducing valves caused by flow channel asymmetry, manufacturing tolerances and dynamic response differences in the prior art. By combining flow channel structure optimization with feedforward control based on virtual state estimation model, active equalization adjustment of the flow of each valve can be achieved without the need to add multiple physical sensors, so as to suppress downstream pressure fluctuations and improve system control accuracy and operational stability.
[0024] Please see Figure 1 , Figure 1 This is a schematic diagram of the flow control system for hydrogen pressure reduction provided in an embodiment of this application. Figure 1 As shown in the embodiment of this application, the flow control system 10 for hydrogen pressure reduction includes: Multiple pressure reducing valves 13 are connected in the hydrogen supply circuit.
[0025] The flow channel structure 14 is used to connect multiple pressure reducing valves 13 to the upstream gas source and the downstream common rail.
[0026] Boundary parameter acquisition unit 11 is used to acquire system boundary parameters, which include at least the upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure.
[0027] The controller 12 is connected to the boundary parameter acquisition unit 11 and multiple pressure reducing valves 13 respectively. The controller 12 inputs the system boundary parameters into the virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve. Based on the difference in the inlet virtual pressure of each pressure reducing valve, the controller 12 generates a differentiated opening compensation command for each pressure reducing valve and outputs the differentiated opening compensation command to the corresponding pressure reducing valve to adjust its opening.
[0028] Upstream gas source refers to hydrogen storage device that provides hydrogen to the system. It is usually a high-pressure hydrogen storage tank, and its outlet pipeline is connected to the system inlet.
[0029] The downstream common rail is a manifold located on the outlet side of multiple pressure reducing valves. It is used to collect the hydrogen output from each pressure reducing valve and supply it to the downstream engine. Its internal pressure is the downstream common rail pressure.
[0030] The flow channel structure refers to the fluid passage component that connects the upstream gas source to multiple pressure reducing valves. Its specific form is determined according to the layout of the pressure reducing valves: when a series layout is adopted, the flow channel structure is a gas distribution chamber, which has multiple valve ports opened sequentially along the airflow direction for installing pressure reducing valves; when a parallel layout is adopted, the flow channel structure is a symmetrical Y-shaped diversion chamber, which divides one intake gas into multiple branches, and pressure reducing valves are installed at the end of each branch.
[0031] Upstream tank pressure refers to the pressure inside the hydrogen cylinder, which is transmitted to the electronic control unit by hydrogen cylinder components (such as the cylinder valve).
[0032] The initial value of the upstream storage tank temperature is the ambient temperature, obtained based on the intake and exhaust system sensors.
[0033] The outlet pressure of the first-stage valve refers to the pipeline pressure after the pressure is reduced by the first-stage pressure reducing valve and before entering the flow channel structure. It is measured in real time by a pressure sensor installed on the outlet pipeline of the first-stage pressure reducing valve.
[0034] Downstream common rail pressure refers to the hydrogen pressure inside the downstream common rail, which is measured in real time by a pressure sensor installed on the downstream common rail.
[0035] The four boundary parameters mentioned above together constitute the input of the virtual state estimation model, which is used to calculate the virtual pressure at the inlet of each pressure reducing valve in real time.
[0036] This application breaks the vicious cycle described above through the following closed-loop process: First, high-precision sensors are integrated and computational fluid dynamics simulations are used to conduct full-domain testing. By fusing experimental and simulation data, the inherent laws governing the pressure field distribution and flow allocation within the system under different inlet pressures, temperatures, and valve commands are revealed. A physical mechanism and data fusion model that can accurately describe the relationship between boundary measurable parameters and the virtual states of each valve is established. Second, the high-fidelity model is engineered and encapsulated, transforming it into an efficient embedded algorithm model. In the actual system, only a few boundary pressure signals, such as upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure, need to be read. The model can then calculate the inlet virtual pressure of each pressure reducing valve in real time. Then, the controller compares the differences in the virtual states of each valve calculated by the model and proactively and proactively issues differentiated opening commands to each valve based on these differences, curbing the trend of uneven flow at the source. Finally, the system utilizes the continuously monitored downstream common rail pressure fluctuation characteristics during operation to fine-tune and learn online the built-in model, enabling it to adapt to the performance degradation or changes of the system over time and ensuring control accuracy throughout its entire lifecycle.
[0037] The system comprises multiple secondary pressure-reducing valves, a flow channel structure, a boundary parameter acquisition unit, and a controller. The pressure-reducing valves are integrated into the hydrogen supply circuit in series or parallel, depending on actual gas supply requirements. The flow channel structure serves as a connecting hub, with its upstream end connected to the hydrogen storage tank and its downstream end converging into the common rail to supply gas to the engine. The boundary parameter acquisition unit specifically includes pressure and temperature sensors installed at the outlet of the hydrogen storage tank for real-time monitoring of the upstream tank pressure and temperature; a pressure sensor installed at the outlet of the primary pressure-reducing valve for acquiring the primary valve outlet pressure; and a pressure sensor installed at the downstream common rail for acquiring the downstream common rail pressure. The controller is connected to each of the aforementioned sensors and the actuators of each secondary pressure-reducing valve via signal lines.
[0038] During system operation, the controller first reads four boundary parameters in real time through the boundary parameter acquisition unit: upstream tank pressure, upstream tank temperature, primary valve outlet pressure, and downstream common rail pressure. These four parameters are then fed into a pre-deployed virtual state estimation model within the controller. This model, trained based on fluid dynamics mechanisms and extensive experimental data, can calculate the instantaneous pressure value at the inlet of each secondary pressure reducing valve in the flow channel structure—the virtual inlet pressure—based on the aforementioned boundary conditions. This allows for soft measurement of the inlet state of each valve without the need for additional sensors.
[0039] Subsequently, the controller compares the calculated virtual inlet pressures of each pressure-reducing valve and identifies the differences between them. Based on these differences, the controller generates differentiated opening compensation commands for each pressure-reducing valve: for valves with higher virtual inlet pressures, the opening command is appropriately increased to overcome the higher flow resistance; for valves with lower virtual inlet pressures, the opening command is correspondingly decreased to avoid excessive flow. These differentiated commands are output to the actuators of each valve in real time, precisely adjusting the valve core position to make the flow in each branch more consistent. The entire process proactively intervenes before pressure fluctuations occur, suppressing flow imbalance caused by uneven inlet pressure at the source and effectively breaking the vicious cycle of "uneven load - difference in opening response - flow imbalance - pressure oscillation".
[0040] In the embodiments of this application, please refer to Figure 2 , Figure 2 This is a front view, or axial sectional view, of the first and second stage valves connected in series in the hydrogen supply circuit provided in the embodiments of this application. Figure 2 As shown in the diagram, this schematic illustrates the connection relationship between the primary valve 23, the first secondary valve 21, the second secondary valve 22, and the gas distribution chamber 24 in the hydrogen supply circuit. "P" indicates a pressure sensor located at the outlet of the primary valve, used to collect the outlet pressure of the primary valve in real time. This pressure is input to the controller as one of the system boundary parameters.
[0041] In another embodiment of this application, please refer to Figure 3 , Figure 3 This is a top view showing the first and second stage valves of the hydrogen supply circuit provided in an embodiment of this application connected in parallel. Figure 3 As shown in the diagram, the schematic diagram further illustrates the gas path and flow channel details in the hydrogen supply circuit, including the connection and medium flow relationship of the primary valve 23, the first secondary valve 21 (proximal secondary valve), the second secondary valve 22 (remote secondary valve), and the gas distribution chamber 24.
[0042] Hydrogen gas enters the gas distribution chamber 24 from upstream after being depressurized by the primary valve 23. The gas distribution chamber 24 is a connecting flow channel within the valve body, which evenly distributes the hydrogen gas from the primary valve outlet to the inlet sides of the first and second secondary valves 21 and 22. The first and second secondary valves 21 and 22 are arranged in parallel, respectively performing secondary depressurization and regulation on the distributed hydrogen gas before finally outputting it to the downstream common rail. Additionally, a concentric annular groove structure is formed between each secondary valve and the valve body; this concentric annular groove is also part of the gas distribution chamber, further optimizing the uniformity of hydrogen distribution and reducing the inlet pressure difference between the near and far secondary valves, providing a basis for the controller to achieve differentiated opening compensation.
[0043] Please refer to Figure 4as well as Figure 5 , Figure 4 This is a schematic diagram of a parallel arrangement of multiple pressure reducing valves provided in an embodiment of this application. Figure 5 This is a schematic diagram of a series arrangement of multiple pressure reducing valves provided in an embodiment of this application.
[0044] like Figure 4 As shown, the flow channel structure includes a symmetrical Y-shaped flow branch cavity. The symmetrical Y-shaped flow branch cavity has a preset bifurcation angle, and the length, curvature and surface roughness of each branch flow channel are consistent.
[0045] Figure 4 Corresponding to Figure 3 In a series-connected hydrogen pressure-reducing system, multiple pressure-reducing valves are sequentially installed on the gas distribution chamber along the gas flow direction (arrow direction). The length-to-diameter ratio (L / D) of the gas distribution chamber is controlled between 2.5 and 4.0. Within this range, the gas flows through the distribution chamber and maintains a uniform pressure gradient, avoiding excessive pressure loss or airflow disturbance along the flow path. Multiple valve ports are opened on the distribution chamber. The inlet diameter d2 of the downstream distal valve port is slightly larger than the inlet diameter d1 of the upstream proximal valve port, with a ratio d2 / d1 of 1.05 to 1.15. The front chamber volume ratio is controlled between 0.9 and 1.1. By increasing the distal flow area, pressure loss along the flow path is compensated, allowing the distal valve to obtain an inlet pressure closer to that of the proximal valve.
[0046] During installation, the air inlet of the distribution chamber is connected to the outlet of the upstream primary valve, and the air outlet is connected to the downstream common rail. Each valve is installed at its corresponding port according to the airflow sequence. The direct effect of this design is that when multiple pressure-reducing valves are opened simultaneously, the inlet pressure of the distal valve is increased, reducing the pressure difference with the proximal valve. This makes the flow rates of each valve more similar under the same opening command, achieving static flow balancing at the hardware level. This initial balancing reduces the burden of dynamic adjustment by the controller, achieving flow balancing without excessive compensation, thereby improving system response speed and control accuracy.
[0047] Because the parallel lengths are the same, parallel valves have no near or far ends.
[0048] In a parallel-configuration hydrogen pressure-reducing system, multiple pressure-reducing valves are installed in parallel on a symmetrical Y-shaped flow-splitting chamber. The inlet of this flow-splitting chamber is connected to the outlet of the upstream primary valve, and the outlet branches into multiple symmetrically distributed branch channels, with a pressure-reducing valve installed at the end of each branch. The flow-splitting chamber adopts a Y-shaped symmetrical structure, with the bifurcation angle controlled within the range of 45 to 60 degrees. This angle range is the optimal value obtained through fluid dynamics simulation and experimental verification. At this angle, the eddy current loss is minimized when the airflow passes through the bifurcation, and the airflow distribution between the two branches is most uniform. If the bifurcation angle is too small, the two branches are too close, and the airflow interferes with each other at the bifurcation; if the bifurcation angle is too large, the airflow turns sharply, and the local resistance increases.
[0049] like Figure 5 As shown, the flow channel structure includes a gas distribution chamber, wherein the gas distribution chamber has a preset length-to-diameter ratio, and the inlet diameter of the pressure reducing valve located at the distal end is larger than the inlet diameter of the pressure reducing valve located at the proximal end.
[0050] Figure 5 Corresponding to Figure 2 In terms of structural implementation, the flow channels of each branch are precision-machined to ensure highly consistent geometric parameters: the centerline length of each branch is equal, the curvature is consistent, and the surface roughness of the inner wall is controlled within ±2% tolerance. The length, curvature, and surface roughness of each branch flow channel are consistent, with tolerances controlled within ±2%, to achieve static flow balance and eliminate flow resistance differences caused by machining errors. During installation, each pressure reducing valve is fixed on the valve seat at the end of the branch, with the valve seat sealing surface perpendicular to the flow channel.
[0051] The direct technical effect of this structural design is that, since the geometric parameters of each branch flow channel are completely identical, their inherent flow resistance characteristics are basically the same. When multiple pressure reducing valves open to the same degree simultaneously, the gas flow rate of each branch is theoretically equal, achieving static flow balancing. This symmetrical design at the hardware level eliminates flow distribution deviations caused by differences in flow channels, ensuring consistent inlet pressure for each valve, providing a reliable hardware foundation for precise controller control, and improving the coordinated control accuracy of parallel valve groups.
[0052] For further details, please refer to Figure 6 , Figure 6 A flowchart is provided for generating differentiated opening degree compensation instructions for each pressure reducing valve for the controller provided in the embodiments of this application.
[0053] S101. Compare the virtual inlet pressure of each pressure reducing valve and calculate the deviation between the virtual inlet pressure of each pressure reducing valve and the reference value.
[0054] The baseline value can be the average of the virtual inlet pressures of all pressure reducing valves, or a pressure reference value preset according to the system's target operating conditions. The controller calculates the deviation of the virtual inlet pressure of each pressure reducing valve from the baseline value through differential calculation. This deviation quantifies the uneven load on each valve due to flow channel asymmetry, manufacturing tolerances, or differences in dynamic response.
[0055] S102. Based on the deviation, generate a compensation command to increase the opening degree for pressure reducing valves whose inlet virtual pressure is higher than the reference value, and generate a compensation command to decrease the opening degree for pressure reducing valves whose inlet virtual pressure is lower than the reference value.
[0056] Specifically, for a pressure-reducing valve with a virtual inlet pressure higher than the baseline value, it means that the actual pressure at the valve inlet is high and the flow resistance is large. The controller generates a compensation command to increase the valve opening, thus appropriately increasing the valve opening to overcome the high flow resistance. For a pressure-reducing valve with a virtual inlet pressure lower than the baseline value, it means that the actual pressure at the valve inlet is low and the flow capacity is relatively excessive. The controller generates a compensation command to decrease the valve opening, thus appropriately limiting the opening to avoid excessive flow. This compensation command, combined with the basic opening command, forms the final control signal, which is output to the actuators of each valve.
[0057] The direct technical effect of the aforementioned compensation command generation process is that the controller no longer sends the same opening command to all pressure reducing valves, but instead makes differentiated adjustments based on the real-time virtual inlet pressure status of each valve. For valves with higher inlet pressure, increasing the opening command can increase their actual flow rate; for valves with lower inlet pressure, decreasing the opening command can suppress excessive flow. This feedforward compensation mechanism based on virtual pressure differences proactively intervenes before pressure fluctuations actually occur, suppressing flow distribution deviations caused by uneven inlet pressure at the source, thereby breaking the vicious cycle and achieving proactive balanced control of multi-valve flow.
[0058] In this embodiment of the application, the differentiated opening compensation instruction is a feedforward control instruction, wherein the controller is further configured to generate a feedback control instruction and output the feedforward control instruction and the feedback control instruction superimposed to each pressure reducing valve.
[0059] In this embodiment, the differential opening compensation command is configured as a feedforward control command. This feedforward control command is generated in real time based on the difference in virtual pressure at the inlet of each pressure reducing valve. It is used to proactively adjust the opening of each valve before pressure fluctuations occur, suppressing the trend of uneven flow at the source. Simultaneously, the controller is also configured to generate a feedback control command. This feedback control command is calculated based on the deviation between the actual detected value and the target value of the downstream common rail pressure, and is used to correct static errors during system operation.
[0060] During the control command output phase, the controller superimposes the feedforward control command and the feedback control command to generate the final control signal, which is then output to the actuators of each pressure reducing valve. The superposition method can be additive, where the final opening command equals the basic opening command plus the feedforward compensation plus the feedback correction; or it can be weighted superposition or other fusion methods, depending on the system control strategy.
[0061] This application demonstrates that feedforward control commands can rapidly respond to real-time differences in inlet pressure of each valve, proactively compensating for flow unevenness caused by flow channel asymmetry or inconsistent valve responses, and suppressing downstream pressure fluctuations. Feedback control commands, on the other hand, can eliminate static deviations accumulated during system operation due to model errors, environmental changes, or component aging, ensuring that the downstream common rail pressure remains stable near the target value. The combined control command possesses both the speed and predictability of feedforward control and the accuracy and robustness of feedback control, thereby achieving high-precision balanced flow control of multiple pressure-reducing valves across the entire operating range.
[0062] In this embodiment of the application, a virtual state estimation model is pre-established in the following manner: Obtain experimental test data, including the measured inlet pressure of each pressure reducing valve under different upstream tank pressures, upstream tank temperatures, and valve openings.
[0063] First, experimental test data was obtained. This data was acquired during the system development phase by setting up a test platform and measuring the pressure values at the inlet of each pressure reducing valve under different upstream tank pressures, temperatures, and valve openings. These data accurately reflect the pressure distribution characteristics of the system under actual operating conditions.
[0064] Obtain computational fluid dynamics simulation data, including simulation results of flow field distribution under operating conditions corresponding to experimental test data.
[0065] Secondly, computational fluid dynamics (CFD) simulation data is obtained. This simulation data is obtained by numerically simulating the internal flow field of the system using CFD software under the same operating conditions as the experimental test data. The simulation results include pressure distribution, velocity distribution, and streamline distribution within the flow channel. The simulation data can compensate for the limitations of experimental tests, which have a limited number of measuring points and cannot comprehensively capture the details of the flow field.
[0066] By fusing experimental test data with simulation data, the mapping relationship between system boundary parameters and the inlet pressure of each pressure reducing valve is fitted, and an initial model is generated.
[0067] In one specific embodiment, the fusion and fitting process employs a multi-layer feedforward neural network. Using system boundary parameters (upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure) as inputs and the inlet pressure of each pressure-reducing valve as outputs, experimental and simulation data are collected as training samples to train the neural network, thereby establishing a mapping relationship between inputs and outputs. Then, the trained neural network undergoes structured pruning and quantization to remove redundant nodes, transforming it into a lightweight model that can run efficiently in an embedded controller.
[0068] Then, the experimental test data and simulation data are fused. Fusion methods can include data fitting, parameter identification, or machine learning. Correlation analysis is performed on corresponding data from both methods under the same operating conditions to fit the mapping relationship between system boundary parameters and the inlet pressure of each pressure reducing valve. This mapping relationship is presented in the form of a mathematical expression, thus generating the initial model. This initial model can accurately calculate the pressure value at the inlet of each pressure reducing valve based on four boundary parameters: upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure.
[0069] The virtual state estimation model is an embedded algorithm model obtained by dimensionality reduction and encapsulation of the initial model. The embedded algorithm model is deployed in the controller and is used to calculate the inlet virtual pressure of each pressure reducing valve in real time.
[0070] Since the initial model may contain complex fluid dynamics equations or high-dimensional data tables, the computational load is large, making it difficult to run directly in real time on an embedded controller. Therefore, the initial model undergoes dimensionality reduction processing, such as by simplifying equations, using lookup tables for interpolation, or neural network compression. In another specific embodiment, the dimensionality reduction processing employs principal component analysis to project the high-dimensional CFD simulation data into a low-dimensional space and extract the main features; or it uses response surface methodology to approximate the original complex model with low-order polynomials. The encapsulated embedded algorithm model is implemented in C language functions, taking boundary parameters as input and outputting the virtual pressure at each valve inlet. This reduces the computational complexity of the model and encapsulates it into an embedded algorithm model suitable for running on controller hardware.
[0071] The embedded algorithm model is ultimately deployed in the controller. When the system is actually running, the controller will input the four boundary parameters collected in real time into the model, which can quickly calculate the inlet virtual pressure of each pressure reducing valve for subsequent differentiated opening compensation control.
[0072] Furthermore, the controller is also configured as follows: The fluctuation amplitude of the downstream common rail pressure is monitored in real time. When the fluctuation amplitude exceeds a preset threshold, the correction coefficient in the virtual state estimation model is adjusted according to the magnitude of the fluctuation amplitude to correct the virtual pressure at the inlet of each pressure reducing valve output by the model.
[0073] In one specific embodiment, the correction factor is a multiplicative correction factor for the virtual inlet pressure of each pressure reducing valve output by the model, with an initial value of 1. When the downstream common rail pressure fluctuation amplitude exceeds a preset threshold, the controller calculates the ratio of the fluctuation amplitude to the threshold and adjusts the correction factor to its original value multiplied by (1 + proportional coefficient × overshoot). If the pressure remains too high, the correction factor is gradually decreased; if the pressure remains too low, the correction factor is gradually increased. The proportional coefficient is calibrated through bench testing.
[0074] During system operation, the controller continuously receives pressure signals collected in real time by the pressure sensor at the downstream common rail pipe, processes these signals, and extracts the fluctuation characteristics of the downstream common rail pressure, specifically calculating the pressure fluctuation amplitude. The fluctuation amplitude can be calculated using the difference between the maximum and minimum values within a sliding window, or by using the root mean square value, depending on the system sampling frequency and control requirements.
[0075] The controller has a preset fluctuation amplitude threshold, which is set according to the system's allowable pressure fluctuation range, for example, ±2% or ±5% of the target pressure. When the fluctuation amplitude calculated in real time does not exceed the preset threshold, it indicates that the output accuracy of the current virtual state estimation model meets the control requirements, and the model parameters remain unchanged. When the fluctuation amplitude exceeds the preset threshold, it indicates that the virtual pressure at the inlet of each pressure reducing valve output by the model deviates from the actual operating condition, causing the control command to fail to effectively suppress pressure fluctuations. At this time, the controller initiates the model correction program.
[0076] The specific operation of the correction procedure is as follows: Adjust the correction coefficients in the virtual state estimation model according to the magnitude of the fluctuation. The correction coefficients can be one or more adjustable parameters in the model, such as the flow coefficient, flow resistance coefficient, or pressure loss coefficient. The adjustment method can be proportional, meaning the larger the fluctuation amplitude, the larger the adjustment of the correction coefficient; or stepwise, meaning the correction coefficient is adjusted by a fixed step size every time the fluctuation amplitude exceeds a preset level. The direction of adjustment of the correction coefficients is preset based on the sign of the fluctuation amplitude or the system characteristics. For example, if the downstream pressure remains consistently high, the virtual pressure value output by the model will be appropriately reduced.
[0077] The adjusted correction coefficients are incorporated into the virtual state estimation model for subsequent inlet virtual pressure calculations. This online adaptive correction mechanism enables the model to be dynamically fine-tuned based on the actual operating state of the system, compensating for model deviations caused by factors such as component aging, environmental changes, or wear, ensuring that the virtual state estimation model maintains high computational accuracy throughout its entire lifecycle, thereby maintaining the stability and accuracy of the system's flow balance control.
[0078] The flow control system and method for hydrogen depressurization provided in this application, compared with the prior art scheme that relies on adding multiple physical sensors for feedback control, solves the problems of uneven flow distribution, downstream pressure fluctuations and decreased system control accuracy caused by flow channel asymmetry, manufacturing tolerances and dynamic response differences in traditional multi-valve layouts. This is achieved by optimizing the hardware design of the flow channel structure, combining feedforward control and feedback control based on a virtual state estimation model for coordinated adjustment, and introducing an online adaptive correction mechanism for the model. It achieves the beneficial effects of active flow balancing of multiple valves, suppressing pressure oscillations, and improving the stability and control accuracy of the system throughout its entire life cycle without increasing hardware costs.
[0079] Please see Figure 7 , Figure 7 A flowchart illustrating a flow control method for hydrogen pressure reduction provided in an embodiment of this application. Figure 7 As shown in the embodiments of this application, the flow control method for hydrogen depressurization includes: S201. Obtain system boundary parameters, which include at least the upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure. S202. Input the system boundary parameters into the virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve; S203. Generate differentiated opening compensation commands for each pressure reducing valve based on the differences in the virtual inlet pressure of each pressure reducing valve. S204. Output the differentiated opening compensation command to the corresponding pressure reducing valve to adjust its opening.
[0080] The above method uses system boundary parameters as input and a pre-established virtual state estimation model to calculate the virtual pressure at the inlet of each pressure reducing valve in real time, thereby achieving the perception of the load state of each valve without adding physical sensors. Then, based on the differences in virtual pressure among the valves, it generates differentiated opening compensation commands to actively adjust the valve openings to compensate for flow unevenness caused by flow channel asymmetry, manufacturing tolerances, or differences in dynamic response. This method proactively intervenes before pressure fluctuations occur, suppressing flow distribution imbalance at its source, and simultaneously employs subsequent online adaptive correction mechanisms (such as... Figure 4 (As shown) Ensure that the model accuracy remains stable over the system running time, thereby achieving high-precision flow balance control across the entire operating range.
[0081] This application aims to fundamentally solve the problem of flow distribution imbalance caused by uneven dynamic load when multiple secondary pressure reducing valves are opened simultaneously in a series or parallel configuration. This problem not only causes uneven wear of components but also leads to system-level pressure oscillations and control instability. Specifically, this application aims to: achieve high-precision real-time estimation of inlet pressure or branch flow of multiple valves without relying on multiple high-cost physical sensors; perform intelligent feedforward compensation of valve opening based on the estimated values to actively suppress flow imbalance; and ensure high stability, high response speed, and long-term operational consistency of the system under all operating conditions through hardware and software co-design.
[0082] The key innovations of this application include: First, providing a flow balance control architecture suitable for series and parallel hydrogen pressure reducing valves, which includes hardware flow channel optimization design and virtual sensing control closed loop; Second, employing virtual pressure or virtual flow sensing methods, based on a pre-calibrated model, to predict the state of each valve or branch in real time according to upstream measurable parameters for control decision-making; Third, adopting a collaborative feedforward-feedback control strategy to dynamically adjust the opening degree of each valve according to the differences in virtual states; Fourth, introducing an online adaptive correction mechanism for the model to ensure long-term operational accuracy; Fifth, providing specific hardware design parameters and matching criteria for series and parallel layouts.
[0083] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0084] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.
[0085] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0086] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0087] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0088] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A flow control system for hydrogen pressure reduction, characterized in that, include: Multiple pressure reducing valves are connected to the hydrogen supply circuit; The flow channel structure is used to connect the plurality of pressure reducing valves to the upstream gas source and the downstream common rail; A boundary parameter acquisition unit is used to acquire system boundary parameters, which include at least the upstream tank pressure, the upstream tank temperature, the first-stage valve outlet pressure, and the downstream common rail pressure. The controller is connected to the boundary parameter acquisition unit and each of the pressure reducing valves. The controller inputs the system boundary parameters into the virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve. Based on the difference in the inlet virtual pressure of each pressure reducing valve, the controller generates a differentiated opening compensation command for each pressure reducing valve and outputs the differentiated opening compensation command to the corresponding pressure reducing valve to adjust its opening.
2. The system according to claim 1, characterized in that, The flow channel structure includes a gas distribution chamber. The gas distribution chamber has a preset length-to-diameter ratio, and the inlet diameter of the pressure reducing valve located at the far end is larger than the inlet diameter of the pressure reducing valve located at the near end.
3. The system according to claim 1, characterized in that, The flow channel structure includes a symmetrical Y-shaped flow splitter cavity. The symmetrical Y-shaped flow branch cavity has a preset bifurcation angle, and the length, curvature and surface roughness of each branch flow channel are consistent.
4. The system according to claim 1, characterized in that, The controller is also configured to: Compare the inlet virtual pressure of each pressure reducing valve, and calculate the deviation between the inlet virtual pressure of each pressure reducing valve and the reference value, where the reference value is the average value of the inlet virtual pressure of all pressure reducing valves. Based on the deviation, a compensation command to increase the opening degree is generated for pressure reducing valves with inlet virtual pressure higher than the reference value, and a compensation command to decrease the opening degree is generated for pressure reducing valves with inlet virtual pressure lower than the reference value.
5. The system according to claim 4, characterized in that, The differentiated opening compensation command is a feedforward control command. The controller is further configured as follows: A feedback control command is generated, and the feedforward control command and the feedback control command are superimposed and output to each pressure reducing valve.
6. The system according to claim 1, characterized in that, The boundary parameter acquisition unit includes a pressure sensor and a temperature sensor installed at the upstream storage tank, a pressure sensor installed at the outlet of the primary valve, and a pressure sensor installed at the downstream common rail.
7. The system according to claim 1, characterized in that, The virtual state estimation model is pre-established using the following method: Acquire experimental test data, which includes the measured inlet pressure of each pressure reducing valve under different upstream tank pressures, upstream tank temperatures, and valve openings; Obtain computational fluid dynamics simulation data, which includes simulation results of flow field distribution under the operating conditions corresponding to the experimental test data; The experimental test data and the simulation data are fused together to fit the mapping relationship between the system boundary parameters and the inlet pressure of each pressure reducing valve, and an initial model is generated.
8. The system according to claim 1, characterized in that, The controller is also configured to: Real-time monitoring of the fluctuation amplitude of downstream common rail pressure; When the fluctuation amplitude exceeds a preset threshold, the correction coefficient in the virtual state estimation model is adjusted according to the magnitude of the fluctuation amplitude to correct the virtual pressure at the inlet of each pressure reducing valve output by the model.
9. The system according to claim 7, characterized in that, The virtual state estimation model is an embedded algorithm model obtained by dimensionality reduction and encapsulation of the initial model. The embedded algorithm model is deployed in the controller and is used to calculate the inlet virtual pressure of each pressure reducing valve in real time.
10. A flow control method for hydrogen depressurization, characterized in that, include: Obtain system boundary parameters, which include at least the upstream tank pressure, upstream tank temperature, first-stage valve outlet pressure, and downstream common rail pressure; The system boundary parameters are input into the virtual state estimation model to obtain the inlet virtual pressure of each pressure reducing valve; Based on the differences in the virtual inlet pressure of each pressure reducing valve, generate differentiated opening compensation commands for each pressure reducing valve; The differentiated opening compensation command is output to the corresponding pressure reducing valve to adjust its opening.