Control method for communicating pipe of hydrogen production system

Abstract

The invention relates to the technical field of water electrolysis hydrogen production, in particular to a hydrogen production system communicating pipe control method which comprises the following steps: synchronously acquiring liquid level data of a hydrogen separator and an oxygen separator, oxygen concentration data in hydrogen and hydrogen concentration data in oxygen; calculating a liquid level difference based on the collected liquid level data, and performing signal preprocessing on the gas concentration data; the liquid level difference and the preprocessed gas concentration data are input into a control model to be processed; and comparing the liquid level difference with a liquid level difference threshold value through a control model, and comparing the gas concentration data with a corresponding gas concentration threshold value. According to the method, a multi-threshold coupling judgment mechanism of the gas purity parameter and the liquid level difference is introduced, and then the hierarchical control instruction is triggered according to the standard exceeding grade, so that the problems that single liquid level balance control is mostly adopted in a traditional method, and due to the lack of an active blocking mechanism for oxygen in hydrogen and hydrogen diffusion in oxygen, the potential safety hazard is caused are solved. And the risk that the gas purity is out of control is high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrogen production by electrolysis of water, and particularly relates to a hydrogen production system communication pipe control method. BACKGROUND

[0002] With the acceleration of global energy transformation and the promotion of the "double carbon" target, hydrogen energy, as an important clean secondary energy, is increasingly prominent. Hydrogen production by electrolysis of water, especially green hydrogen production driven by renewable energy, is a key path to achieve industrial decarbonization and large-scale consumption of renewable energy. In the mainstream technology route of alkaline electrolysis of water for hydrogen production, the safety of system operation and the quality of hydrogen product are the core factors restricting its large-scale commercial application. Traditional methods mostly use single liquid level balance control, which lacks an active blocking mechanism for hydrogen diffusion in oxygen and oxygen diffusion in hydrogen, resulting in high risk of gas purity out of control. SUMMARY

[0003] In order to make up for the above shortcomings, the present application provides a hydrogen production system communication pipe control method, which aims to improve the problem that traditional methods mostly use single liquid level balance control, which lacks an active blocking mechanism for hydrogen diffusion in oxygen and oxygen diffusion in hydrogen, resulting in high risk of gas purity out of control.

[0004] The present application provides the following technical scheme, a hydrogen production system communication pipe control method, comprising the following steps:

[0005] S1, synchronously collecting liquid level data, oxygen concentration data in hydrogen and hydrogen concentration data in oxygen of a hydrogen separator and an oxygen separator;

[0006] S2, calculating a liquid level difference based on the collected liquid level data, and performing signal preprocessing on the gas concentration data;

[0007] S3, inputting the liquid level difference and the preprocessed gas concentration data into a control model for processing;

[0008] S4, comparing the liquid level difference with a liquid level difference threshold value through the control model, and comparing the gas concentration data with corresponding gas concentration threshold values;

[0009] S5, determining a working condition grade of the system based on the comparison results, wherein the working condition grade includes a safety priority working condition defined by the gas concentration data exceeding the corresponding gas concentration threshold value;

[0010] S6, generating a corresponding control instruction according to the determined working condition grade, and generating an instruction to close an adjusting valve on the communication pipe when the safety priority working condition is determined;

[0011] S7, executing the control instruction, implementing a control strategy by adjusting the opening degree of the adjusting valve on the communication pipe, and optimizing and adjusting the control model parameters based on system operation data.

[0012] By adopting the technical scheme, the gas purity parameter and the multi-threshold value coupling judgment mechanism of the liquid level difference are introduced, and then the hierarchical control instruction is triggered according to the exceeding standard level, thereby improving the problem that the traditional method mostly adopts single liquid level balance control, and due to the lack of active blocking mechanism for hydrogen diffusion in oxygen and oxygen diffusion in hydrogen, the risk of gas purity out of control is high.

[0013] Preferably, the synchronous acquisition comprises:

[0014] The synchronous sampling time sequence based on the system clock is set;

[0015] The sampling interval is dynamically adjusted according to the system load change;

[0016] The time sequence consistency of data of each sensor is ensured through the synchronous trigger function of the data acquisition card.

[0017] Preferably, the signal preprocessing comprises:

[0018] The oxygen concentration data in the hydrogen is filtered in real time by using a low-pass filter, and high-frequency measurement noise is eliminated;

[0019] The hydrogen concentration data in the oxygen is filtered in real time by using a low-pass filter;

[0020] The cutoff frequency of the filter is set so that it can effectively filter out random interference components in the sensor signal.

[0021] Preferably, the control model processing comprises:

[0022] A liquid level-gas purity coupling control model based on state space representation is constructed;

[0023] The preprocessed liquid level difference data is input into the state observer of the model;

[0024] The preprocessed gas concentration data is input into the mass balance equation of the model;

[0025] Preliminary control parameters are calculated through a model predictive control algorithm.

[0026] Preferably, the threshold comparison comprises:

[0027] The oxygen concentration data in the hydrogen is compared with a preset mild exceeding threshold value by using a first comparator;

[0028] The oxygen concentration data in the hydrogen is compared with a preset serious exceeding threshold value by using a second comparator;

[0029] The hydrogen concentration data in the oxygen is compared with a corresponding mild exceeding threshold value by using a third comparator;

[0030] The hydrogen concentration data in the oxygen is compared with a corresponding serious exceeding threshold value using a fourth comparator.

[0031] Preferably, the working condition level includes:

[0032] The liquid level difference is input to a first threshold comparator and compared with a first liquid level difference threshold value;

[0033] The liquid level difference is input to a second threshold comparator and compared with a second liquid level difference threshold value;

[0034] Based on the output results of the multiple comparators, the specific level of the liquid level abnormal working condition is determined by a logic judgment unit.

[0035] Preferably, the control instruction includes:

[0036] When it is determined that the safety priority working condition, a step response control instruction of the hydrogen side outlet pressure regulating valve is generated by a first control loop;

[0037] A proportional control instruction of the oxygen side outlet pressure regulating valve is generated by a second control loop;

[0038] An emergency shutdown instruction of the communication pipe regulating valve is generated by a safety interlock system.

[0039] Preferably, the control instruction further includes:

[0040] When the liquid level abnormal working condition is the first level, a regulating valve control instruction of a first predetermined opening range is generated by a PID controller;

[0041] When the liquid level abnormal working condition is the second level, a regulating valve control instruction of a second predetermined opening range is generated by a fuzzy control algorithm.

[0042] Preferably, the optimization adjustment of the control model parameters includes:

[0043] A historical database containing system dynamic response characteristics is established;

[0044] The parameters of the control model are iteratively optimized using a reinforcement learning algorithm;

[0045] The gain parameters of the controller are updated in real time by an online parameter identification technology.

[0046] Preferably, the S1-S7 further includes:

[0047] The dynamic response data of the system are recorded by a data acquisition system;

[0048] The execution time of each control instruction is marked using a time stamp;

[0049] Time-stamped data is stored into a ring buffer for model parameter optimization and system performance analysis.

[0050] The present application has the following advantages:

[0051] 1、The present application introduces a multi-threshold coupling judgment mechanism of gas purity parameters and liquid level difference, and triggers a hierarchical control instruction according to the exceeding level, thereby improving the single liquid level balance control of the traditional method, and lacking an active blocking mechanism for hydrogen diffusion in oxygen and oxygen diffusion in hydrogen, thereby causing a high risk of gas purity out of control.

[0052] 2、The present application synchronously collects liquid level data and gas concentration data and inputs them into a control model for processing, thereby realizing cooperative perception of liquid level state and gas purity state, thereby improving the single liquid level parameter monitoring of the traditional method, and unable to grasp the gas cross-diffusion situation in real time, thereby causing the problem of safety hidden danger early warning lag.

[0053] 3、The present application determines the system working condition level based on the comparison result and generates a corresponding control instruction, thereby matching the control response with the abnormal severity, thereby improving the fixed threshold control of the traditional method, and unable to distinguish fault levels and take indiscriminate response, thereby causing the problem of over-adjustment or insufficient response.

[0054] 4、The present application executes the control instruction and optimizes and adjusts the control model parameters based on system running data, thereby making the control strategy self-adaptively updated with the system running state, thereby improving the fixed parameter control of the traditional method, and unable to adapt to the change of system dynamic characteristics, thereby causing the problem of gradually declining control performance. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 A method flow chart of a hydrogen production system communication pipe control method is proposed for the present application;

[0056] Figure 2 A control model processing structure diagram of a hydrogen production system communication pipe control method is proposed for the present application;

[0057] Figure 3 A threshold comparison logic diagram of a hydrogen production system communication pipe control method is proposed for the present application;

[0058] Figure 4 A control instruction generation logic diagram of a hydrogen production system communication pipe control method is proposed for the present application;

[0059] Figure 5 A parameter optimization and data flow diagram of a hydrogen production system communication pipe control method is proposed for the present application. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1:

[0062] In a first embodiment of the present invention, the present invention provides a method for controlling the connecting pipe of a hydrogen production system, such as... Figures 1-5 As shown, it includes the following steps:

[0063] S1. Simultaneously collect liquid level data, oxygen concentration data in hydrogen, and hydrogen concentration data in oxygen from the hydrogen separator and oxygen separator.

[0064] Furthermore, synchronous data acquisition includes:

[0065] Configure synchronous sampling timing based on the system clock;

[0066] The sampling interval is dynamically adjusted according to changes in system load.

[0067] The synchronous triggering function of the data acquisition card ensures the timing consistency of data from each sensor.

[0068] Specifically, the synchronous acquisition process first establishes a synchronous sampling timing sequence based on the system clock. The system clock generates a reference clock signal, and each sensor data acquisition unit establishes a unified sampling time reference based on this clock signal. The sampling timing sequence satisfies the following relationship: ;in Indicates the first Each sampling time, Indicates the initial sampling time. Indicates the reference sampling period. This is the sampling sequence number. This timing design ensures that all sensor data has a consistent time label.

[0069] The sampling interval is dynamically adjusted based on changes in system load. The system load is measured by the current value passing through the electrolytic cell. Reflects the sampling period The relationship with the load is as follows: ;in As the reference sampling period, To adjust the coefficient, This is the system's maximum operating current. This refers to the real-time monitoring value of the electrolytic cell current. When the system load is close to its maximum value, a shorter sampling period is used to improve the control response speed; when the load is low, the sampling period is appropriately extended to reduce the system's computational burden.

[0070] The synchronization trigger function of the data acquisition card ensures the timing consistency of data from all sensors. The data acquisition card receives a synchronization trigger signal generated by the system clock, simultaneously initiating data conversion from the liquid level sensor, the oxygen concentration sensor in hydrogen, and the hydrogen concentration sensor in oxygen. The synchronization trigger signal is strictly synchronized with the sampling timing, ensuring that the time deviation of data acquisition from each sensor is less than a preset minimum time tolerance value.

[0071] The output signal of the liquid level sensor is converted from analog to digital to obtain a digital value of the liquid level, and the output signal of the gas concentration sensor is conditioned and converted from analog to digital to obtain a digital value of the gas concentration. All these data are simultaneously stored in the data buffer under the control of a synchronous trigger signal and marked with a uniform timestamp.

[0072] This synchronous acquisition mechanism ensures that the data processed by the subsequent control model is completely aligned in the time dimension, avoiding control deviations caused by asynchronous data acquisition times. A unified time base provides the necessary foundation for accurately calculating parameter change trends, and dynamically adjusting the sampling interval optimizes system resource utilization while maintaining control accuracy. Hardware synchronization triggering of the data acquisition card eliminates timing errors between sensor data at the physical level, establishing a reliable data foundation for subsequent precise control.

[0073] S2. Calculate the liquid level difference based on the collected liquid level data, and perform signal preprocessing on the gas concentration data;

[0074] Further signal preprocessing includes:

[0075] A low-pass filter is used to filter the oxygen concentration data in hydrogen in real time to eliminate high-frequency measurement noise.

[0076] A low-pass filter is used to perform real-time filtering of hydrogen concentration data in oxygen.

[0077] Set the cutoff frequency of the filter to effectively filter out random interference components in the sensor signal.

[0078] Specifically, the signal preprocessing stage is used to improve the quality of the acquired data and provide reliable input for subsequent control decisions. This stage includes two main parts: liquid level difference calculation and gas concentration data filtering.

[0079] The liquid level difference is calculated using a differential algorithm. Let the liquid level data of the hydrogen separator be... The oxygen separator liquid level data is as follows: Then the liquid level difference The calculation formula is: ;in This indicates the liquid level measurement value of the hydrogen separator. This indicates the measured liquid level in the oxygen separator. This represents the real-time liquid level difference between the two separators. This liquid level difference serves as the direct input parameter for subsequent liquid level balance control.

[0080] A second-order Butterworth low-pass filter is used for filtering the gas concentration data. The filter's transfer function is:

[0081] ;

[0082] in Describe the complex variable of the Laplace transform. This indicates the filter cutoff frequency. This filter is used for the oxygen concentration signal in hydrogen gas. and hydrogen concentration signal in oxygen Perform filtering processing separately.

[0083] The filter cutoff frequency is determined based on the sensor characteristics, and its calculation formula is as follows: ;in Indicates the cutoff frequency, based on the sensor sampling frequency. Confirmed, satisfied This relationship ensures that high-frequency noise is effectively filtered out while retaining useful components of the signal.

[0084] The discrete implementation of filtering in the time domain is as follows:

[0085] ;

[0086] in This represents the input signal at the current sampling time. This represents the output signal at the current sampling time. , This represents the input signal at the first two sampling times. , This represents the output signal at the first two sampling times. , , , , This represents the filter coefficients determined by the cutoff frequency.

[0087] The input signal flow is as follows: raw liquid level data directly enters the differential calculation unit, while raw gas concentration data first enters the filtering unit. The output signal flow is as follows: the liquid level difference calculation unit outputs Δh to the control model, and the filtering unit outputs the filtered signal. and The signal is fed into the control model. This processing method eliminates high-frequency random interference introduced during sensor measurement, ensuring the smoothness and reliability of gas purity parameters and creating conditions for accurate evaluation of system operating conditions. Real-time calculation of the liquid level difference provides a direct basis for liquid level balance control, and effective filtering of gas concentration data prevents control command oscillations caused by measurement noise.

[0088] S3. Input the liquid level difference and pre-treated gas concentration data into the control model for processing;

[0089] Furthermore, the control model processes the following:

[0090] Construct a liquid level-gas purity coupled control model based on state-space representation;

[0091] The preprocessed liquid level difference data is input into the model's state observer;

[0092] The preprocessed gas concentration data is input into the mass balance equation of the model;

[0093] Preliminary control parameters are calculated using a model predictive control algorithm.

[0094] Specifically, the liquid level-gas purity coupled control model establishes the system's dynamic characteristics using the state-space method. This model integrates liquid level balance and gas purity control within a unified framework, achieving multi-variable collaborative control.

[0095] The state-space representation consists of state equations and output equations. The state equations describe how the system state evolves over time: ;in This represents the system state vector, which includes liquid level difference, gas concentration, and internal state variables that cannot be directly measured. This represents the control input vector, which contains the valve opening command. It represents the system state matrix and describes the coupling relationships between the state variables. This represents the control input matrix, which describes the degree of influence of the control quantity on the state.

[0096] The output equation defines the mathematical expression of the system observations: ;in This represents the system output vector, which contains measurable liquid level difference and gas concentration data. This represents the output matrix, establishing the mapping relationship between state variables and observations.

[0097] The state observer receives the preprocessed liquid level difference data Δh and reconstructs the complete system state using the following algorithm: ;in This represents the state estimate. This represents the observer gain matrix. The observer continuously corrects the state estimate using the deviation between the actual output and the estimated output, ensuring accurate tracking of the system's true state even in sensor measurement noise environments.

[0098] The mass balance equation is specifically designed for gas concentration data, describing the mass transfer process between oxygen in hydrogen and hydrogen in oxygen: ;in This represents a gas concentration vector, containing OTH and HTO concentration values. Indicates the mass transfer coefficient. This represents the gas cross-diffusion flux caused by the exchange of alkali solution in the connecting tube.

[0099] Model predictive control algorithms predict system behavior over a future period based on current state estimates, and calculate control parameters by optimizing an objective function. The objective function comprehensively considers both liquid level deviation and gas purity deviation. ;in Indicates the length of the prediction time domain. This indicates the length of the control time domain. Indicates the future The system output predicted value at time 10:00. This indicates that a reference value will be output. and These are the weight matrices for the output error and the control increment, respectively. This optimization problem is solved once per control cycle, outputting a preliminary sequence of control parameters; only the first control parameter is actually executed.

[0100] Input data includes pre-treated liquid level difference Filtered and Concentration value. The output data is the initial opening command of the control valve. The state observer uses the liquid level difference data to correct the state estimate, the mass balance equation handles the dynamic characteristics of gas purity, and the model predictive control algorithm coordinates multiple control objectives to generate the optimal control command. This approach establishes a quantitative coupling relationship between liquid level and gas purity, providing accurate control parameters for subsequent staged control.

[0101] S4. By controlling the model, compare the liquid level difference with the liquid level difference threshold, and compare the gas concentration data with the corresponding gas concentration threshold.

[0102] Furthermore, the threshold comparison includes:

[0103] The first comparator is used to compare the oxygen concentration data in hydrogen with a preset slightly excessive threshold.

[0104] The second comparator is used to compare the oxygen concentration data in hydrogen with a preset severe exceedance threshold.

[0105] The third comparator is used to compare the hydrogen concentration data in oxygen with the corresponding slightly excessive threshold.

[0106] The fourth comparator is used to compare the hydrogen concentration data in oxygen with the corresponding severe exceedance threshold.

[0107] Specifically, the threshold comparison process is performed by a dedicated comparator circuit or a software comparison module. For the oxygen concentration data in hydrogen, a threshold for slight exceedance is set. The volume fraction was 0.2%, which is significantly above the acceptable threshold. The volume fraction is 0.4%. The measured oxygen concentration in hydrogen gas is... Input to the first comparator and Compare, and simultaneously input the second comparator with Compare. This represents the measured oxygen concentration in the hydrogen gas collected in the current cycle. This indicates that the oxygen concentration in hydrogen gas slightly exceeds the threshold. This indicates that the oxygen concentration in the hydrogen gas has seriously exceeded the standard threshold.

[0108] For the hydrogen concentration data in oxygen, a threshold for slight exceedance is set. The volume fraction was 2%, which is significantly higher than the threshold. The volume fraction was 4%. The measured hydrogen concentration in oxygen was... Input third comparator and Compare, and simultaneously input the fourth comparator with Compare. This represents the measured value of hydrogen concentration in the oxygen collected in the current cycle. This indicates that the hydrogen concentration in oxygen slightly exceeds the threshold. This indicates that the hydrogen concentration in oxygen has seriously exceeded the threshold.

[0109] Each comparator outputs a Boolean value, and the comparison logic is as follows:

[0110] First comparator output :when Output true if true, otherwise output false.

[0111] Second comparator output :when Output true if true, otherwise output false.

[0112] Third comparator output :when Output true if true, otherwise output false.

[0113] Fourth comparator output :when Output true if true, otherwise output false.

[0114] The same mechanism was used for comparing liquid level differences. Measured liquid level difference values. Input the fifth comparator and the mild anomaly threshold The input to the sixth comparator is compared with the severe anomaly threshold. Compare. Set to 50mm. Set to 100mm. The comparator output is:

[0115] Fifth comparator output :when The output is true.

[0116] Sixth comparator output :when The output is true.

[0117] All comparators are refreshed synchronously in each control cycle, and the output results are transmitted to the logic judgment unit for determining the operating condition level. This comparison mechanism provides discretized input conditions for the hierarchical control strategy, converting continuous sensor signals into explicit logical states.

[0118] S5. Determine the system's operating condition level based on the comparison results. The operating condition level includes the safety priority operating condition defined by the gas concentration data exceeding the corresponding gas concentration threshold.

[0119] Furthermore, the operating condition levels include:

[0120] The liquid level difference is input into the first threshold comparator and compared with the first liquid level difference threshold.

[0121] The liquid level difference is input into the second threshold comparator and compared with the second liquid level difference threshold.

[0122] Based on the output of multiple comparators, the specific level of abnormal liquid level conditions is determined by a logic judgment unit.

[0123] Specifically, the condition level determination is performed by a logic judgment unit. The input to the logic judgment unit is the Boolean output of the six comparators in S4: the first comparator output... Second comparator output Third comparator output Fourth comparator output Fifth comparator output The output of the sixth comparator . This indicates whether the oxygen concentration in hydrogen exceeds the slightly excessive threshold. This indicates whether the oxygen concentration in hydrogen exceeds the severely excessive threshold. This indicates whether the hydrogen concentration in the oxygen exceeds the slightly excessive threshold. This indicates whether the hydrogen concentration in oxygen exceeds the severely excessive threshold. This indicates whether the absolute value of the liquid level difference exceeds the threshold for minor anomalies. This indicates whether the absolute value of the liquid level difference exceeds the severe anomaly threshold.

[0124] The logic judgment unit first calculates the gas purity severely exceeding the standard indicator. and signs of severe liquid level abnormality : ;in A value of true indicates that the gas concentration has seriously exceeded the standard. A value of true indicates a severe anomaly in the liquid level difference. (Symbol) Represents a logical OR operation.

[0125] Next, calculate the gas purity slightly exceeding the standard. and signs of slight abnormal liquid level :

[0126] ;

[0127] in A value of true indicates that the gas concentration has slightly exceeded the standard. A value of true indicates a slight abnormality in the liquid level difference.

[0128] Operating condition level Determined by the following logical judgment rules:

[0129] like or ,but This indicates a severely abnormal operating condition;

[0130] Otherwise if or ,but 1 indicates a slightly abnormal operating condition;

[0131] otherwise This indicates normal operating conditions.

[0132] This is an integer variable, taking values ​​of 0, 1, or 2, corresponding to normal operating conditions, slightly abnormal operating conditions, and severely abnormal operating conditions, respectively. When The system then enters a safety-priority operating mode, prioritizing the handling of gas purity risks.

[0133] The logic judgment unit synchronously reads all comparator outputs in each control cycle, performs the above logic operation, and then outputs the result. The judgment process ensures that the highest priority response is triggered regardless of the liquid level when the gas concentration is severely exceeded, thus prioritizing safety. The output operating condition level is transmitted to the control command generation module as the basis for decision-making.

[0134] S6. Generate corresponding control commands based on the determined operating condition level. When the operating condition is determined to be a safety priority condition, generate a command to close the regulating valve on the connecting pipe.

[0135] Furthermore, the control commands include:

[0136] When the safety priority condition is determined, a step response control command for the hydrogen side outlet pressure regulating valve is generated through the first control loop;

[0137] The proportional control command for the oxygen-side outlet pressure regulating valve is generated through the second control loop.

[0138] An emergency shut-off command for the regulating valve in the connecting pipe is generated through the safety interlock system.

[0139] The control commands also include:

[0140] When the abnormal liquid level condition is at level one, the PID controller generates a control command for the regulating valve within the first predetermined opening range.

[0141] When the abnormal liquid level condition is at level two, a control command for the regulating valve with a second predetermined opening range is generated through a fuzzy control algorithm.

[0142] Specifically, corresponding control commands are generated based on the determined operating condition level. The input to the control command generation module is the operating condition level. Liquid level difference Oxygen concentration in hydrogen gas Hydrogen concentration in oxygen The output is the opening command for the regulating valve in the connecting pipe. Hydrogen-side outlet pressure regulating valve opening command Oxygen side outlet pressure regulating valve opening command .

[0143] When it is determined to be a safety priority operating condition, i.e. At this time, the control command generation module executes the safety interlock logic. First, it generates an emergency closing command for the connecting pipe regulating valve: . This indicates the opening degree command of the regulating valve in the connecting pipe, in percentage. This indicates that the valve is completely closed.

[0144] Meanwhile, the hydrogen-side outlet pressure regulating valve employs step response control. Its opening command... Calculated by the following formula: ;in This indicates the hydrogen-side outlet valve opening command for the current control cycle. This indicates the opening command from the previous control cycle. This is the preset step change amount. It is a sign function, the value of which is determined by the sign of the liquid level difference: if but ,like but This command causes the hydrogen-side outlet valve to actuate rapidly to compensate for the liquid level regulation function after the connecting pipe is closed.

[0145] The oxygen-side outlet pressure regulating valve uses proportional control, and its opening command... Proportional to the liquid level difference:

[0146] ;

[0147] in This is the proportional control coefficient for the oxygen-side outlet valve. This indicates the oxygen-side outlet valve opening command for the current control cycle.

[0148] When the operating condition level is abnormal liquid level, that is... or However, if the problem is caused by a severe abnormal liquid level and the gas concentration is not severely exceeded, then a regulating control should be adopted for the regulating valve of the connecting pipe.

[0149] If the abnormal liquid level is classified as Level 1 The PID controller generates commands for the regulating valve. The PID control algorithm formula is: ;in This is the opening command for the regulating valve in the connecting pipe during the current control cycle. For level difference deviation, Set the liquid level difference reference value to 0. This is the proportionality coefficient. The integral coefficient is... The differential coefficients are... For control cycle. This command causes the valve opening to be smoothly adjusted within a first predetermined opening range of 0% to 50%.

[0150] If the abnormal liquid level is classified as Level 2 And by If triggered, a fuzzy control algorithm is employed. The input to the fuzzy controller is the liquid level difference. and its rate of change The output is the change in the opening of the regulating valve in the connecting pipe. The opening instruction is calculated as follows: The fuzzy rule base contains the following rules: If For the upright and For the sake of righteousness, then This algorithm enables the valve opening to be rapidly adjusted within a second predetermined range of 50% to 70% to correct severe level deviations with maximum flow capacity.

[0151] All control commands , , Ultimately, the output is limited to the physical opening range of each valve and then sent to the corresponding actuator.

[0152] S7. Execute control commands, implement control strategies by adjusting the opening of the regulating valve on the connecting pipe, and optimize and adjust the control model parameters based on system operation data;

[0153] Furthermore, optimizing and adjusting the control model parameters includes:

[0154] Establish a historical database containing the dynamic response characteristics of the system;

[0155] The parameters of the control model are iteratively optimized using reinforcement learning algorithms;

[0156] The controller's gain parameters are updated in real time using online parameter identification technology.

[0157] Specifically, the control command execution module receives control command vectors from S6. . This represents the control command vector for the current control cycle. This indicates the opening command of the regulating valve in the connecting pipe. This indicates the opening command of the hydrogen-side outlet pressure regulating valve. This indicates the opening command for the oxygen-side outlet pressure regulating valve. The execution module converts the digital command into a 4-20mA analog signal to drive the electrical positioners of each regulating valve, ensuring the valve opening meets the command requirements.

[0158] The system operation data recording unit synchronously collects system response data after command execution, including liquid level difference. Oxygen concentration in hydrogen gas Hydrogen concentration in oxygen The actual opening degree of each valve is recorded. A timestamp is appended to each data point in the recording unit. This generates time-stamped data packets. .

[0159] A historical database is established to store the aforementioned data packets. The database uses a circular buffer structure with a buffer size of [size missing]. When the data volume exceeds It automatically overwrites the oldest data. This indicates the maximum number of data packets that the circular buffer can store.

[0160] Reinforcement learning algorithms are used to analyze the parameters of the PID controller. , , Perform iterative optimization. Define the value function. : ;

[0161] in This represents the cumulative reward value. Indicates the optimization cycle length. This is the discount factor, with a value ranging from 0 to 1. Indicates the first Instant rewards for each step. Calculated jointly by liquid level deviation and gas concentration deviation:

[0162] ;

[0163] These are the weighting coefficients. The optimization objective is to adjust the PID parameters to achieve... Maximize. In each optimization cycle of the algorithm... Update the PID parameter matrix by sampling data from the historical database. .

[0164] The controller gain parameters are updated in real time using online parameter identification technology. The system model parameters are identified using the recursive least squares method. A parameter vector is defined. Its recursive update formula is:

[0165] ;

[0166] ;

[0167] ;

[0168] in express The parameter vector estimate at time t. express The system output at any given time, express The vector of observed data at time t, Represents the gain matrix. Represents the covariance matrix. Forgetting factor. Identified parameters. Used to adjust the controller gain in real time to ensure that control performance is maintained under different loads.

[0169] The optimized parameters are written into the control model via an online update mechanism, completing the parameter adjustment closed loop. The entire process is executed cyclically in each control cycle, achieving continuous self-optimization of the control strategy.

[0170] S1-S7 also include:

[0171] The system's dynamic response data is recorded through a data acquisition system;

[0172] Use timestamps to mark the execution time of each control command;

[0173] Time-stamped data is stored in a circular buffer for model parameter optimization and system performance analysis.

[0174] Specifically, the data acquisition system records the system's dynamic response data in each control cycle. The recorded dataset includes control command vectors, system state vectors, and timestamps. The control command vector is denoted as... ,in This is a command to adjust the opening of the regulating valve in the connecting pipe. This is the opening command for the hydrogen-side outlet pressure throttle valve. This is the opening command for the oxygen-side outlet pressure regulating valve. The system state vector is denoted as... ,in This is the current liquid level difference. This is the current oxygen concentration in hydrogen gas. It represents the current hydrogen concentration in oxygen.

[0175] Attach a timestamp to each data packet using the system clock. The timestamp is calculated using the following formula:

[0176] ;

[0177] in It is the first The absolute time of each control cycle It is the base time for system initialization. It is a control cycle. It is a control cycle count. The timestamp marks the execution time of each control command and the corresponding system status acquisition time.

[0178] Timestamped data packets are stored in a circular buffer. The circular buffer is a fixed-size first-in-first-out queue, and its storage index... Update according to the following rules: ;in It is the index of the data packet's storage location in the circular buffer. It controls the cycle count. This is the total capacity of the circular buffer. When... Exceed When new data is added, the oldest data will be overwritten. Each storage location stores one complete data record. .

[0179] The stored data is used for model parameter optimization and system performance analysis. The parameter optimization module periodically reads historical phase sequences from the buffer. ,in This refers to the analysis window size. The performance analysis module calculates key metrics such as the root mean square value of the liquid level difference. : ; In the past The degree of fluctuation in liquid level difference within each control cycle. These quantified performance indicators, together with the original data, constitute the basic input of the optimization algorithm.

[0180] The data logging process provides the system with a complete historical operational archive. Timestamps ensure that all data has an accurate time-series relationship, supporting retrospective analysis of control effects. The circular buffer structure ensures that the system can run for a long time without exhausting storage space, always retaining key operational data from the most recent period. The stored data directly serves the online learning and parameter tuning of the control model, enabling the system to self-improve based on historical operational experience.

[0181] Example 2:

[0182] Under the power fluctuation operation of a renewable energy water electrolysis hydrogen production station, the electrolyzer load needs to be rapidly adjusted according to the output of wind and solar power generation, leading to dynamic imbalance of liquid level in the hydrogen and oxygen separator and increased turbulence in the gas-liquid two-phase flow. At this time, traditional control methods, relying solely on the liquid level difference as a single parameter to adjust the connecting pipe, cannot simultaneously monitor and suppress the diffusion of hydrogen-oxygen and oxygen-hydrogen impurity gases through the connecting pipe with the alkali exchange. Furthermore, they lack a mechanism for graded response based on the abnormal amplitude of the liquid level difference and the severity of gas exceedances. This results in a comprehensive technical problem under dynamic operating conditions: uncontrolled gas purity, delayed safety risks, and a mismatch between control response and fault severity. To solve these problems, this invention provides a connecting pipe control method for a hydrogen production system, the structure of which is as follows: Figure 1 As shown. The specific implementation process of this method is as follows:

[0183] By executing steps S1 to S7, intelligent and coordinated control of the connecting pipe between the hydrogen and oxygen separators in the water electrolysis hydrogen production system was achieved. The core function of this method is to break through the traditional single control dimension that only aims at liquid level balance, and to construct a comprehensive control system that integrates real-time monitoring, multi-parameter coupling analysis, hierarchical decision-making, safety interlocking, and self-optimization.

[0184] Synchronous data acquisition via S1 provides the control system with time-consistent basic data on liquid level and gas purity. The signal preprocessing stage of S2, particularly the filtering of oxygen concentration data in hydrogen and hydrogen concentration data in oxygen, effectively eliminates measurement noise, providing reliable data quality for subsequent accurate judgment of operating conditions. S3 inputs the processed liquid level difference and gas concentration data into the liquid level-gas purity coupled control model for further processing. This model is the core of the coordinated control, establishing a dynamic correlation between liquid level status and gas purity.

[0185] S4 uses a multi-threshold comparator to compare the liquid level difference and gas concentration parameters in parallel, converting continuous physical quantities into discrete logical states. S5 then determines the specific operating condition level of the system based on these comparison results and according to preset rules. This hierarchical mechanism, particularly the definition of a "safety-priority operating condition," enables the system to clearly distinguish between different states such as normal, slightly abnormal, and severely abnormal.

[0186] S6 generates corresponding control commands based on the operating condition level determined by S5. When the operating condition is determined to be safety-priority, the system will prioritize handling gas purity risks, generating a command to close the regulating valve of the connecting pipe, fundamentally cutting off the path of gas diffusion through alkaline exchange. Simultaneously, it temporarily compensates for the liquid level balance function by adjusting the outlet pressure regulating valves on the hydrogen and oxygen sides. For different levels of liquid level anomalies, different algorithms such as PID control or fuzzy control are used to generate corresponding regulating valve opening commands, achieving precise matching between control response and fault severity.

[0187] S7 is responsible for executing control commands and continuously optimizing the system. It not only translates control commands into the actual actions of the regulating valve, but also records the system's dynamic response data and uses techniques such as reinforcement learning and online parameter identification to iteratively optimize the control model parameters, enabling the system to adapt to changes in equipment characteristics during long-term operation and continuously improve control performance.

[0188] Through the closed-loop execution of the above steps, the dual objectives of dynamic liquid level balance and gas purity assurance are achieved. Under various operating conditions, especially when dynamic load changes occur, the risk of gas cross-diffusion can be detected and suppressed in advance, significantly improving the safety, stability and adaptability of the hydrogen production system.

[0189] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for controlling the connecting pipe of a hydrogen production system, characterized in that, Includes the following steps: S1. Simultaneously collect liquid level data, oxygen concentration data in hydrogen, and hydrogen concentration data in oxygen from the hydrogen separator and oxygen separator. S2. Calculate the liquid level difference based on the collected liquid level data, and perform signal preprocessing on the gas concentration data; S3. Input the liquid level difference and pre-treated gas concentration data into the control model for processing; S4. By controlling the model, compare the liquid level difference with the liquid level difference threshold, and compare the gas concentration data with the corresponding gas concentration threshold. S5. Determine the operating condition level of the system based on the comparison results. The operating condition level includes a safety priority operating condition defined by the gas concentration data exceeding the corresponding gas concentration threshold. S6. Generate corresponding control commands based on the determined operating condition level. When the operating condition is determined to be a safety priority condition, generate a command to close the regulating valve on the connecting pipe. S7. Execute control commands, implement control strategies by adjusting the opening of the regulating valve on the connecting pipe, and optimize and adjust the control model parameters based on system operation data.

2. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The synchronous acquisition includes: Configure synchronous sampling timing based on the system clock; The sampling interval is dynamically adjusted according to changes in system load. The synchronous triggering function of the data acquisition card ensures the timing consistency of data from each sensor.

3. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The signal preprocessing includes: A low-pass filter is used to filter the oxygen concentration data in the hydrogen in real time to eliminate high-frequency measurement noise. The hydrogen concentration data in the oxygen is filtered in real time using a low-pass filter. Set the cutoff frequency of the filter to effectively filter out random interference components in the sensor signal.

4. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The control model performs the following processing: Construct a liquid level-gas purity coupled control model based on state-space representation; The preprocessed liquid level difference data is input into the state observer of the model; The preprocessed gas concentration data is input into the mass balance equation of the model. Preliminary control parameters are calculated using a model predictive control algorithm.

5. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The threshold comparison includes: The oxygen concentration data in the hydrogen is compared with a preset slight exceedance threshold using a first comparator. The oxygen concentration data in the hydrogen is compared with a preset severe exceedance threshold using a second comparator. The hydrogen concentration data in the oxygen is compared with the corresponding slightly excessive threshold using a third comparator. The fourth comparator is used to compare the hydrogen concentration data in the oxygen with the corresponding severe exceedance threshold.

6. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The operating condition levels include: The liquid level difference is input into the first threshold comparator and compared with the first liquid level difference threshold. The liquid level difference is input into the second threshold comparator and compared with the second liquid level difference threshold. Based on the output of multiple comparators, the specific level of abnormal liquid level conditions is determined by a logic judgment unit.

7. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The control commands include: When the safety priority condition is determined, a step response control command for the hydrogen side outlet pressure regulating valve is generated through the first control loop; The proportional control command for the oxygen-side outlet pressure regulating valve is generated through the second control loop. An emergency shut-off command for the regulating valve in the connecting pipe is generated through the safety interlock system.

8. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The control commands also include: When the abnormal liquid level condition is at level one, the PID controller generates a control command for the regulating valve within the first predetermined opening range. When the abnormal liquid level condition is at level two, a control command for the regulating valve with a second predetermined opening range is generated through a fuzzy control algorithm.

9. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, The optimization and adjustment of the control model parameters includes: Establish a historical database containing the dynamic response characteristics of the system; The parameters of the control model are iteratively optimized using reinforcement learning algorithms; The controller's gain parameters are updated in real time using online parameter identification technology.

10. The method for controlling the connecting pipe of a hydrogen production system according to claim 1, characterized in that, S1-S7 also includes: The system's dynamic response data is recorded through a data acquisition system; Use timestamps to mark the execution time of each control command; Time-stamped data is stored in a circular buffer for model parameter optimization and system performance analysis.

Images