Alkaline electrolysis system energy consumption-temperature-purity collaborative prediction control method

Abstract

The application relates to the technical field of hydrogen production control by electrolysis of water, and discloses an alkali electrolysis system energy consumption-temperature-purity collaborative prediction control method, which acquires operation data and control targets of an alkali electrolysis system, generates a plurality of candidate operation conditions according to the operation data, the control targets and equipment operation ranges; predicts electrolytic cell voltage, hydrogen production, oxygen production and unit hydrogen production energy consumption corresponding to the candidate operation conditions through an electrochemical model; predicts electrolytic cell temperature state through a thermal model; predicts hydrogen purity and oxygen purity through a gas purity model; performs coupling evaluation on the unit hydrogen production energy consumption, the electrolytic cell temperature state, the hydrogen purity and the oxygen purity to form a collaborative prediction result, determines a target operation condition from the candidate operation conditions meeting constraint conditions, and outputs corresponding control parameters; and corrects model parameters according to actual operation data. The alkali electrolysis system operation condition selection rationality can be improved.

Description

Technical Field

[0001] This invention relates to the field of hydrogen production control technology through water electrolysis, specifically a method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system. Background Technology

[0002] Alkaline electrolysis systems are currently a relatively mature method for producing hydrogen through water electrolysis. They typically include an alkaline electrolyzer, DC power supply, alkali circulation system, water replenishment system, gas-liquid separation system, cooling system, pressure regulation system, gas purity detection device, and control system. During operation, current, voltage, temperature, pressure, anode-cathode pressure difference, alkali circulation flow rate, cooling capacity, and water replenishment status all collectively affect the electrolyzer's energy consumption, hydrogen production, temperature safety, and gas purity. With the development of renewable energy-based hydrogen production and dynamic load hydrogen production scenarios, alkaline electrolysis systems not only need to meet target hydrogen production volumes but also need to consider unit hydrogen production energy consumption, temperature constraints, and gas purity requirements.

[0003] Existing alkaline electrolysis systems often employ control methods based on fixed current, fixed power, empirical table lookup, or feedback from a single operating parameter to determine operating conditions. These methods typically focus only on single or a few parameters such as current, voltage, hydrogen production, or temperature, making it difficult to comprehensively reflect the coupling relationship between electrochemical processes, thermal processes, and gas purity changes. When system load, ambient temperature, alkaline solution circulation status, pressure status, or hydrogen production demand changes, the original operating conditions may no longer be suitable for the current system state, easily leading to problems such as increased energy consumption per unit of hydrogen production, lag in temperature control, insufficient gas purity margin, or inappropriate operating condition selection. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for coordinated predictive control of energy consumption, temperature, and purity in alkaline electrolysis systems. This method solves the problem that existing alkaline electrolysis systems cannot uniformly reflect the coupling relationship between energy consumption, temperature, and gas purity, leading to insufficient rationality in the selection of operating conditions.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system, comprising:

[0006] S1. Obtain the operating data and control targets of the alkaline electrolysis system, wherein the control targets include target hydrogen production, target hydrogen purity, and target oxygen purity;

[0007] S2. Generate multiple candidate operating conditions based on the operating data, the control target, and the equipment operating range. Each candidate operating condition includes a current setting value, a pressure setting value, an alkaline solution circulation flow setting value, a cooling water flow setting value, and a makeup water setting value.

[0008] S3. Input the candidate operating conditions into the electrochemical model to predict the corresponding electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production.

[0009] S4. Input the candidate operating conditions and the prediction results of the electrochemical model into the thermal model to predict the corresponding electrolytic cell temperature state;

[0010] S5. Input the candidate operating conditions, the prediction results of the electrochemical model and the prediction results of the thermal model into the gas purity model to predict the corresponding hydrogen purity and oxygen purity.

[0011] S6. The unit hydrogen production energy consumption, the temperature state of the electrolyzer, the hydrogen purity and the oxygen purity are coupled and evaluated to form a synergistic prediction result;

[0012] S7. Select candidate operating conditions that meet the constraints from the collaborative prediction results according to the control target and the equipment operating range, and determine the target operating condition based on the comprehensive target value;

[0013] S8. Convert the target operating condition into control parameters and output them to the DC power supply, alkali circulation system, cooling system, pressure regulation system and water replenishment system;

[0014] S9. Obtain the actual operating data after the target operating condition is executed, and correct the electrochemical model, the thermal model and the gas purity model according to the error between the prediction result and the actual operating data.

[0015] Further, in step S1, the step of obtaining the operational data includes:

[0016] Time alignment is performed on the data collected from the alkaline electrolytic cell, DC power supply, alkaline solution circulation system, cooling system, pressure regulation system, water replenishment system, and gas purity detection device.

[0017] Perform outlier removal and smoothing on the time-aligned collected data;

[0018] A state dataset is generated based on the processed collected data to characterize the current operating state.

[0019] Furthermore, in step S2, the step of generating multiple candidate operating conditions includes:

[0020] Based on the current operating data, adjust the current setting value, the pressure setting value, the alkali solution circulation flow rate setting value, the cooling water flow rate setting value, and the water replenishment setting value respectively within the operating range of the equipment;

[0021] The adjusted settings are combined to form candidate combinations;

[0022] Candidate operating conditions are obtained by eliminating candidate combinations that exceed the acceleration / deceleration limit, power limit, and flow limit.

[0023] Further, in step S3, the electrochemical model calculates the electrochemical polarization loss, resistance loss, and Faraday efficiency based on the current density, pressure, electrolyzer temperature, and alkaline circulation state, and predicts the electrolyzer voltage, hydrogen production, oxygen production, and unit hydrogen production energy consumption based on the electrochemical polarization loss, resistance loss, and Faraday efficiency.

[0024] Further, in step S4, the thermal model predicts the temperature state of the electrolytic cell based on the input power, reaction heat generation, environmental heat dissipation, alkaline solution circulation heat exchange, and cooling heat exchange. The temperature state of the electrolytic cell includes the temperature change trend and the predicted temperature value.

[0025] Further, in step S5, the prediction step of the gas purity model includes: extracting the anode-cathode pressure difference from the operating data; inputting the hydrogen production, oxygen production, electrolyzer temperature state, pressure setpoint and alkali circulation flow rate setpoint in the candidate operating conditions, the anode-cathode pressure difference, and gas-liquid separation efficiency into the gas purity model to predict the risk of gas cross-permeation and the risk of gas-liquid entrainment; and predicting the hydrogen purity and the oxygen purity based on the risk of gas cross-permeation and the risk of gas-liquid entrainment.

[0026] Further, in step S6, the step of forming the collaborative prediction result includes:

[0027] The electrolyzer temperature state predicted by the thermal model is used as the temperature input of the gas purity model, and as the temperature input of the electrochemical model during the next control cycle or the second iteration correction in the same control cycle.

[0028] The hydrogen and oxygen production predicted by the electrochemical model are used as the gas production inputs of the gas purity model.

[0029] The unit hydrogen production energy consumption, the electrolyzer temperature status, the hydrogen purity, and the oxygen purity are correlated according to the same candidate operating condition to form the collaborative prediction result.

[0030] Further, in step S7, the constraints include:

[0031] The predicted hydrogen production is not less than the target hydrogen production.

[0032] The predicted hydrogen purity is not lower than the target hydrogen purity;

[0033] The predicted oxygen purity is not lower than the target oxygen purity;

[0034] The predicted temperature value will not exceed the upper temperature limit;

[0035] The predicted pressure is within the allowable pressure range;

[0036] The current, voltage, alkaline solution circulation flow rate, cooling water flow rate, and makeup water volume are all within their respective allowable ranges.

[0037] Further, in step S9, the step of correcting the electrochemical model, the thermal model, and the gas purity model includes:

[0038] The actual voltage, actual temperature, actual hydrogen production, actual hydrogen purity, and actual oxygen purity are compared with the corresponding prediction results to obtain the prediction error.

[0039] When the prediction error exceeds a set threshold, the model parameters are corrected.

[0040] The model parameters include polarization parameters, ohmic impedance parameters, and Faraday efficiency parameters in the electrochemical model; equivalent heat capacity, heat dissipation coefficient, and cooling heat transfer coefficient in the thermal model; and gas mixing correction coefficient and gas-liquid separation efficiency parameters in the gas purity model.

[0041] This invention also provides a coordinated predictive control system for energy consumption, temperature, and purity in an alkaline electrolysis system, used to execute the aforementioned coordinated predictive control method for energy consumption, temperature, and purity in an alkaline electrolysis system. The system includes:

[0042] The data acquisition module is used to acquire the operating data and control objectives of the alkaline electrolysis system;

[0043] The candidate operating condition generation module is used to generate multiple candidate operating conditions based on the operating data, the control target, and the equipment operating range.

[0044] The electrochemical prediction module is used to predict the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production using an electrochemical model.

[0045] The thermal prediction module is used to predict the temperature state of the electrolyzer using a thermal model.

[0046] The purity prediction module is used to predict the purity of hydrogen and oxygen using a gas purity model.

[0047] The coupled evaluation module is used to perform coupled evaluation of the unit hydrogen production energy consumption, the temperature state of the electrolyzer, the hydrogen purity and the oxygen purity to form a synergistic prediction result;

[0048] The optimization control module is used to filter candidate operating conditions that meet the constraints from the collaborative prediction results, determine the target operating condition, and output control parameters.

[0049] The model correction module is used to correct the electrochemical model, the thermal model, and the gas purity model based on actual operating data.

[0050] This invention provides a method for the coordinated predictive control of energy consumption, temperature, and purity in an alkaline electrolysis system. It offers the following advantages:

[0051] 1. This invention uses electrochemical models, thermal models, and gas purity models to predict the unit hydrogen production energy consumption, electrolyzer temperature state, hydrogen purity, and oxygen purity under candidate operating conditions, respectively. The prediction results are coupled and evaluated to form a synergistic prediction result, which can uniformly reflect the coupling relationship between the electrochemical process, thermal process, and gas purity change process in the alkaline electrolysis system, thereby improving the rationality of operating condition selection.

[0052] 2. When determining the target operating condition, this invention uses the target hydrogen production, target hydrogen purity, target oxygen purity, upper temperature limit, pressure range, and equipment operating range as constraints. The target operating condition is determined based on the comprehensive target value among the candidate operating conditions that meet the constraints. This can avoid the problem of gas purity decreasing or temperature exceeding limits due to simply pursuing low energy consumption.

[0053] 3. This invention predicts the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production using an electrochemical model, enabling the controller to estimate the energy consumption results of different candidate operating conditions before executing control, thereby reducing the energy consumption per unit of hydrogen production while meeting the requirements for hydrogen production and gas purity.

[0054] 4. This invention predicts the temperature state of the electrolytic cell under candidate operating conditions through a thermal model, and incorporates the predicted temperature value into the constraint screening and comprehensive target value calculation. This can identify the risk of temperature exceeding the limit in advance and reduce the equipment safety risk caused by the lag in electrolytic cell temperature control.

[0055] 5. This invention predicts the risks of gas cross-permeation, gas-liquid entrainment, hydrogen purity, and oxygen purity through a gas purity model, and uses hydrogen purity and oxygen purity as constraints. This can reduce the risks of substandard gas purity and hydrogen-oxygen mixing, and improve the safety of the hydrogen production system and the quality of the product gas.

[0056] 6. This invention can generate multiple candidate operating conditions based on changes in hydrogen production demand or new energy input power, and dynamically determine the target operating condition based on the collaborative prediction results, enabling the alkaline electrolysis system to adapt to variable load operating scenarios and improve the system's operational flexibility and economy.

[0057] 7. This invention corrects the errors between actual operating data and predicted results in electrochemical models, thermal models, and gas purity models, thereby reducing the deviation between model parameters and long-term operating conditions, and making subsequent operating condition optimization results closer to the actual system state. Attached Figure Description

[0058] Figure 1 This is a structural block diagram of the alkaline electrolysis system of the present invention;

[0059] Figure 2 This is a flowchart of the energy consumption-temperature-purity coordinated prediction and control method for alkaline electrolysis systems according to the present invention;

[0060] Figure 3 This is a diagram showing the coupling prediction relationship between the electrochemical model, thermal model, and gas purity model of this invention;

[0061] Figure 4 This is a schematic diagram of the closed-loop control and model correction of the present invention. Detailed Implementation

[0062] The technical solution of the present invention will now be clearly and completely described 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.

[0063] This invention provides a method for the coordinated predictive control of energy consumption, temperature, and purity in an alkaline electrolysis system, applicable to the prediction, optimization, and closed-loop control of operating conditions in alkaline electrolysis hydrogen production systems. The alkaline electrolysis system can be a single alkaline electrolyzer system, or a hydrogen production system formed by multiple alkaline electrolyzers operating in parallel or in groups. The alkaline electrolysis system includes at least an alkaline electrolyzer, a DC power supply, an alkaline solution circulation system, a cooling system, a pressure regulation system, a water replenishment system, a gas-liquid separation system, a gas purity detection device, and a controller.

[0064] The system includes: a DC power supply to provide electrical energy to the alkaline electrolyzer; an alkaline solution circulation system to drive the alkaline solution to circulate between the alkaline electrolyzer and the gas-liquid separation system; a cooling system to remove heat generated during the operation of the alkaline electrolyzer; a pressure regulation system to regulate the pressure on the hydrogen and oxygen sides; a water replenishment system to replenish the water required for the reaction in the alkaline electrolysis system; a gas-liquid separation system to separate the gas-liquid mixtures on the hydrogen and oxygen sides respectively; a gas purity detection device to detect the purity of hydrogen and oxygen; and a controller to execute the energy consumption-temperature-purity coordinated predictive control method for the alkaline electrolysis system of this invention. Figure 1 As shown.

[0065] In a preferred embodiment, the controller includes a data acquisition module, a candidate operating condition generation module, an electrochemical prediction module, a thermal prediction module, a purity prediction module, a coupling evaluation module, an optimization control module, and a model correction module. The data acquisition module is communicatively connected to voltage and current acquisition devices, temperature sensors, pressure sensors, differential pressure sensors, flow sensors, liquid level sensors, gas purity detection devices, and a host scheduling system. The candidate operating condition generation module generates multiple candidate operating conditions based on operating data, control objectives, and equipment operating range. The electrochemical prediction module runs an electrochemical model. The thermal prediction module runs a thermal model. The purity prediction module runs a gas purity model. The coupling evaluation module correlates the prediction results of the electrochemical model, thermal model, and gas purity model according to the same candidate operating condition to form a collaborative prediction result. The optimization control module filters candidate operating conditions that meet the constraints from the collaborative prediction results and determines the target operating condition. The model correction module corrects the electrochemical model, thermal model, and gas purity model based on the actual operating data after the target operating condition is executed.

[0066] During the operation of an alkaline electrolysis system, the operating status of the alkaline electrolyzer is jointly affected by current, voltage, electrolyzer temperature, pressure, anode-cathode pressure difference, alkaline solution circulation flow rate, cooling water flow rate, makeup water volume, hydrogen production, hydrogen purity, and oxygen purity. Changes in current affect hydrogen production, electrolyzer voltage, and heat generation; changes in electrolyzer temperature affect electrolysis efficiency, electrolyzer impedance, and gas cross-permeability; pressure and anode-cathode pressure difference affect gas-liquid separation efficiency and gas purity. Therefore, this invention does not employ a single current, single power, or single temperature feedback control method. Instead, it incorporates energy consumption, temperature, and purity into the same predictive control process, selecting the target operating condition from multiple candidate operating conditions that satisfies the constraints and has the best overall target value.

[0067] like Figure 2 As shown, the method of the present invention includes the following steps:

[0068] S1. Obtain the operating data and control objectives of the alkaline electrolysis system.

[0069] The operating data includes current, voltage, electrolyzer temperature, pressure, anode-cathode pressure difference, alkali circulation flow rate, cooling water flow rate, makeup water volume, hydrogen production, hydrogen purity, and oxygen purity. Furthermore, the operating data may also include ambient temperature, alkali concentration, alkali level, hydrogen-side level, oxygen-side level, cooling water inlet temperature, cooling water outlet temperature, hydrogen-side gas-liquid separation pressure, oxygen-side gas-liquid separation pressure, DC power supply efficiency, alkali pump operating frequency, cooling water pump operating frequency, and valve opening.

[0070] Control objectives include target hydrogen production, target hydrogen purity, and target oxygen purity. Furthermore, control objectives may also include target operating modes, target power ranges, target temperature ranges, target pressure ranges, target energy consumption levels, or target load change rates. Target operating modes may include energy-priority mode, purity-priority mode, hydrogen production-priority mode, stable operation mode, electricity price response mode, and equipment protection mode.

[0071] In step S1, the data acquisition module acquires data from each sensor and actuator according to the control cycle. The control cycle can be set according to the response speed of the alkaline electrolysis system. The data acquisition module performs timestamp verification on data from different sources and maps data within the same control cycle to a unified time axis. For data with a high sampling frequency, smoothing is performed using moving average, median filtering, or low-pass filtering; for data with a low sampling frequency, the most recent valid value is preserved or interpolation is used to fill in the gaps; data that clearly exceeds the physical range is marked as abnormal data and removed. The processed operating data forms a status dataset.

[0072] In one specific implementation, the state dataset includes current current, current voltage, current electrolyzer temperature, current pressure, current anode-cathode pressure difference, current alkali circulation flow rate, current cooling water flow rate, current makeup water volume, current hydrogen production, current hydrogen purity, current oxygen purity, and a valid identifier for each data point. When a certain operating data point is temporarily unavailable, the controller can use historical data predictions or model estimates as a substitute. However, in subsequent optimization processes, the uncertainty weight corresponding to that data is increased to prevent the target operating condition selection from deviating from the actual operating state due to data loss.

[0073] S2. Generate multiple candidate operating conditions based on operating data, control objectives, and equipment operating range.

[0074] The equipment operating range includes allowable current range, allowable voltage range, allowable power range, allowable pressure range, allowable anode-cathode pressure difference range, allowable alkali circulation flow rate range, allowable cooling water flow rate range, allowable makeup water flow rate range, allowable temperature range, and actuator speed limits. Candidate operating conditions include current setpoints, pressure setpoints, alkali circulation flow rate setpoints, cooling water flow rate setpoints, and makeup water flow rate setpoints. Furthermore, candidate operating conditions may also include target voltage, target power, hydrogen-side pressure setpoints, oxygen-side pressure setpoints, alkali pump frequency setpoints, cooling water pump frequency setpoints, and pressure regulating valve opening setpoints.

[0075] The candidate operating condition generation module uses the current operating data as a benchmark and determines the initial search range for the current setpoint based on the target hydrogen production. If the current hydrogen production is lower than the target hydrogen production, the candidate set for the current setpoint is preferentially expanded towards the load increase direction; if the current hydrogen production is higher than the target hydrogen production, the candidate set for the current setpoint is preferentially expanded towards the load decrease direction. For pressure setpoints, alkali circulation flow rate setpoints, cooling water flow rate setpoints, and makeup water setpoints, the candidate operating condition generation module generates multiple candidate values ​​within the corresponding equipment operating range using a preset step size or an adaptive step size, and combines these candidate values ​​into a candidate combination. The preset step size can be determined based on the DC power supply adjustment accuracy, pump frequency adjustment accuracy, valve opening adjustment accuracy, and control cycle; the adaptive step size can be determined based on the deviation between the current operating data and the control target.

[0076] To avoid an excessive number of candidate combinations, the candidate operating condition generation module can perform pre-screening based on basic equipment constraints. Pre-screening includes eliminating candidate combinations that exceed the allowable output capacity of the DC power supply, exceed the actuator's acceleration / deceleration limits, may cause the anode-cathode pressure difference to exceed limits, have significantly insufficient cooling capacity, and fail to meet the water replenishment requirements for safe liquid levels. Multiple candidate operating conditions are obtained through pre-screening.

[0077] In a preferred embodiment, the candidate operating condition generation module can further determine the high-efficiency operating condition range based on historical operating data, and increase the sampling density within the high-efficiency operating condition range, while decreasing the sampling density within the low-efficiency or high-risk operating condition range. This reduces the computational load while ensuring optimization accuracy.

[0078] S3. Input the candidate operating conditions into the electrochemical model to predict the corresponding electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production.

[0079] The electrochemical model is used to describe the electrochemical response of the alkaline electrolyzer under candidate operating conditions. The inputs to the electrochemical model include current setpoints, pressure setpoints, alkali circulation flow rate setpoints, current electrolyzer temperature, alkali concentration, and model parameters related to the electrode and diaphragm states. The outputs of the electrochemical model include electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production.

[0080] In one specific implementation, the electrochemical model calculates electrochemical polarization loss, resistive loss, and Faraday efficiency based on current density, pressure, electrolyzer temperature, and alkaline circulation status. It then predicts the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production based on these parameters. Electrochemical polarization loss characterizes the voltage loss caused by electrode reaction kinetics; resistive loss characterizes the ohmic losses generated by the electrolyte, membrane, electrodes, and connecting structures; and Faraday efficiency characterizes the effective proportion of the input current actually converted into hydrogen and oxygen.

[0081] Within the current control cycle, the electrochemical model uses the current electrolyzer temperature as the initial temperature state for prediction. After obtaining the predicted temperature value, the thermal model uses this value as the temperature input for the gas purity model, and it can also be used as the temperature input for the electrochemical model during the next control cycle or the second iteration correction within the same control cycle. This approach avoids conflicts in the model calculation order and allows changes in thermal state to constrain subsequent energy consumption and purity predictions.

[0082] The energy consumption per unit of hydrogen production can be determined based on the correlation between input power and hydrogen production. Specifically, the controller first determines the input power of the electrolyzer based on the current setpoint under candidate operating conditions and the predicted electrolyzer voltage, and then determines the energy consumption per unit of hydrogen production based on the predicted hydrogen production. Since the energy consumption per unit of hydrogen production is a conventional energy consumption evaluation metric in this field, this embodiment does not limit its fixed calculation formula, as long as it can reflect the electrical energy consumption corresponding to the unit of hydrogen production.

[0083] In a preferred embodiment, the electrochemical model can employ a mechanistic model, a semi-empirical model, a lookup table model, a data-driven model, or a hybrid model. The mechanistic model can predict the electrolyzer voltage based on reversible voltage, activation overvoltage, ohmic overvoltage, and concentration overvoltage. The semi-empirical model can fit the relationship between current, temperature, pressure, and electrolyzer voltage based on experimental data. The lookup table model can look up voltage, power, and hydrogen production based on historical operating data under different current, temperature, and pressure conditions. The data-driven model can employ a regression model, a neural network model, a support vector machine model, a random forest model, or a gradient boosting tree model. The hybrid model can first obtain the basic prediction results through the mechanistic model, and then correct the deviations of the basic prediction results through a data correction model.

[0084] S4. Input the candidate operating conditions and the prediction results of the electrochemical model into the thermal model to predict the corresponding electrolytic cell temperature state.

[0085] The thermal model is used to predict the thermal response of an alkaline electrolyzer under candidate operating conditions. The inputs to the thermal model include the candidate operating conditions, electrolyzer input power, hydrogen production, oxygen production, current electrolyzer temperature, ambient temperature, alkaline solution circulation flow rate, cooling water flow rate, cooling water inlet temperature, cooling water outlet temperature, and thermal model parameters. The output of the thermal model is the electrolyzer temperature status. The electrolyzer temperature status includes the temperature change trend and predicted temperature value, and can further include the rate of temperature rise, rate of temperature fall, remaining time to reach the upper temperature limit, and the degree of temperature uniformity in different regions.

[0086] In one specific implementation, the thermal model predicts the electrolyzer temperature state based on input power, reaction heat generation, ambient heat dissipation, alkaline solution circulation heat exchange, and cooling heat exchange. Specifically, the thermal model first determines the input power corresponding to the candidate operating conditions based on the electrolyzer voltage and current setpoints predicted by the electrochemical model, then determines the heat generation trend based on irreversible electrochemical losses; subsequently, it determines the heat removed through the alkaline solution circulation and cooling water flow rates; finally, it predicts the electrolyzer temperature state in the next control cycle or multiple future control cycles by combining ambient heat dissipation and the equivalent heat capacity of the alkaline electrolyzer.

[0087] In a preferred embodiment, the thermal model can use the current electrolyzer temperature, input power, alkali circulation flow rate, and cooling water flow rate as state variables, and predict the electrolyzer temperature state through a state-space model, a thermal balance model, or a data-driven time-series model. When the alkaline electrolyzer includes multiple temperature monitoring points, the thermal model can predict the temperature value of each monitoring point separately, and select the highest predicted temperature value as the predicted temperature value in the constraint screening. It can also determine the risk of uneven temperature distribution based on the temperature difference between the monitoring points.

[0088] S5. Input the candidate operating conditions, the prediction results of the electrochemical model and the prediction results of the thermal model into the gas purity model to predict the corresponding hydrogen purity and oxygen purity.

[0089] The gas purity model is used to predict the purity of hydrogen and oxygen under candidate operating conditions. The inputs to the gas purity model include hydrogen production, oxygen production, electrolyzer temperature and state, pressure, anode-cathode pressure difference, alkali circulation flow rate, gas-liquid separation efficiency, diaphragm state parameters, and gas entrainment correction coefficient. The outputs of the gas purity model include hydrogen purity, oxygen purity, gas cross-permeation risk, and gas-liquid entrainment risk.

[0090] In one specific implementation, the gas purity model predicts the risks of gas cross-permeation and gas-liquid entrainment based on hydrogen production, oxygen production, electrolyzer temperature and pressure, anode-cathode pressure difference, alkali circulation flow rate, and gas-liquid separation efficiency. Based on these risks, the model then predicts the purity of hydrogen and oxygen. The gas cross-permeation risk characterizes the likelihood of gas mixing between the hydrogen and oxygen sides; the gas-liquid entrainment risk characterizes the possibility that insufficient gas-liquid separation will lead to alkali or bubble entrainment, affecting gas purity.

[0091] In a preferred embodiment, a gas purity model is used to construct a purity risk index. The purity risk index can be expressed by the following formula:

[0092] ;

[0093] in, This represents the purity risk index corresponding to the i-th candidate operating condition; This represents the normalized anode-cathode pressure difference corresponding to the i-th candidate operating condition; This represents the normalized predicted temperature value corresponding to the i-th candidate operating condition; This represents the normalized gas production intensity corresponding to the i-th candidate operating condition; This represents the normalized alkaline solution circulation flow rate corresponding to the i-th candidate operating condition; This represents the normalized gas-liquid separation efficiency corresponding to the i-th candidate operating condition. , , , and This is the purity risk weighting coefficient.

[0094] In the above formula, as the anode-cathode pressure difference, predicted temperature, and gas production intensity increase, the risk of gas cross-permeation or gas-liquid entrainment usually increases; as the alkaline solution circulation flow rate and gas-liquid separation efficiency increase, gas discharge and gas-liquid separation capabilities are enhanced, and purity risk decreases. This purity risk index can be used to correct the predicted results of hydrogen and oxygen purity, and can also be used for subsequent candidate operating condition screening.

[0095] In practice, the purity risk weight coefficient can be obtained by calibrating historical operating data. Specifically, the controller collects data on anode-cathode pressure difference, predicted temperature, gas production intensity, alkali circulation flow rate, gas-liquid separation efficiency, actual hydrogen purity, and actual oxygen purity under different operating conditions. Samples whose actual hydrogen or oxygen purity does not meet the requirements are marked as high-risk samples, while samples whose purity meets the requirements and has a large margin are marked as low-risk samples. The purity risk weight coefficient is then determined through regression, classification, or least squares fitting.

[0096] Furthermore, the gas purity model can establish a mapping relationship between purity risk indicators and the decrease in hydrogen and oxygen purity based on historical purity detection data. It then subtracts the corresponding decrease from baseline hydrogen and oxygen purity to obtain predicted hydrogen and oxygen purity under candidate operating conditions. The baseline hydrogen and oxygen purity can be determined based on current actual purity, average purity under recent stable operating conditions, or design purity. The mapping relationship can employ piecewise linear mapping, regression mapping, lookup table mapping, or data-driven mapping. Therefore, the gas purity model can not only output risk levels but also predictable hydrogen and oxygen purity that can be used for constraint screening.

[0097] S6. Couple the evaluation of unit hydrogen production energy consumption, electrolyzer temperature status, hydrogen purity and oxygen purity to form a synergistic prediction result.

[0098] The coupling relationship of the three models is as follows: Figure 3 As shown, the electrochemical model, thermal model, and gas purity model are not independent unidirectional models, but are coupled by sharing state variables and prediction results. Specifically, within the current control cycle, the electrochemical model uses the current electrolyzer temperature as the initial temperature state to predict the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production; the thermal model predicts the electrolyzer temperature state based on candidate operating conditions and the prediction results of the electrochemical model; and the gas purity model predicts hydrogen and oxygen purity based on candidate operating conditions, the hydrogen and oxygen production predicted by the electrochemical model, and the electrolyzer temperature state predicted by the thermal model. For application scenarios requiring improved computational accuracy, the controller can perform a second iteration within the same control cycle, re-inputting the electrolyzer temperature state predicted by the thermal model into the electrochemical model to re-predict the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production, and updating the input of the gas purity model.

[0099] The coupled evaluation module correlates the unit hydrogen production energy consumption, electrolyzer temperature status, hydrogen purity, and oxygen purity under the same candidate operating condition to form a collaborative prediction result. The collaborative prediction result includes at least the candidate operating condition number, candidate operating condition parameters, predicted electrolyzer voltage, predicted hydrogen production, predicted oxygen production, unit hydrogen production energy consumption, predicted temperature value, temperature change trend, predicted hydrogen purity, predicted oxygen purity, purity risk index, and valid identifier of the candidate operating condition.

[0100] In a preferred embodiment, the coupled evaluation module normalizes the prediction results for different dimensions. For unit hydrogen production energy consumption, a lower value is better; for predicted temperature values, the closer to the target temperature range, the better, and values ​​exceeding the upper temperature limit are directly judged as not meeting the constraints; for hydrogen and oxygen purity, higher than the target purity is better; for the amplitude of operating condition changes, a smaller amplitude is more conducive to system stability. After normalization, the collaborative prediction results can be directly used for constraint screening and comprehensive target value calculation.

[0101] S7. Select candidate operating conditions that meet the constraints from the collaborative prediction results according to the control objectives and equipment operating range, and determine the target operating condition based on the comprehensive target value.

[0102] The optimization control module first performs constraint screening on each candidate operating condition. Constraints include: predicted hydrogen production not lower than the target hydrogen production; predicted hydrogen purity not lower than the target hydrogen purity; predicted oxygen purity not lower than the target oxygen purity; predicted temperature not exceeding the upper temperature limit; predicted pressure within the allowable pressure range; current within the allowable current range; voltage within the allowable voltage range; alkali circulation flow rate within the allowable flow rate range; cooling water flow rate within the allowable range of the cooling system; and makeup water volume within the allowable range of the makeup water system. For candidate operating conditions that do not meet any of the constraints, the optimization control module eliminates them; for candidate operating conditions that meet the constraints, the optimization control module calculates the comprehensive target value.

[0103] In a preferred embodiment, the overall target value is expressed by the following formula:

[0104] ;

[0105] in, This represents the comprehensive target value for the i-th candidate operating condition; This represents the normalized unit hydrogen production energy consumption corresponding to the i-th candidate operating condition; This represents the normalized purity deviation corresponding to the i-th candidate operating condition; This represents the normalized temperature deviation corresponding to the i-th candidate operating condition; This represents the normalized operating cost corresponding to the i-th candidate operating condition; This represents the normalized operating condition change penalty corresponding to the i-th candidate operating condition; , , , and These represent the weight coefficients of the corresponding items.

[0106] The purity deviation is determined based on the margins of the predicted hydrogen purity relative to the target hydrogen purity and the predicted oxygen purity relative to the target oxygen purity. When the predicted hydrogen or oxygen purity is lower than the corresponding target purity, the candidate operating condition has been eliminated during the constraint screening stage; when both the predicted hydrogen and oxygen purity are higher than the corresponding target purity, the larger the purity margin, the smaller the purity deviation. The temperature deviation is determined based on the degree of deviation of the predicted temperature value from the target temperature range. Operating costs can be estimated based on electricity prices, water consumption, auxiliary equipment power consumption, and maintenance costs. The penalty for changes in operating conditions is determined based on the magnitude of changes in current, pressure, flow rate, and cooling adjustment between the candidate operating condition and the current operating condition, to avoid frequent and significant changes in control parameters.

[0107] The optimization control module selects the candidate operating condition with the smallest comprehensive target value from the candidate operating conditions that meet the constraints as the target operating condition. When there are multiple candidate operating conditions with the same or similar comprehensive target values, the candidate operating condition with the smaller penalty for operating condition changes is given priority; if the penalties for operating condition changes are also similar, the candidate operating condition with the larger purity margin is given priority; if the purity margins are also similar, the candidate operating condition with the predicted temperature value far from the upper temperature limit is given priority.

[0108] The weighting coefficients in the overall target value can differ under different operating modes. For example, in energy-priority mode, increasing w... E In purity-first mode, increase w P In stable operating mode, improve w U In device protection mode, improve w T Furthermore, the upper temperature limit and the anode-cathode pressure difference limit are tightened. Through weight adjustment, the same set of collaborative predictive control methods can adapt to different operational objectives.

[0109] S8. Convert the target operating condition into control parameters and output them to the DC power supply, alkali circulation system, cooling system, pressure regulation system and water replenishment system.

[0110] Once the target operating conditions are determined, the optimization control module converts these conditions into control parameters that each actuator can recognize. These control parameters include the target current, target voltage, or target power of the DC power supply; the frequency of the alkali pump or the alkali circulation flow rate of the alkali circulation system; the frequency of the cooling water pump, the opening degree of the cooling water valve, or the cooling water flow rate of the cooling system; the setpoints for the hydrogen and oxygen sides or the opening degree of the pressure regulating valve of the pressure regulating system; and the start / stop status of the makeup water pump, the opening degree of the makeup water valve, or the makeup water volume of the makeup water system.

[0111] In one specific implementation, the optimization control module does not directly output the target control parameters in a step manner, but instead generates a smooth control sequence based on the actuator's acceleration and deceleration limits. For example, when the difference between the target current and the current current exceeds the maximum allowable change in a single cycle, the optimization control module decomposes the target current into multiple control cycles and executes them step by step; when the difference between the target pressure and the current pressure exceeds the maximum allowable change in a single cycle, the optimization control module gradually adjusts the pressure regulating valve opening; when the difference between the target cooling water flow rate and the current cooling water flow rate is large, the optimization control module prioritizes increasing the cooling water pump frequency and then adjusts the valve opening to reduce temperature fluctuations.

[0112] When the predicted temperature value corresponding to the target operating condition approaches the upper temperature limit, the optimization control module can increase the cooling water flow rate or decrease the current setpoint in advance. When the purity risk index corresponding to the target operating condition approaches the upper risk limit, the optimization control module can reduce the anode-cathode pressure difference, increase the alkali circulation flow rate, or adjust the pressure setpoint to improve the gas purity margin. Thus, the output of the control parameters not only meets the current target operating condition but also reflects the proactive suppression of temperature and purity risks.

[0113] S9. Obtain the actual operating data after the target operating condition is executed, and correct the electrochemical model, thermal model and gas purity model based on the error between the prediction results and the actual operating data.

[0114] After the alkaline electrolysis system operates according to the target conditions, the data acquisition module continues to collect actual operating data. The closed-loop control and model correction process is as follows: Figure 4 As shown. Actual operating data includes actual voltage, actual temperature, actual hydrogen production, actual hydrogen purity, and actual oxygen purity. It can also further include actual pressure, actual anode-cathode pressure difference, actual alkali circulation flow rate, actual cooling water flow rate, and actual makeup water volume. The model correction module compares the actual operating data with the predicted results corresponding to the target operating conditions to obtain the prediction error.

[0115] When the prediction error does not exceed the set threshold, the model correction module keeps the model parameters unchanged or uses a small-step update method for smooth correction. When the prediction error exceeds the set threshold, the model correction module corrects the model parameters. The model parameters that can be corrected include polarization parameters, ohmic impedance parameters, and Faraday efficiency parameters in the electrochemical model; equivalent heat capacity, heat dissipation coefficient, and cooling heat transfer coefficient in the thermal model; and gas impurity correction coefficient and gas-liquid separation efficiency parameters in the gas purity model.

[0116] In a preferred embodiment, the model parameters are updated using the following logic:

[0117] ;

[0118] in, Represents the model parameters for the k-th control cycle; This represents the model parameters for the (k+1)th control cycle; Indicates the correction step size; The sensitivity matrix represents the relationship between error and model parameters; This represents the error vector between the predicted results and the actual operating data. The error vector can include the voltage error between the actual voltage and the predicted electrolyzer voltage, the temperature error between the actual temperature and the predicted temperature, the hydrogen production error between the actual hydrogen production and the predicted hydrogen production, the hydrogen purity error between the actual hydrogen purity and the predicted hydrogen purity, and the oxygen purity error between the actual oxygen purity and the predicted oxygen purity. The sensitivity matrix can be obtained offline through historical operating samples, or it can be estimated online through the correspondence between model parameter perturbations and prediction error changes within adjacent control cycles.

[0119] In actual operation, to avoid abrupt changes in model parameters caused by abnormal sampling data, the model correction module can set an upper limit for the correction magnitude and abnormal sample filtering rules. When a certain actual operating data shows sensor failure, communication interruption, or obvious deviation from physical laws, the model correction module will not use this data to update the model parameters. When prediction errors in the same direction occur for multiple consecutive control cycles, the model correction module increases the correction weight of the corresponding parameter; when the prediction error changes frequently in the positive and negative directions, the model correction module reduces the correction step size to improve model stability.

[0120] In step S9, the revised electrochemical model, thermal model, and gas purity model are used for candidate operating condition prediction and optimization in the next control cycle. As operating data accumulates, the model prediction results gradually approach the actual performance state of the current alkaline electrolysis system, thereby improving the reliability of the target operating condition.

[0121] Example 1: This example uses an alkaline electrolysis hydrogen production system as an example to illustrate the specific operation process of the present invention. The system includes an alkaline electrolyzer, a DC power supply, an alkaline solution circulation pump, a cooling water pump, a heat exchanger, a hydrogen-side pressure regulating valve, an oxygen-side pressure regulating valve, a makeup water pump, a hydrogen gas-liquid separator, an oxygen gas-liquid separator, an online gas analyzer, and a controller.

[0122] After the system starts, the controller first acquires the current operating data, including the current current, current voltage, current electrolyzer temperature, current hydrogen-side pressure, current oxygen-side pressure, current anode-cathode pressure difference, current alkali circulation flow rate, current cooling water flow rate, current makeup water volume, current hydrogen production, current hydrogen purity, and current oxygen purity. Simultaneously, the controller receives control targets from the host scheduling system, including target hydrogen production, target hydrogen purity, and target oxygen purity.

[0123] The controller generates multiple candidate operating conditions based on the equipment's operating range. Each candidate operating condition includes current setpoints, pressure setpoints, alkali circulation flow rate setpoints, cooling water flow rate setpoints, and makeup water setpoints. For each candidate operating condition, the controller first predicts the electrolyzer voltage, hydrogen production, oxygen production, and unit hydrogen production energy consumption using an electrochemical model; then it predicts the electrolyzer temperature state using a thermal model; and finally, it predicts the hydrogen and oxygen purity using a gas purity model.

[0124] The controller correlates the unit hydrogen production energy consumption, electrolyzer temperature status, hydrogen purity, and oxygen purity corresponding to the same candidate operating condition to form a collaborative prediction result. Subsequently, the controller eliminates candidate operating conditions where the predicted hydrogen production is lower than the target hydrogen production, the predicted hydrogen purity is lower than the target hydrogen purity, the predicted oxygen purity is lower than the target oxygen purity, or the predicted temperature exceeds the upper temperature limit or pressure exceeds the limit. For the remaining candidate operating conditions, the controller calculates a comprehensive target value and selects the candidate operating condition with the smallest comprehensive target value as the target operating condition.

[0125] Once the target operating conditions are determined, the controller outputs the target current or target power to the DC power supply, the target frequency to the alkali circulation pump, the target frequency to the cooling water pump, the target opening degree to the pressure regulating valve, and a water replenishment control command to the water replenishment pump. After execution, the controller continues to collect actual voltage, actual temperature, actual hydrogen production, actual hydrogen purity, and actual oxygen purity, and compares them with the predicted results. When the prediction error exceeds a set threshold, the controller corrects the electrochemical model, thermal model, and gas purity model to make the collaborative prediction results for the next control cycle closer to the actual operating conditions.

[0126] Example 2: This example illustrates the application of the present invention in scenarios involving fluctuations in the input power of new energy sources. When the alkaline electrolysis system is connected to wind or photovoltaic power, the upper-level dispatch system can issue control targets based on the predicted available power for the next short period. In addition to target hydrogen production, target hydrogen purity, and target oxygen purity, the control targets also include an upper limit on available power and load change restrictions.

[0127] When generating candidate operating conditions, the controller incorporates the available power limit as part of the equipment's operating range. The predicted input power corresponding to a candidate operating condition must not exceed the available power limit, and the change in target current relative to the current must not exceed the load change limit. When renewable energy power decreases, the controller generates multiple load reduction candidate operating conditions and predicts the unit hydrogen production energy consumption, electrolyzer temperature state, hydrogen purity, and oxygen purity under each candidate operating condition using electrochemical, thermal, and gas purity models. If a candidate operating condition has low energy consumption but the predicted hydrogen purity is lower than the target hydrogen purity, it is eliminated. If a candidate operating condition has a high purity margin but requires an excessively large operating condition change, its overall target value increases due to the operating condition change penalty.

[0128] When the power output of the renewable energy source increases, the controller generates multiple candidate operating conditions for load increase. For each candidate operating condition, the thermal model predicts the temperature change trend in advance, and the gas purity model predicts the risks of gas cross-permeation and gas-liquid entrainment. If the predicted temperature value after load increase is close to the upper temperature limit, the controller can prioritize the candidate operating condition that simultaneously increases the cooling water flow rate or the alkaline solution circulation flow rate; if the purity risk index increases after load increase, the controller can prioritize the candidate operating condition with a more gradual pressure change, a smaller anode-cathode pressure difference, or a higher alkaline solution circulation capacity. Thus, the alkaline electrolysis system can achieve dynamic adjustment under the condition of fluctuation in renewable energy input power, while taking into account energy consumption, temperature, and purity.

[0129] Example 3: This example illustrates the application of the present invention in a scenario where multiple alkaline electrolyzers operate in parallel. When a hydrogen production station includes multiple alkaline electrolyzers, the controller can acquire the operating data of each alkaline electrolyzer and establish corresponding electrochemical, thermal, and gas purity models for each. After the upper-level scheduling system issues the total target hydrogen production, total power limit, and gas purity requirements, the controller first determines the candidate load allocation scheme for each alkaline electrolyzer based on its health status, current temperature, current purity margin, and unit hydrogen production energy consumption level.

[0130] For each candidate load allocation scheme, the controller inputs the candidate operating conditions of each alkaline electrolyzer into the corresponding model for prediction, obtaining the unit hydrogen production energy consumption, electrolyzer temperature state, hydrogen purity, and oxygen purity for each alkaline electrolyzer. The controller eliminates any candidate load allocation scheme that does not meet the requirements for cell temperature or purity, and selects the scheme with the lowest total unit hydrogen production energy consumption, the highest temperature margin for each alkaline electrolyzer, and the smallest load variation range for each alkaline electrolyzer as the target load allocation scheme. Subsequently, the controller converts the target load allocation scheme into the target operating conditions for each alkaline electrolyzer and outputs the control parameters accordingly.

[0131] When the temperature of a certain alkaline electrolyzer approaches its upper temperature limit, or the purity of hydrogen or oxygen approaches its lower purity limit, the controller can transfer part of the load of that alkaline electrolyzer to other alkaline electrolyzers with larger temperature and purity margins. Thus, while meeting the overall target hydrogen production, the hydrogen production station can achieve a more stable operating state.

[0132] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for coordinated predictive control of energy consumption, temperature, and purity in an alkaline electrolysis system, characterized in that, include: S1. Obtain the operating data and control targets of the alkaline electrolysis system, wherein the control targets include target hydrogen production, target hydrogen purity, and target oxygen purity; S2. Generate multiple candidate operating conditions based on the operating data, the control target, and the equipment operating range. Each candidate operating condition includes a current setting value, a pressure setting value, an alkaline solution circulation flow setting value, a cooling water flow setting value, and a makeup water setting value. S3. Input the candidate operating conditions into the electrochemical model to predict the corresponding electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production. S4. Input the candidate operating conditions and the prediction results of the electrochemical model into the thermal model to predict the corresponding electrolytic cell temperature state; S5. Input the candidate operating conditions, the prediction results of the electrochemical model and the prediction results of the thermal model into the gas purity model to predict the corresponding hydrogen purity and oxygen purity. S6. The unit hydrogen production energy consumption, the temperature state of the electrolyzer, the hydrogen purity and the oxygen purity are coupled and evaluated to form a synergistic prediction result; S7. Select candidate operating conditions that meet the constraints from the collaborative prediction results according to the control target and the equipment operating range, and determine the target operating condition based on the comprehensive target value; S8. Convert the target operating condition into control parameters and output them to the DC power supply, alkali circulation system, cooling system, pressure regulation system and water replenishment system; S9. Obtain the actual operating data after the target operating condition is executed, and correct the electrochemical model, the thermal model and the gas purity model according to the error between the prediction result and the actual operating data.

2. The method for coordinated predictive control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S1, the step of obtaining the running data includes: Time alignment is performed on the data collected from the alkaline electrolytic cell, DC power supply, alkaline solution circulation system, cooling system, pressure regulation system, water replenishment system, and gas purity detection device. Perform outlier removal and smoothing on the time-aligned collected data; A state dataset is generated based on the processed collected data to characterize the current operating state.

3. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S2, the steps for generating multiple candidate operating conditions include: Based on the current operating data, adjust the current setting value, the pressure setting value, the alkali circulation flow rate setting value, the cooling water flow rate setting value, and the water replenishment setting value respectively within the operating range of the equipment; The adjusted settings are combined to form candidate combinations; Candidate operating conditions are obtained by eliminating candidate combinations that exceed the acceleration / deceleration limit, power limit, and flow limit.

4. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S3, the electrochemical model calculates the electrochemical polarization loss, resistive loss, and Faraday efficiency based on the current density, pressure, electrolyzer temperature, and alkaline circulation state, and predicts the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production based on the electrochemical polarization loss, resistive loss, and Faraday efficiency.

5. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S4, the thermal model predicts the temperature state of the electrolytic cell based on the input power, reaction heat generation, environmental heat dissipation, alkaline solution circulation heat exchange, and cooling heat exchange. The temperature state of the electrolytic cell includes the temperature change trend and the predicted temperature value.

6. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S5, the prediction step of the gas purity model includes: extracting the anode-cathode pressure difference from the operating data; inputting the hydrogen production, oxygen production, electrolyzer temperature state, pressure setpoint and alkali circulation flow rate setpoint in the candidate operating conditions, the anode-cathode pressure difference, and gas-liquid separation efficiency into the gas purity model to predict the risk of gas cross-permeation and gas-liquid entrainment; and predicting the hydrogen purity and oxygen purity based on the risk of gas cross-permeation and gas-liquid entrainment.

7. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S6, the steps for forming the collaborative prediction result include: The electrolyzer temperature state predicted by the thermal model is used as the temperature input of the gas purity model, and as the temperature input of the electrochemical model during the next control cycle or the second iteration correction in the same control cycle. The hydrogen and oxygen production predicted by the electrochemical model are used as the gas production inputs of the gas purity model. The unit hydrogen production energy consumption, the electrolyzer temperature status, the hydrogen purity, and the oxygen purity are correlated according to the same candidate operating condition to form the collaborative prediction result.

8. The method for coordinated prediction and control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S7, the constraints include: The predicted hydrogen production is not less than the target hydrogen production. The predicted hydrogen purity is not lower than the target hydrogen purity; The predicted oxygen purity is not lower than the target oxygen purity; The predicted temperature value will not exceed the upper temperature limit; The predicted pressure is within the allowable pressure range; The current, voltage, alkaline solution circulation flow rate, cooling water flow rate, and makeup water volume are all within their respective allowable ranges.

9. The method for coordinated predictive control of energy consumption, temperature, and purity in an alkaline electrolysis system according to claim 1, characterized in that, In step S9, the step of correcting the electrochemical model, the thermal model, and the gas purity model includes: The actual voltage, actual temperature, actual hydrogen production, actual hydrogen purity, and actual oxygen purity are compared with the corresponding prediction results to obtain the prediction error. When the prediction error exceeds a set threshold, the model parameters are corrected. The model parameters include polarization parameters, ohmic impedance parameters, and Faraday efficiency parameters in the electrochemical model; equivalent heat capacity, heat dissipation coefficient, and cooling heat transfer coefficient in the thermal model; and gas mixing correction coefficient and gas-liquid separation efficiency parameters in the gas purity model.

10. A coordinated predictive control system for energy consumption, temperature, and purity in an alkaline electrolysis system, characterized in that, The system is used to perform the energy consumption-temperature-purity coordinated prediction and control method for an alkaline electrolysis system according to any one of claims 1 to 9, the system comprising: The data acquisition module is used to acquire the operating data and control objectives of the alkaline electrolysis system; The candidate operating condition generation module is used to generate multiple candidate operating conditions based on the operating data, the control target, and the equipment operating range. The electrochemical prediction module is used to predict the electrolyzer voltage, hydrogen production, oxygen production, and energy consumption per unit of hydrogen production using an electrochemical model. The thermal prediction module is used to predict the temperature state of the electrolyzer using a thermal model. The purity prediction module is used to predict the purity of hydrogen and oxygen using a gas purity model. The coupled evaluation module is used to perform coupled evaluation of the unit hydrogen production energy consumption, the temperature state of the electrolyzer, the hydrogen purity and the oxygen purity to form a synergistic prediction result; The optimization control module is used to filter candidate operating conditions that meet the constraints from the collaborative prediction results, determine the target operating condition, and output control parameters. The model correction module is used to correct the electrochemical model, the thermal model, and the gas purity model based on actual operating data.

Images