Coordinated optimization control method based on wind-solar-hydrogen combined power generation system

Abstract

The application discloses a collaborative optimization control method based on a wind-solar-hydrogen combined power generation system, relates to the technical field of new energy collaborative control, and comprises the following steps: collecting a wind power actual measurement sequence, a photovoltaic power actual measurement sequence, an electrolytic cell working state signal, a hydrogen storage tank pressure actual measurement value and a fuel cell power actual measurement sequence; quantifying the uncertainty of wind-solar power to generate a prediction interval; obtaining adjustable margins and hydrogen storage state indexes by evaluating the states of all devices; and constructing a multi-time scale collaborative optimization model. An improved model predictive control algorithm based on the dynamic adjustment of the constraint boundary of the hydrogen storage state index is used for solving, so that a future multi-period optimization control instruction set is obtained and is issued to corresponding controllers for execution. The method fully considers the uncertainty of wind-solar power and the change of the hydrogen storage state, realizes the collaborative operation of multiple devices and the adaptive adjustment of control parameters, improves the stability and control robustness of the system operation, and enhances the adaptability of the combined power generation system to fluctuating working conditions.

Description

Technical Field

[0001] This invention belongs to the field of new energy collaborative control technology, specifically a collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system. Background Technology

[0002] Traditional wind-solar-hydrogen storage combined power generation systems often rely on deterministic power prediction results for optimization calculations, failing to quantify the uncertainties in wind and solar power sequences, thus failing to reflect power fluctuation risks in the model. Furthermore, the status of each system component is monitored independently, and the adjustability margin of the electrolyzer, hydrogen storage status, and adjustable power of the fuel cell are not integrated with power prediction information, making it difficult to construct a multi-dimensional, multi-timescale collaborative optimization model.

[0003] Conventional model predictive control algorithms use fixed constraint boundaries, making it impossible to adjust optimization conditions in real time based on hydrogen storage state indicators. When hydrogen storage is too high or too low, the control strategy does not match the actual system capacity. Optimization instructions are generated only for a single control cycle, failing to cover multiple future cycles, resulting in insufficient continuity of control actions and weak overall system coordination and disturbance rejection capabilities.

[0004] This invention quantifies the uncertainty of wind and solar power and constructs a multi-timescale model in conjunction with equipment status. It adopts an improved model predictive control algorithm that dynamically adjusts the constraint boundary based on the hydrogen storage status to solve multi-cycle optimization instructions and issue them for execution, thereby improving the system's adaptive and collaborative operation capabilities. Summary of the Invention

[0005] This invention aims to solve at least one of the technical problems existing in the prior art;

[0006] Therefore, this invention proposes a collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system, including:

[0007] From the combined power generation system consisting of wind turbine, photovoltaic array, electrolyzer, hydrogen storage tank and fuel cell, the measured power sequence of wind power, the measured power sequence of photovoltaic array, the working status signal of electrolyzer, the measured pressure value of hydrogen storage tank and the measured power sequence of fuel cell are periodically collected;

[0008] Uncertainty quantification is performed on the measured wind power and photovoltaic power sequences to generate wind power prediction intervals and photovoltaic power prediction intervals. State assessment is performed on the electrolyzer working status signal, the measured value of hydrogen storage tank pressure and the measured fuel cell power sequence to generate the electrolyzer adjustability margin, hydrogen storage status index and fuel cell adjustability power.

[0009] Based on the wind power prediction range, photovoltaic power prediction range, electrolyzer adjustable margin, hydrogen storage state index and fuel cell adjustable power, a multi-timescale collaborative optimization model is constructed.

[0010] An improved model predictive control algorithm is used to solve the multi-timescale collaborative optimization model. The improved model predictive control algorithm dynamically adjusts the constraint boundary of the optimization problem based on the hydrogen storage state index.

[0011] Solving the multi-timescale collaborative optimization model yields an optimized control instruction set for multiple future control cycles;

[0012] The optimized control instruction set is sent to the corresponding electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve in the combined power generation system for execution.

[0013] Furthermore, the periodic acquisition of wind power measurement sequences, photovoltaic power measurement sequences, electrolyzer operating status signals, hydrogen storage tank pressure measurement values, and fuel cell power measurement sequences includes:

[0014] The wind turbine generator monitoring and data acquisition system reads the wind turbine speed, pitch angle and generator output electrical parameters at a fixed sampling frequency, and calculates and generates the measured wind power sequence.

[0015] The photovoltaic array's combiner box and inverter monitoring unit read the DC side current and voltage and the AC side output electrical parameters at a fixed sampling frequency, and calculate and generate the photovoltaic power measured sequence.

[0016] The input / output module of the electrolytic cell control cabinet reads the input current, input voltage, operating temperature, and alarm status words of the electrolytic cell, and combines them to generate the operating status signal of the electrolytic cell.

[0017] The measured pressure value of the hydrogen storage tank is read and output in real time by a pressure transmitter installed on the hydrogen storage tank.

[0018] The power management unit of the fuel cell system reads its net output electrical parameters at a fixed sampling frequency, and calculates and generates the measured power sequence of the fuel cell.

[0019] Furthermore, uncertainty quantification is performed on the measured wind power sequence and the measured photovoltaic power sequence to generate wind power prediction intervals and photovoltaic power prediction intervals, including:

[0020] Empirical mode decomposition was performed on the measured wind power sequence to obtain multiple wind electronic mode components that reflect the fluctuation characteristics at different time scales;

[0021] An autoregressive integral moving average prediction model is established for each of the wind electronic mode components, and the prediction uncertainty band of each wind electronic mode component is calculated based on the prediction residual distribution of the autoregressive integral moving average prediction model.

[0022] The prediction uncertainty of all wind electronic mode components is superimposed and reconstructed in time series to form a total wind power prediction range covering future periods;

[0023] A similar decomposition, prediction, and uncertainty band calculation process is performed on the measured photovoltaic power sequence to generate the corresponding photovoltaic power prediction interval.

[0024] Furthermore, a status assessment is performed on the electrolyzer operating status signal, the measured pressure of the hydrogen storage tank, and the measured power sequence of the fuel cell to generate the electrolyzer adjustability margin, hydrogen storage status index, and fuel cell adjustability power, including:

[0025] The electrolytic cell's operating status signal is analyzed to obtain the current operating point, temperature, and historical operating time of the electrolytic cell. Based on the electrolytic cell performance degradation model, the maximum allowable power ramp-up rate and the maximum allowable power reduction depth of the electrolytic cell in the current and short future time are calculated. The maximum allowable power ramp-up rate and the maximum allowable power reduction depth together characterize the adjustability margin of the electrolytic cell.

[0026] Based on the measured pressure of the hydrogen storage tank and its historical trend, combined with the upper and lower limits of the design pressure of the hydrogen storage tank, the deviation of the current pressure from the median of safe operation and the trend coefficient based on the recent pressure change rate are calculated. The hydrogen storage status index is composed of the pressure deviation and the pressure trend coefficient by weighted combination.

[0027] By analyzing the fluctuation characteristics of the measured power sequence of the fuel cell, and combining the rated power of the fuel cell with the current efficiency curve, the maximum power that the fuel cell can be adjusted up and the minimum power that can be adjusted down in real time under the premise of maintaining high efficiency operation are determined. The maximum power that can be adjusted up and the minimum power that can be adjusted down in real time together define the adjustable power of the fuel cell.

[0028] Furthermore, based on the aforementioned wind power prediction range, photovoltaic power prediction range, electrolyzer adjustability margin, hydrogen storage state index, and fuel cell adjustability power, a multi-timescale collaborative optimization model is constructed, including:

[0029] An optimization objective function is established with the goal of minimizing the total operating cost of the combined power generation system. The optimization objective function includes a wind curtailment penalty term, a hydrogen supply and demand imbalance penalty term, and an operating wear and tear cost term for the electrolyzer and fuel cell.

[0030] A constraint system is established, which includes: range constraints that require the actual output power of wind turbines and photovoltaic arrays to fluctuate within their corresponding wind power prediction range and photovoltaic power prediction range; equipment constraints that require the power adjustment of the electrolyzer to be within its adjustable margin range; pressure changes in the hydrogen storage tank to meet the pressure safety trajectory constraints dynamically defined by the hydrogen storage status index; power constraints that require the output power adjustment of the fuel cell to be within its adjustable power range; and power balance constraints of the system at multiple predefined time points.

[0031] The objective function and the constraint system together constitute the multi-timescale collaborative optimization model.

[0032] Furthermore, the improved model predictive control algorithm is used to solve the multi-timescale collaborative optimization model. The improved model predictive control algorithm dynamically adjusts the constraint boundaries of the optimization problem based on the hydrogen storage state index, including:

[0033] At the beginning of each control cycle, the improved model predictive control algorithm is invoked. The improved model predictive control algorithm takes the currently collected system state as the initial point and performs rolling solution on the multi-time scale collaborative optimization model within the preset prediction time domain.

[0034] Before each rolling solution, the upper and lower boundaries of the pressure safety trajectory constraint of the hydrogen storage tank in the multi-timescale collaborative optimization model are dynamically calculated and updated based on the real-time updated value of the hydrogen storage state index. When the hydrogen storage state index shows that the pressure is too high, the upper pressure constraint is tightened; when the pressure is too low, the lower pressure constraint is tightened.

[0035] After dynamically updating the constraint boundaries, an interior point solver is used to solve the updated multi-timescale collaborative optimization model to obtain the preliminary optimized control sequence in the current prediction time domain.

[0036] The instruction corresponding to the first control cycle in the preliminary optimized control sequence is extracted and used as the execution instruction for the current cycle. This rolling optimization process is repeated in the next control cycle.

[0037] Furthermore, the working principle of the improved model predictive control algorithm includes:

[0038] Within the rolling optimization framework of the standard model predictive control algorithm, a constraint boundary dynamic adjuster is embedded.

[0039] The constraint boundary dynamic adjuster receives the hydrogen storage status index in real time and has a built-in correlation function. The correlation function defines the mapping relationship between different numerical ranges of the hydrogen storage status index and the tightness of the hydrogen storage tank pressure constraint boundary.

[0040] In each optimization cycle, the constraint boundary dynamic adjuster calculates the temporary upper and lower limits of the pressure constraint based on the current hydrogen storage state index through the correlation function, and uses the temporary upper and lower limits to replace the original fixed boundary of the hydrogen storage tank pressure safety trajectory constraint in the optimization model.

[0041] After the boundary replacement is completed, the improved model predictive control algorithm continues to perform the prediction and optimization steps of the standard model predictive control algorithm;

[0042] After the optimization solution is completed, the constraint boundary dynamic adjuster restores the boundary of the hydrogen storage tank pressure safety trajectory constraint to the original fixed boundary, in preparation for the dynamic adjustment of the next cycle.

[0043] Furthermore, by solving the multi-timescale collaborative optimization model, an optimized control instruction set for multiple future control cycles is obtained, including:

[0044] The optimized control instruction set includes a target power instruction sequence for the electrolyzer, a target power instruction sequence for the fuel cell, and a target pressure regulation instruction for the hydrogen storage tank.

[0045] Run the solution process to obtain the global optimal solution or a satisfactory solution that meets the engineering requirements of the multi-timescale collaborative optimization model;

[0046] From the global optimal solution or satisfactory solution, time series variables spanning the entire prediction time domain are extracted. These time series variables include the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank.

[0047] Discretize the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank, and extract them according to the length of the control cycle to generate the target power command sequence of the electrolyzer, the target power command sequence of the fuel cell, and the target pressure adjustment command of the hydrogen storage tank that correspond one-to-one with each future control cycle.

[0048] The target power command sequence of the electrolyzer, the target power command sequence of the fuel cell, and the target pressure regulation command of the hydrogen storage tank are packaged in chronological order to form the optimized control command set.

[0049] Furthermore, the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank are discretized and sampled, and truncated according to the length of the control cycle, including:

[0050] Determine the basic control cycle duration of the combined power generation system;

[0051] Using the duration of the basic control cycle as an interval, the planned power curve of the electrolytic cell is sampled at equal intervals on the time axis. The power value of each sampling point is used as the set value of the electrolytic cell power command within the corresponding control cycle, thereby forming the target power command sequence of the electrolytic cell.

[0052] Using the same basic control cycle duration and sampling method, the planned power curve of the fuel cell and the planned pressure curve of the hydrogen storage tank are processed to form the target power command sequence of the fuel cell and the target pressure regulation command of the hydrogen storage tank, respectively.

[0053] Furthermore, the optimized control command set is issued to the corresponding electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve in the combined power generation system for execution, including:

[0054] From the optimized control instruction set, extract the instantaneous target power instruction for the electrolyzer, the instantaneous target power instruction for the fuel cell, and the instantaneous target pressure regulation instruction for the hydrogen storage tank corresponding to the current control cycle;

[0055] The instantaneous electrolytic cell target power command is encapsulated into a data frame conforming to the electrolytic cell controller communication protocol and sent to the electrolytic cell controller via fieldbus.

[0056] The instantaneous fuel cell target power command is encapsulated into a data frame conforming to the fuel cell controller communication protocol and sent to the fuel cell controller via fieldbus;

[0057] The instantaneous hydrogen storage tank target pressure adjustment command is converted into an analog voltage signal or a pulse width modulation signal and output to the drive circuit of the hydrogen storage tank pressure regulating valve.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] Uncertainty quantification was performed on the measured power sequences of wind and solar power to generate predicted power ranges for both. Status assessments were conducted on the operating data of electrolyzers, hydrogen storage tanks, and fuel cells to obtain the adjustability margin of the electrolyzer, hydrogen storage status indicators, and the adjustable power of the fuel cell. Based on these assessments, a multi-timescale collaborative optimization model was constructed. The power prediction range comprehensively characterizes the fluctuation range of renewable energy output, and the status indicators of each device intuitively reflect real-time adjustment capabilities. The integration of multi-dimensional information makes the model input more closely match the actual operating characteristics of the system, and the multi-timescale structure adapts to the dynamic changes required by different control cycles.

[0060] An improved model predictive control algorithm is employed to solve a multi-timescale collaborative optimization model. The constraint boundaries of the optimization problem are dynamically adjusted based on hydrogen storage state indicators, generating a set of optimized control instructions for multiple future control cycles. These instructions are then sent to the electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve for execution. The constraint boundaries adaptively change in real time with the hydrogen storage state, ensuring the optimization process matches the system's energy storage capacity. The multi-cycle instruction set enables continuous control actions, suppressing disturbances caused by fluctuations in wind and solar power output, and maintaining stable and coordinated operation of all system components. Attached Figure Description

[0061] Figure 1 This is a state diagram of the collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system as described in this invention.

[0062] Figure 2 A flowchart for the periodic acquisition of multi-source hydrogen energy data;

[0063] Figure 3 A flowchart for generating forecast intervals for wind and solar power. Detailed Implementation

[0064] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0065] See Figure 1 This invention provides a collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system, the overall implementation process of which is as follows:

[0066] From a combined power generation system consisting of a wind turbine, photovoltaic array, electrolyzer, hydrogen storage tank, and fuel cell, measured sequences of wind power, photovoltaic power, electrolyzer operating status signals, hydrogen storage tank pressure, and fuel cell power are periodically collected. Uncertainty quantification is performed on the measured wind power and photovoltaic power sequences to generate wind power prediction intervals and photovoltaic power prediction intervals. Simultaneously, state assessments are performed on the electrolyzer operating status signals, hydrogen storage tank pressure, and fuel cell power sequences to generate electrolyzer adjustability margin, hydrogen storage state index, and fuel cell adjustable power. Based on the wind power prediction interval, photovoltaic power prediction interval, electrolyzer adjustability margin, hydrogen storage state index, and fuel cell adjustable power, a multi-timescale collaborative optimization model is constructed. An improved model predictive control algorithm is used to solve this multi-timescale collaborative optimization model, where the improved model predictive control algorithm dynamically adjusts the constraint boundaries of the optimization problem based on the hydrogen storage state index. Solving the multi-timescale collaborative optimization model yields an optimized control instruction set for multiple future control cycles. The optimized control command set is sent to the corresponding electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve in the combined power generation system for execution.

[0067] In one embodiment of the present invention, see [reference] Figure 2 The system employs a wind turbine generator monitoring and data acquisition system to read rotor speed, pitch angle, and generator output electrical parameters at a fixed sampling frequency, calculating and generating a measured wind power sequence. Similarly, the photovoltaic array's combiner box and inverter monitoring unit reads DC-side current and voltage, and AC-side output electrical parameters at a fixed sampling frequency, calculating and generating a measured photovoltaic power sequence. The electrolyzer control cabinet's input / output modules read the electrolyzer's input current, input voltage, operating temperature, and alarm status words, combining them to generate an electrolyzer operating status signal. A pressure transmitter installed on the hydrogen storage tank reads and outputs the measured hydrogen storage tank pressure in real time. Finally, the fuel cell system's power management unit reads its net output electrical parameters at a fixed sampling frequency, calculating and generating a measured fuel cell power sequence.

[0068] In practical implementation, the wind turbine generator's monitoring and data acquisition system reads the rotor speed, pitch angle, and generator output electrical parameters at a fixed sampling frequency. These parameters include the instantaneous values ​​of the three-phase output voltage, output current, and active power. The following formula is used to calculate the actual output power of the wind turbine generator based on these measurements:

[0069]

[0070] in: This represents the average output power of a wind turbine within a sampling window. This indicates the total number of sampling points within the sampling window. , , They represent the first The instantaneous value of the three-phase output voltage corresponding to each sampling point , , They represent the first The instantaneous value of the three-phase output current corresponding to each sampling point; repeat the above calculation process at fixed time intervals, arrange the multiple average output powers obtained in time sequence, and generate a wind power measured sequence.

[0071] In some embodiments, the DC-side current and voltage and the AC-side output electrical parameters are read at a fixed sampling frequency through the combiner box and inverter monitoring unit of the photovoltaic array. The DC-side current and voltage include the instantaneous values ​​of the DC current and DC voltage at the output terminal of the photovoltaic array. The AC-side output electrical parameters include the instantaneous values ​​of the three-phase output voltage, the instantaneous output current, and the instantaneous total active power at the inverter grid connection point. The instantaneous DC output power of the photovoltaic array is obtained by multiplying the instantaneous DC current and the instantaneous DC voltage measured on the DC side. The instantaneous AC output power of the photovoltaic array is used as the instantaneous AC output power of the photovoltaic array. These values ​​are recorded at fixed time intervals and arranged in chronological order to generate a photovoltaic power measurement sequence.

[0072] In practical implementation, the input / output module of the electrolytic cell control cabinet reads the input current, input voltage, operating temperature, and alarm status word of the electrolytic cell. The input current is the measured current output of the electrolytic cell's DC power supply, the input voltage is the measured voltage output of the electrolytic cell's DC power supply, the operating temperature is the average of the temperature measurements at multiple points inside the electrolytic cell stack, and the alarm status word is the fault alarm bit field code summarized by the electrolytic cell control system. The input current, input voltage, operating temperature, and alarm status word are packaged into a set of structured data packets at fixed sampling times. Multiple sets of structured data packets are continuously collected and stored in chronological order, and combined to generate the electrolytic cell operating status signal.

[0073] It is understandable that a pressure transmitter installed on the hydrogen storage tank reads and outputs the measured pressure value of the hydrogen storage tank in real time. The pressure transmitter adopts the absolute pressure measurement principle, directly sensing the physical quantity change of the gas pressure inside the hydrogen storage tank and converting it into a standard industrial signal output. The standard industrial signal is quantized into a digital pressure value through an analog-to-digital conversion module. The digital pressure value is the measured pressure value of the hydrogen storage tank. The measured pressure value of the hydrogen storage tank is recorded and stored as time series data at fixed time intervals. Optionally, the power management unit of the fuel cell system reads its net output electrical parameters at a fixed sampling frequency. The net output electrical parameters include the instantaneous values ​​of the three-phase output voltage, output current, and net output active power at the fuel cell system grid connection point. The instantaneous value of net output active power is directly used as the instantaneous value of the fuel cell output power. The data is repeatedly collected at fixed time intervals and arranged in time sequence to generate a measured power sequence of the fuel cell.

[0074] In one embodiment of the present invention, see [reference] Figure 3 Empirical mode decomposition (EMD) was performed on the measured wind power sequence to obtain multiple wind electronic mode components reflecting fluctuation characteristics at different time scales. An autoregressive integral moving average (ARM) prediction model was established for each wind electronic mode component, and the prediction uncertainty band for each wind electronic mode component was calculated based on the prediction residual distribution of the ARM model. The prediction uncertainty bands of all wind electronic mode components were superimposed and reconstructed over time to form a total wind power prediction interval covering future periods. A similar decomposition, prediction, and uncertainty band calculation process was performed on the measured photovoltaic (PV) power sequence to generate the corresponding PV power prediction interval. The electrolyzer operating status signal was analyzed to obtain the current operating point, temperature, and historical operating time of the electrolyzer. Based on the electrolyzer performance degradation model, the maximum allowable power ramp-up rate and the maximum allowable power downsizing depth of the electrolyzer in the current and short future periods were calculated. The maximum allowable power ramp-up rate and the maximum allowable power downsizing depth jointly characterize the electrolyzer's adjustability margin. Based on the measured pressure values ​​and historical trends of the hydrogen storage tank, and considering the upper and lower limits of the tank's design pressure, the deviation of the current pressure from the median safe operating pressure and the trend coefficient based on the recent pressure change rate are calculated. A weighted combination of the pressure deviation and the pressure trend coefficient constitutes the hydrogen storage status index. The fluctuation characteristics of the measured power sequence of the fuel cell are analyzed. Combined with the fuel cell's rated power and current efficiency curve, the maximum and minimum power values ​​that can be instantly increased and decreased while maintaining high-efficiency operation are determined. These maximum and minimum power values ​​together define the adjustable power of the fuel cell.

[0075] In specific implementation, empirical mode decomposition is performed on the measured wind power sequence to decompose the non-stationary wind power measured sequence into a series of intrinsic mode function components with different time scale fluctuation characteristics and a residual component. These intrinsic mode function components are the wind electronic mode components. An autoregressive integral moving average prediction model is established for each wind electronic mode component. The model is trained using historical wind electronic mode component data to obtain the parameter estimates of the autoregressive integral moving average prediction model. Based on the probability distribution of the prediction residual sequence of the autoregressive integral moving average prediction model, the prediction uncertainty band of each wind electronic mode component is calculated based on the confidence level. The prediction uncertainty bands of all wind electronic mode components are superimposed in time series and added to the prediction values ​​of the residual component to reconstruct the total wind power prediction interval covering future periods. The same empirical mode decomposition, autoregressive integral moving average modeling, and uncertainty band calculation process is performed on the measured photovoltaic power sequence to generate the corresponding photovoltaic power prediction interval.

[0076] In some embodiments, the electrolyzer's operating status signal is analyzed to extract the current input current, input voltage, operating temperature, and historical cumulative operating hours. Based on a pre-established electrolyzer performance degradation model, the maximum allowable power ramp-up rate and the maximum allowable power reduction depth of the electrolyzer under the current operating temperature and aging level are calculated. The maximum allowable power ramp-up rate and the maximum allowable power reduction depth together characterize the electrolyzer's adjustability margin. Based on the historical time series of the measured pressure values ​​of the hydrogen storage tank, combined with the upper and lower design pressure limits of the hydrogen storage tank, the normalized deviation of the current pressure relative to the safe operating median value is calculated. A trend coefficient is calculated based on the pressure change rate of the most recent samples. The pressure deviation and the pressure trend coefficient are combined using a linear weighting method to form the hydrogen storage status index.

[0077] It is understood that by analyzing the statistical characteristics of the fluctuations in the measured power sequence of the fuel cell over a period of time, including the root mean square value of the power change and the maximum fluctuation amplitude, and combining the rated power of the fuel cell and the efficiency curve under the current load, the power range of the fuel cell in the high-efficiency operating range is identified. Within this range, the maximum power that the fuel cell can be adjusted up and the minimum power that can be adjusted down are determined. The maximum power that can be adjusted up and the minimum power that can be adjusted down together define the adjustable power of the fuel cell.

[0078] In practical implementation, for a given wind electronic mode component The calculation method for its prediction uncertainty bandwidth is as follows:

[0079]

[0080] in: Indicates the time of prediction wind electronic mode components The half-width of the predicted uncertainty band, This represents the inverse cumulative distribution function of the standard normal distribution. The significance level is used to control for the confidence level. Represents wind electronic mode components The autoregressive integral moving average prediction model at the prediction time The estimated standard deviation of the predicted residual at the given location.

[0081] In one embodiment of the present invention, an optimization objective function is established with the goal of minimizing the total operating cost of the combined power generation system. The objective function includes penalties for wind and solar curtailment, penalties for hydrogen supply and demand imbalance, and operating wear and tear costs of the electrolyzer and fuel cell. A constraint system is set, including: interval constraints requiring the actual output power of the wind turbine and photovoltaic array to fluctuate within their corresponding wind power prediction intervals and photovoltaic power prediction intervals; equipment constraints requiring the power adjustment of the electrolyzer to be within its adjustable margin; pressure changes in the hydrogen storage tank to meet pressure safety trajectory constraints dynamically defined by hydrogen storage status indicators; power constraints requiring the output power adjustment of the fuel cell to be within its adjustable power range; and power balance constraints of the system at multiple predefined time points. The optimization objective function and the constraint system together constitute a multi-timescale collaborative optimization model.

[0082] In specific implementation, an optimization objective function is established with the goal of minimizing the total operating cost of the combined power generation system. The optimization objective function is composed of a linear weighted sum of a wind curtailment penalty term, a hydrogen energy supply and demand imbalance penalty term, and an operating wear and tear cost term for the electrolyzer and fuel cell. The wind curtailment penalty term is calculated based on the power difference between the actual output power of the wind turbine and the lower limit of the wind power prediction range, and the power difference between the actual output power of the photovoltaic array and the lower limit of the photovoltaic power prediction range. The hydrogen energy supply and demand imbalance penalty term is calculated based on the degree to which the actual pressure of the hydrogen storage tank deviates from the expected pressure trajectory determined by the hydrogen energy dispatch plan. The operating wear and tear cost term for the electrolyzer and fuel cell is calculated by multiplying the absolute value of the power change of the electrolyzer and the absolute value of the power change of the fuel cell by the equivalent aging cost coefficient per unit power change.

[0083] In some embodiments, a constraint system is set, including interval constraints, equipment constraints, pressure safety trajectory constraints, power constraints, and power balance constraints. The interval constraints require that the actual output power of the wind turbine is not less than the lower limit and not greater than the upper limit of the wind power prediction interval, and the actual output power of the photovoltaic array is not less than the lower limit and not greater than the upper limit of the photovoltaic power prediction interval. The equipment constraints require that the power adjustment of the electrolyzer does not exceed the range of the maximum allowable power ramp-up rate and the maximum allowable power down depth defined by the electrolyzer's adjustability margin in any control cycle. The pressure safety trajectory constraints require that the measured or predicted pressure value of the hydrogen storage tank always lies between the upper and lower boundaries of the pressure safety trajectory dynamically defined by the hydrogen storage status index. The power constraints require that the output power adjustment of the fuel cell does not exceed the instantaneous power adjustment range defined by the adjustable power of the fuel cell in any control cycle. The power balance constraints require that at multiple predefined time points, the sum of the actual output power of the wind turbine, the actual output power of the photovoltaic array, and the actual output power of the fuel cell minus the actual power consumed by the electrolyzer equals the external load demand power.

[0084] In practical implementation, the process of minimizing the objective function can be mathematically expressed as finding a set of decision variables that minimizes the following expression:

[0085]

[0086] in: This represents the total operating cost of the combined power generation system over an optimization cycle. This represents the set of all discrete time steps included in the optimization cycle. , , These are the penalty coefficients for wind and solar power curtailment, hydrogen energy supply and demand imbalance, and operating wear and tear cost coefficients, respectively. Indicates at time step The sum of curtailed wind power and curtailed solar power. Indicates at time step The difference between the actual pressure and the expected pressure in the hydrogen storage tank. Indicates at time step The change in the power of the electrolytic cell compared to the previous time step. Indicates at time step The change in fuel cell power compared to the previous time step.

[0087] Table 1: Schematic diagram of power balance constraints at predefined time nodes in the multi-timescale collaborative optimization model

[0088]

[0089] As can be understood, Table 1 illustrates examples of power balance constraints at three different time points. The requirements are that the actual output power of the wind turbine plus the actual output power of the photovoltaic array plus the actual output power of the fuel cell minus the actual power consumed by the electrolyzer equals the load demand of 1600kW. The actual output power of the wind turbine must be between the upper and lower limits of the wind power prediction range at that time, and the actual output power of the photovoltaic array must be between the upper and lower limits of the photovoltaic power prediction range at that time. The power balance constraints at other time points follow the same principle.

[0090] In one embodiment of the present invention, at the beginning of each control cycle, an improved model predictive control algorithm is invoked. This algorithm uses the currently acquired system state as the initial point and performs a rolling solution of the multi-timescale co-optimization model within a preset prediction time domain. Before each rolling solution, the upper and lower boundaries of the pressure safety trajectory constraints of the hydrogen storage tank in the multi-timescale co-optimization model are dynamically calculated and updated based on the real-time updated values ​​of the hydrogen storage state indicators. When the hydrogen storage state indicators show that the pressure is too high, the upper pressure constraint is tightened; when the pressure is too low, the lower pressure constraint is tightened. After dynamically updating the constraint boundaries, an interior-point solver is used to solve the updated multi-timescale co-optimization model to obtain a preliminary optimized control sequence within the current prediction time domain. The instructions corresponding to the first control cycle in the preliminary optimized control sequence are extracted and used as the execution instructions for the current cycle, and this rolling optimization process is repeated in the next control cycle. A constraint boundary dynamic adjuster is embedded within the rolling optimization framework of the standard model predictive control algorithm. The constraint boundary dynamic adjuster receives hydrogen storage state indicators in real time and has a built-in correlation function that defines the mapping relationship between different numerical ranges of the hydrogen storage state indicators and the tightness of the hydrogen storage tank pressure constraint boundary. In each optimization cycle, the constraint boundary dynamic adjuster calculates the temporary upper and lower limits of the pressure constraint based on the current hydrogen storage state indicators using the correlation function, and replaces the original fixed boundary of the hydrogen storage tank pressure safety trajectory constraint in the optimization model with these temporary upper and lower limits. After the boundary replacement is completed, the improved model predictive control algorithm continues to execute the prediction and optimization steps of the standard model predictive control algorithm. After the optimization solution is completed, the constraint boundary dynamic adjuster restores the boundary of the hydrogen storage tank pressure safety trajectory constraint to the original fixed boundary, preparing for dynamic adjustment in the next cycle.

[0091] In practical implementation, at the beginning of each control cycle, the improved model predictive control algorithm is invoked. This algorithm uses the currently acquired system state as its initial point. The currently acquired system state includes the latest values ​​of the measured wind power sequence, the latest values ​​of the measured photovoltaic power sequence, the current readings of the electrolyzer operating status signal, the current readings of the measured hydrogen storage tank pressure, and the latest values ​​of the measured fuel cell power sequence. The multi-timescale collaborative optimization model is solved in a rolling manner within a preset prediction time domain. Before each rolling solution, the values ​​of the hydrogen storage state indicators are updated in real time. The upper and lower boundaries of the pressure safety trajectory constraints of the hydrogen storage tank in the multi-timescale collaborative optimization model are dynamically calculated and updated. When the hydrogen storage status index shows that the pressure is too high, the upper pressure constraint is tightened; when the pressure is too low, the lower pressure constraint is tightened. After dynamically updating the constraint boundaries, the interior point method solver is used to solve the updated multi-timescale collaborative optimization model to obtain the preliminary optimized control sequence in the current prediction time domain. The instruction corresponding to the first control cycle in the preliminary optimized control sequence is extracted as the execution instruction for the current cycle, and this rolling optimization process is repeated in the next control cycle.

[0092] In practical implementation, a constraint boundary dynamic adjuster is embedded within the rolling optimization framework of the standard model predictive control algorithm. This constraint boundary dynamic adjuster receives the hydrogen storage state index in real time and has a built-in correlation function. This correlation function defines the mapping relationship between different numerical ranges of the hydrogen storage state index and the tightness of the hydrogen storage tank pressure constraint boundary. In each optimization cycle, the constraint boundary dynamic adjuster calculates the temporary upper and lower limits of the pressure constraint based on the current hydrogen storage state index using the correlation function, and uses these temporary upper and lower limits to replace the original fixed boundary of the hydrogen storage tank pressure safety trajectory constraint in the optimization model. After the boundary replacement is completed, the improved model predictive control algorithm continues to execute the prediction and optimization steps of the standard model predictive control algorithm. After the optimization solution is completed, the constraint boundary dynamic adjuster restores the boundary of the hydrogen storage tank pressure safety trajectory constraint to the original fixed boundary, preparing for dynamic adjustment in the next cycle.

[0093] In some embodiments, the input to the correlation function is a hydrogen storage state index. The output is the temporary upper boundary offset of the hydrogen storage tank pressure safety trajectory constraint. and temporary lower boundary offset The calculation method is as follows:

[0094]

[0095] in: and It concerns hydrogen storage status indicators. The piecewise linear function, whose function parameters were determined through offline simulation and debugging, and the hydrogen storage state index A higher value indicates that the hydrogen storage tank pressure is closer to the high-risk area, while a lower value indicates that the hydrogen storage tank pressure is closer to the low-risk area; the temporary pressure upper boundary is equal to the original fixed upper boundary minus The temporary lower boundary pressure is equal to the original fixed lower boundary pressure plus... .

[0096] Table 2: Mapping Table of Hydrogen Storage Status Indicators Zoning and Constraint Boundary Adjustment

[0097]

[0098] As can be understood, referring to Table 2, which shows the constraint boundary adjustment strategies corresponding to the division of the hydrogen storage state index into five intervals, when the hydrogen storage state index falls in the interval [0.8, 1.0], it belongs to the extremely high pressure risk level. At this time, the temporary upper boundary of the hydrogen storage tank pressure safety trajectory constraint is lowered by 50% based on the original fixed upper boundary, and the temporary lower boundary is raised by 50% based on the original fixed lower boundary. When the hydrogen storage state index falls in the interval [0.0, 0.2), it belongs to the extremely low pressure risk level. At this time, the temporary upper boundary is raised by 40% based on the original fixed upper boundary, and the temporary lower boundary is lowered by 40% based on the original fixed lower boundary.

[0099] Optionally, when the interior point method solver processes the multi-timescale collaborative optimization model after dynamically updating the constraint boundaries, it substitutes the temporary upper pressure boundary and the temporary lower pressure boundary as the right-hand side of the inequality constraint into the solution to ensure that the optimization result satisfies the dynamically adjusted pressure safety trajectory constraint.

[0100] In one embodiment of the present invention, the optimized control instruction set includes a target power instruction sequence for the electrolyzer, a target power instruction sequence for the fuel cell, and a target pressure regulation instruction for the hydrogen storage tank. The solution process is run to obtain the global optimal solution or a satisfactory solution that meets engineering requirements from the multi-timescale collaborative optimization model. From the global optimal solution or the satisfactory solution, time-series variables spanning the entire prediction time domain are extracted. These time-series variables include the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank. The planned power curves of the electrolyzer, the fuel cell, and the hydrogen storage tank are discretized and sampled, and truncated according to the length of the control cycle to generate a target power instruction sequence for the electrolyzer, a target power instruction sequence for the fuel cell, and a target pressure regulation instruction for the hydrogen storage tank corresponding to each future control cycle. The target power instruction sequences for the electrolyzer, the fuel cell, and the target pressure regulation instructions for the hydrogen storage tank are packaged in chronological order to form the optimized control instruction set. The basic control cycle length of the combined power generation system is determined. Using the duration of the basic control cycle as the interval, the planned power curve of the electrolyzer is sampled at equal intervals along the time axis. The power value at each sampling point is used as the setpoint for the electrolyzer power command within the corresponding control cycle, thus forming the electrolyzer target power command sequence. Using the same basic control cycle duration and sampling method, the planned power curve of the fuel cell and the planned pressure curve of the hydrogen storage tank are processed to form the fuel cell target power command sequence and the hydrogen storage tank target pressure regulation command, respectively. From the optimized control command set, the immediate electrolyzer target power command, the immediate fuel cell target power command, and the immediate hydrogen storage tank target pressure regulation command corresponding to the current control cycle are extracted. The immediate electrolyzer target power command is encapsulated into a data frame conforming to the electrolyzer controller communication protocol and sent to the electrolyzer controller via the fieldbus. The immediate fuel cell target power command is encapsulated into a data frame conforming to the fuel cell controller communication protocol and sent to the fuel cell controller via the fieldbus. The immediate hydrogen storage tank target pressure regulation command is converted into an analog voltage signal or a pulse width modulation signal and output to the drive circuit of the hydrogen storage tank pressure regulating valve.

[0101] In specific implementation, the optimized control instruction set includes a target power instruction sequence for the electrolyzer, a target power instruction sequence for the fuel cell, and a target pressure regulation instruction for the hydrogen storage tank. The solution process is run to obtain the global optimal solution or a satisfactory solution that meets engineering requirements from the multi-timescale collaborative optimization model. From the global optimal solution or satisfactory solution, time-series variables spanning the entire prediction time domain are extracted. These time-series variables include the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank. The planned power curves of the electrolyzer, the fuel cell, and the hydrogen storage tank are discretized and sampled, and truncated according to the length of the control cycle to generate the target power instruction sequence for the electrolyzer, the target power instruction sequence for the fuel cell, and the target pressure regulation instruction for the hydrogen storage tank, each corresponding to a future control cycle. The target power instruction sequence for the electrolyzer, the target power instruction sequence for the fuel cell, and the target pressure regulation instruction for the hydrogen storage tank are packaged in chronological order to form the optimized control instruction set.

[0102] In some embodiments, the basic control cycle duration of the combined power generation system is determined, and the basic control cycle duration is set to five minutes. The planned power curve of the electrolyzer is sampled at equal intervals along the time axis, with the basic control cycle duration as the interval. The power value at each sampling point is used as the set value of the electrolyzer power command within the corresponding control cycle, thereby forming the electrolyzer target power command sequence. The planned power curve of the fuel cell and the planned pressure curve of the hydrogen storage tank are processed using the same basic control cycle duration and sampling method to form the fuel cell target power command sequence and the hydrogen storage tank target pressure regulation command, respectively.

[0103] In specific implementation, the immediate target power command for the electrolyzer, the immediate target power command for the fuel cell, and the immediate target pressure regulation command for the hydrogen storage tank corresponding to the current control cycle are extracted from the optimized control command set. The immediate target power command for the electrolyzer is encapsulated into a data frame conforming to the communication protocol of the electrolyzer controller and sent to the electrolyzer controller via the fieldbus. The immediate target power command for the fuel cell is encapsulated into a data frame conforming to the communication protocol of the fuel cell controller and sent to the fuel cell controller via the fieldbus. The immediate target pressure regulation command for the hydrogen storage tank is converted into an analog voltage signal or a pulse width modulation signal and output to the drive circuit of the pressure regulating valve of the hydrogen storage tank.

[0104] It is understandable that when discretizing the curve, the sampling time... The calculation method is as follows:

[0105]

[0106] in: Indicates the first The sampling time corresponding to each control cycle This indicates the start time of the current control cycle. Indicates the duration of the basic control cycle. Values , To predict the total number of control cycles contained in the time domain; at each sampling time... At this point, read the power value from the planned power curve of the electrolyzer. , take it as the first The target power command for the electrolyzer in each control cycle is used, and similarly, the power value on the planned power curve of the fuel cell is read. As the first The target power command for the fuel cell in each control cycle is used to read the pressure value from the planned pressure curve of the hydrogen storage tank. As the first The target pressure regulation command for the hydrogen storage tank in each control cycle.

[0107] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system, characterized in that, The method includes: From the combined power generation system consisting of wind turbine, photovoltaic array, electrolyzer, hydrogen storage tank and fuel cell, the measured power sequence of wind power, the measured power sequence of photovoltaic array, the working status signal of electrolyzer, the measured pressure value of hydrogen storage tank and the measured power sequence of fuel cell are periodically collected; Uncertainty quantification is performed on the measured wind power and photovoltaic power sequences to generate wind power prediction intervals and photovoltaic power prediction intervals. State assessment is performed on the electrolyzer working status signal, the measured value of hydrogen storage tank pressure and the measured fuel cell power sequence to generate the electrolyzer adjustability margin, hydrogen storage status index and fuel cell adjustability power. Based on the wind power prediction range, photovoltaic power prediction range, electrolyzer adjustable margin, hydrogen storage state index and fuel cell adjustable power, a multi-timescale collaborative optimization model is constructed. An improved model predictive control algorithm is used to solve the multi-timescale collaborative optimization model. The improved model predictive control algorithm dynamically adjusts the constraint boundary of the optimization problem based on the hydrogen storage state index. Solving the multi-timescale collaborative optimization model yields an optimized control instruction set for multiple future control cycles; The optimized control instruction set is sent to the corresponding electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve in the combined power generation system for execution.

2. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, The periodically acquired wind power measurement sequence, photovoltaic power measurement sequence, electrolyzer operating status signal, hydrogen storage tank pressure measurement value, and fuel cell power measurement sequence include: The wind turbine generator monitoring and data acquisition system reads the wind turbine speed, pitch angle and generator output electrical parameters at a fixed sampling frequency, and calculates and generates the measured wind power sequence. The photovoltaic array's combiner box and inverter monitoring unit read the DC side current and voltage and the AC side output electrical parameters at a fixed sampling frequency, and calculate and generate the photovoltaic power measured sequence. The input / output module of the electrolytic cell control cabinet reads the input current, input voltage, operating temperature, and alarm status words of the electrolytic cell, and combines them to generate the operating status signal of the electrolytic cell. The measured pressure value of the hydrogen storage tank is read and output in real time by a pressure transmitter installed on the hydrogen storage tank. The power management unit of the fuel cell system reads its net output electrical parameters at a fixed sampling frequency, and calculates and generates the measured power sequence of the fuel cell.

3. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, Uncertainty quantification is performed on the measured wind power and photovoltaic power sequences to generate wind power prediction intervals and photovoltaic power prediction intervals, including: Empirical mode decomposition was performed on the measured wind power sequence to obtain multiple wind electronic mode components that reflect the fluctuation characteristics at different time scales; An autoregressive integral moving average prediction model is established for each of the wind electronic mode components, and the prediction uncertainty band of each wind electronic mode component is calculated based on the prediction residual distribution of the autoregressive integral moving average prediction model. The prediction uncertainty of all wind electronic mode components is superimposed and reconstructed in time series to form a total wind power prediction range covering future periods; A similar decomposition, prediction, and uncertainty band calculation process is performed on the measured photovoltaic power sequence to generate the corresponding photovoltaic power prediction interval.

4. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, A status assessment is performed on the electrolyzer operating status signal, the measured pressure of the hydrogen storage tank, and the measured power sequence of the fuel cell to generate the electrolyzer adjustability margin, hydrogen storage status index, and fuel cell adjustability power, including: The electrolytic cell's operating status signal is analyzed to obtain the current operating point, temperature, and historical operating time of the electrolytic cell. Based on the electrolytic cell performance degradation model, the maximum allowable power ramp-up rate and the maximum allowable power reduction depth of the electrolytic cell in the current and short future time are calculated. The maximum allowable power ramp-up rate and the maximum allowable power reduction depth together characterize the adjustability margin of the electrolytic cell. Based on the measured pressure of the hydrogen storage tank and its historical trend, combined with the upper and lower limits of the design pressure of the hydrogen storage tank, the deviation of the current pressure from the median of safe operation and the trend coefficient based on the recent pressure change rate are calculated. The hydrogen storage status index is composed of the pressure deviation and the pressure trend coefficient by weighted combination. By analyzing the fluctuation characteristics of the measured power sequence of the fuel cell, and combining the rated power of the fuel cell with the current efficiency curve, the maximum power that the fuel cell can be adjusted up and the minimum power that can be adjusted down in real time under the premise of maintaining high efficiency operation are determined. The maximum power that can be adjusted up and the minimum power that can be adjusted down in real time together define the adjustable power of the fuel cell.

5. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, Based on the aforementioned wind power prediction range, photovoltaic power prediction range, electrolyzer adjustability margin, hydrogen storage state index, and fuel cell adjustability power, a multi-timescale collaborative optimization model is constructed, including: An optimization objective function is established with the goal of minimizing the total operating cost of the combined power generation system. The optimization objective function includes a wind curtailment penalty term, a hydrogen supply and demand imbalance penalty term, and an operating wear and tear cost term for the electrolyzer and fuel cell. A constraint system is established, which includes: range constraints that require the actual output power of wind turbines and photovoltaic arrays to fluctuate within their corresponding wind power prediction range and photovoltaic power prediction range; equipment constraints that require the power adjustment of the electrolyzer to be within its adjustable margin range; pressure changes in the hydrogen storage tank to meet the pressure safety trajectory constraints dynamically defined by the hydrogen storage status index; power constraints that require the output power adjustment of the fuel cell to be within its adjustable power range; and power balance constraints of the system at multiple predefined time points. The objective function and the constraint system together constitute the multi-timescale collaborative optimization model.

6. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, The improved model predictive control algorithm is used to solve the multi-timescale collaborative optimization model. The improved model predictive control algorithm dynamically adjusts the constraint boundaries of the optimization problem based on the hydrogen storage state index, including: At the beginning of each control cycle, the improved model predictive control algorithm is invoked. The improved model predictive control algorithm takes the currently collected system state as the initial point and performs rolling solution on the multi-time scale collaborative optimization model within the preset prediction time domain. Before each rolling solution, the upper and lower boundaries of the pressure safety trajectory constraint of the hydrogen storage tank in the multi-timescale collaborative optimization model are dynamically calculated and updated based on the real-time updated value of the hydrogen storage state index. When the hydrogen storage state index shows that the pressure is too high, the upper pressure constraint is tightened; when the pressure is too low, the lower pressure constraint is tightened. After dynamically updating the constraint boundaries, an interior point solver is used to solve the updated multi-timescale collaborative optimization model to obtain the preliminary optimized control sequence in the current prediction time domain. The instruction corresponding to the first control cycle in the preliminary optimized control sequence is extracted and used as the execution instruction for the current cycle. This rolling optimization process is repeated in the next control cycle.

7. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 6, characterized in that, The working principle of the improved model predictive control algorithm includes: Within the rolling optimization framework of the standard model predictive control algorithm, a constraint boundary dynamic adjuster is embedded. The constraint boundary dynamic adjuster receives the hydrogen storage status index in real time and has a built-in correlation function. The correlation function defines the mapping relationship between different numerical ranges of the hydrogen storage status index and the tightness of the hydrogen storage tank pressure constraint boundary. In each optimization cycle, the constraint boundary dynamic adjuster calculates the temporary upper and lower limits of the pressure constraint based on the current hydrogen storage state index through the correlation function, and uses the temporary upper and lower limits to replace the original fixed boundary of the hydrogen storage tank pressure safety trajectory constraint in the optimization model. After the boundary replacement is completed, the improved model predictive control algorithm continues to perform the prediction and optimization steps of the standard model predictive control algorithm; After the optimization solution is completed, the constraint boundary dynamic adjuster restores the boundary of the hydrogen storage tank pressure safety trajectory constraint to the original fixed boundary, in preparation for the dynamic adjustment of the next cycle.

8. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, Solving the multi-timescale collaborative optimization model yields an optimized control instruction set for multiple future control cycles, including: The optimized control instruction set includes a target power instruction sequence for the electrolyzer, a target power instruction sequence for the fuel cell, and a target pressure regulation instruction for the hydrogen storage tank. Run the solution process to obtain the global optimal solution or a satisfactory solution that meets the engineering requirements of the multi-timescale collaborative optimization model; From the global optimal solution or satisfactory solution, time series variables spanning the entire prediction time domain are extracted. These time series variables include the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank. Discretize the planned power curve of the electrolyzer, the planned power curve of the fuel cell, and the planned pressure curve of the hydrogen storage tank, and extract them according to the length of the control cycle to generate the target power command sequence of the electrolyzer, the target power command sequence of the fuel cell, and the target pressure adjustment command of the hydrogen storage tank that correspond one-to-one with each future control cycle. The target power command sequence of the electrolyzer, the target power command sequence of the fuel cell, and the target pressure regulation command of the hydrogen storage tank are packaged in chronological order to form the optimized control command set.

9. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 8, characterized in that, Discretize and sample the planned power curves of the electrolyzer, the fuel cell, and the hydrogen storage tank, and extract them according to the length of the control cycle, including: Determine the basic control cycle duration of the combined power generation system; Using the duration of the basic control cycle as an interval, the planned power curve of the electrolytic cell is sampled at equal intervals on the time axis. The power value of each sampling point is used as the set value of the electrolytic cell power command within the corresponding control cycle, thereby forming the target power command sequence of the electrolytic cell. Using the same basic control cycle duration and sampling method, the planned power curve of the fuel cell and the planned pressure curve of the hydrogen storage tank are processed to form the target power command sequence of the fuel cell and the target pressure regulation command of the hydrogen storage tank, respectively.

10. The collaborative optimization control method based on a wind-solar-hydrogen storage combined power generation system according to claim 1, characterized in that, The optimized control command set is issued to the corresponding electrolyzer controller, fuel cell controller, and hydrogen storage tank pressure regulating valve in the combined power generation system for execution, including: From the optimized control instruction set, extract the instantaneous target power instruction for the electrolyzer, the instantaneous target power instruction for the fuel cell, and the instantaneous target pressure regulation instruction for the hydrogen storage tank corresponding to the current control cycle; The instantaneous electrolytic cell target power command is encapsulated into a data frame conforming to the electrolytic cell controller communication protocol and sent to the electrolytic cell controller via fieldbus. The instantaneous fuel cell target power command is encapsulated into a data frame conforming to the fuel cell controller communication protocol and sent to the fuel cell controller via fieldbus; The instantaneous hydrogen storage tank target pressure adjustment command is converted into an analog voltage signal or a pulse width modulation signal and output to the drive circuit of the hydrogen storage tank pressure regulating valve.

Images