Multi-energy system optimization method and system based on hydrogen energy storage

Abstract

The invention discloses a multi-energy system optimization method and system based on hydrogen energy storage, and belongs to the field of multi-energy system optimization, and the method comprises the steps: obtaining the historical wind speed, wind direction and solar irradiance data of an anemometer tower and a photovoltaic power station, training a time sequence analysis model through the historical wind speed, wind direction and solar irradiance data, and obtaining a time sequence analysis model; wind energy power and solar energy power prediction curves in a future time period are obtained; the method comprises the following steps: acquiring pressure and temperature data of a hydrogen storage tank, inputting the pressure and temperature data into a pre-established leakage rate dynamic compensation algorithm to obtain a real-time hydrogen loss amount, and if the hydrogen loss amount is greater than a preset threshold value, reducing the operation pressure or temperature of the hydrogen storage tank; and according to the wind and light power prediction curve, the hydrogen production equipment operation data, the hydrogen storage tank operation parameters, the hydrogen conversion equipment control signal, the constraint model and the income cost model as input of a linear programming algorithm, an equipment capacity configuration scheme for maximizing the system economic benefits is obtained.

Description

Technical Field

[0001] This invention belongs to the field of multi-energy system optimization, and in particular relates to a method and system for optimizing multi-energy systems based on hydrogen energy storage. Background Technology

[0002] When constructing optimization methods for multi-energy systems based on hydrogen energy storage, the primary technical challenge is accurately matching the equipment capacity of the hydrogen energy storage system with the operational mode of the wind-solar-hydrogen energy storage station. Specifically, the equipment capacity of the hydrogen energy storage system directly affects the system's energy storage efficiency and response speed, while the operational mode of the wind-solar-hydrogen energy storage station determines the energy input and output rhythm. Improper matching between these two factors can lead to low system operating efficiency and even energy waste.

[0003] The design of hydrogen energy storage system capacity requires comprehensive consideration of various factors, including energy consumption and efficiency in hydrogen production, storage, conversion, and transportation. The technical parameters and operating status of each stage affect the overall system performance. For example, electrolysis efficiency during hydrogen production, leakage rate during storage, and energy loss during conversion all need to be precisely considered in the capacity design. Meanwhile, the operation mode of wind-solar-hydrogen energy storage stations involves effectively addressing the volatility and instability of wind and solar energy. How to efficiently store hydrogen when wind and solar resources are abundant and rationally release hydrogen energy when resources are scarce to maintain stable system operation is a complex technical challenge. This requires not only accurate predictive models to anticipate fluctuations in wind and solar resources but also flexible scheduling strategies to optimize the hydrogen storage and release process. Summary of the Invention

[0004] In view of this, the present invention provides a method and system for optimizing a multi-energy system based on hydrogen energy storage, which can accurately predict fluctuations in wind and solar resources using a predictive model, and optimize the storage and release process of hydrogen through scheduling strategies.

[0005] The technical solution for implementing the present invention is as follows:

[0006] An optimization method for multi-energy systems based on hydrogen energy storage includes:

[0007] Historical wind speed, wind direction, and solar irradiance data from wind measurement towers and photovoltaic power plants are obtained. Time series analysis models are trained using historical wind speed, wind direction, and solar irradiance data to obtain prediction curves for wind power and solar power in future periods.

[0008] The start-up and shutdown status of hydrogen production equipment at each moment is determined based on wind power and solar power prediction curves, and the operating data of hydrogen production equipment is obtained.

[0009] A dynamic compensation model for leakage rate is constructed based on the gradient boosting decision tree algorithm. After collecting the pressure and temperature data of the hydrogen storage tank, the data is input into the dynamic compensation model for leakage rate to obtain optimized operating parameters of the hydrogen storage tank.

[0010] Collect current, voltage, and hydrogen flow data of the hydrogen conversion equipment, plot the efficiency curve of the hydrogen conversion equipment based on the current, voltage, and hydrogen flow data, and input the hydrogen demand, system load data, and efficiency curve data into the dynamic programming model to obtain the control signal of the hydrogen conversion equipment.

[0011] Based on the wind power and solar power prediction curves, hydrogen production equipment operation data, hydrogen storage tank operation parameters, and hydrogen conversion equipment control signals as inputs to the linear programming algorithm, an equipment capacity configuration scheme that maximizes the system's economic benefits is generated.

[0012] Optimize multi-energy systems based on equipment capacity configuration schemes that maximize system economic benefits.

[0013] Preferably, the process of obtaining the wind power and solar power prediction curves for future time periods includes:

[0014] Acquire historical wind speed data, historical wind direction data, and historical solar irradiance data recorded by the wind measurement tower and the photovoltaic power station. Perform data quality checks on the historical wind speed data, historical wind direction data, and historical solar irradiance data to determine if there are missing values. If there are missing values, use linear interpolation to fill in the missing values.

[0015] Outlier detection was performed on the historical wind speed data, historical wind direction data, and historical solar irradiance data after the filling process was completed. The box plot method was used to determine whether there were outliers in the data. If outliers were found, the Laida criterion was used to remove them and obtain valid data.

[0016] Based on the geographical location information of the wind measurement towers and the photovoltaic power stations, spatial matching is performed on the valid data. The data of the wind measurement towers with the closest spatial distance are matched with the data of the photovoltaic power stations to obtain the matched dataset.

[0017] The matched dataset is time-aligned to unify the data time resolution and obtain an aligned dataset. A long short-term memory network is then used to extract features from the aligned dataset to obtain a multi-dimensional feature vector.

[0018] Based on the multidimensional feature vector, wind power prediction model and solar power prediction model are constructed, and wind power and solar power prediction curves for future periods are obtained based on the wind power prediction model and solar power prediction model.

[0019] Preferably, the process of determining the start-up and shutdown status of the hydrogen production equipment at each moment based on the predicted wind and solar power curves for the future time period, and obtaining the operating data of the hydrogen production equipment, includes:

[0020] Based on the wind power and solar power prediction curves for the future period, wind power data and solar power data for each moment in the prediction time series are obtained.

[0021] Based on the preset electrolyzer start-up threshold, the wind power data and solar power data at each time point are compared with the electrolyzer start-up threshold to obtain the start-up and shutdown status of the hydrogen production equipment at each time point.

[0022] Based on the start-up and shutdown status of the hydrogen production equipment at each time point, the rated power of the electrolyzer is obtained, and the operating status of the electrolyzer at each time point is determined.

[0023] By using the operating status of the electrolyzer at each time point, and employing the random forest algorithm in conjunction with wind power data and solar power data, the predicted power of the electrolyzer at each time point is obtained.

[0024] Based on the predicted power of the electrolyzer at each time point and combined with the rated power of the electrolyzer, the minimum value between the two is taken to obtain the operating data of the hydrogen production equipment at each time point.

[0025] Preferably, the process of obtaining optimized hydrogen storage tank operating parameters includes:

[0026] The data acquisition unit obtains the pressure value from the pressure sensor and the temperature value from the temperature sensor of the hydrogen storage tank to obtain the real-time pressure value and real-time temperature value.

[0027] The real-time pressure and temperature values ​​are input into the leakage rate dynamic compensation model based on the gradient boosting decision tree algorithm to obtain the hydrogen loss compensation value.

[0028] The initial value of hydrogen loss is calculated based on the real-time pressure value and the real-time temperature value, and the amount of hydrogen loss is obtained based on the initial value of hydrogen loss and the hydrogen loss compensation value.

[0029] A hydrogen loss threshold is set, and the amount of hydrogen loss is compared with the hydrogen loss threshold to obtain a comparison result. Based on the comparison result, the operating parameters of the hydrogen storage tank are optimized to obtain optimized operating parameters of the hydrogen storage tank.

[0030] Preferably, the process of obtaining the control signal for the hydrogen conversion device includes:

[0031] Collect operating data of the hydrogen conversion equipment, including current intensity, voltage values, and hydrogen flow rate data;

[0032] Based on the collected current intensity, voltage values ​​and hydrogen flow data, the conversion efficiency of the equipment under different operating conditions was calculated, and the efficiency curve was plotted.

[0033] Obtain hydrogen demand data and system load data, and use the hydrogen demand data, system load data and efficiency curve data as input to the dynamic programming model to obtain the optimal hydrogen conversion timing.

[0034] The control signal for the hydrogen conversion device is obtained based on the optimal hydrogen conversion timing.

[0035] Preferably, the process of obtaining the optimal hydrogen conversion timing includes:

[0036] Historical hydrogen demand data, system load data, and equipment efficiency curve data are acquired and preprocessed using data cleaning methods to obtain preprocessed hydrogen demand data, preprocessed system load data, and preprocessed equipment efficiency curve data.

[0037] Based on the preprocessed hydrogen demand data and preprocessed system load data, a time series forecasting algorithm is used to obtain the predicted hydrogen demand data and the predicted system load data for future time periods.

[0038] Based on the hydrogen demand forecast data and the system load forecast data at future time, a linear regression algorithm is used to obtain the efficiency values ​​of the equipment under different loads.

[0039] The efficiency values ​​of the equipment under different loads and the predicted hydrogen demand data at future times are calculated to obtain the optimal operating power of the hydrogen production equipment, the capacity of the hydrogen storage equipment, and the power of the hydrogen release equipment.

[0040] The operating power, hydrogen storage capacity, and hydrogen release power of the optimal hydrogen production equipment are compared with the real-time operating power, hydrogen storage capacity, and hydrogen release power of the hydrogen production equipment to obtain the comparison results.

[0041] Based on the comparison results, the hydrogen production equipment is adjusted, and the updated hydrogen capacity in the hydrogen storage equipment is obtained based on the adjusted equipment parameter data. Based on the updated hydrogen capacity in the hydrogen storage equipment, the optimal hydrogen conversion timing is obtained.

[0042] Preferably, the process of generating a device capacity configuration scheme that maximizes the system's economic benefits includes:

[0043] A set of hydrogen production equipment operation information is obtained based on the wind power and solar power prediction curves for the future time period.

[0044] Based on the hydrogen production of each electrolyzer in the hydrogen production equipment operation information set, and combined with the minimum and maximum allowable hydrogen storage capacity of the hydrogen storage tank, a set of parameters for the hydrogen storage tank is obtained.

[0045] Based on the hydrogen storage capacity of each hydrogen storage tank in the parameter set, the start-up and shutdown thresholds of the hydrogen fuel cell are set, and a set of control commands for the hydrogen conversion device is generated.

[0046] Based on the set of parameters of the hydrogen storage tank and the set of operating information of the hydrogen production equipment, a constraint model is constructed that includes the upper and lower limits of equipment power and the capacity limit of the hydrogen storage tank. Based on the constraint model, the operating constraints of the equipment are obtained.

[0047] Based on the equipment operation constraints, a revenue-cost model is established by integrating the power generation revenue per unit time and the hydrogen production cost per unit time, and the net system revenue per unit time is obtained based on the revenue-cost model.

[0048] Based on the linear programming algorithm, the operating constraints of the equipment and the net system revenue per unit time are calculated to obtain the equipment capacity configuration scheme that maximizes the total net system revenue within a set time.

[0049] On the other hand, the present invention also provides a multi-energy system optimization system based on hydrogen energy storage, comprising:

[0050] The data acquisition module is used to acquire historical wind speed, wind direction, and solar irradiance data of the wind measurement tower and photovoltaic power station. The time series analysis model is trained using the historical wind speed, wind direction, and solar irradiance data to obtain the wind power and solar power prediction curves for future periods.

[0051] The time series analysis module is used to determine the start-up and shutdown status of the hydrogen production equipment at each moment based on the wind power and solar power prediction curves for the future period, and to obtain the operating data of the hydrogen production equipment.

[0052] The hydrogen loss monitoring module is used to build a dynamic compensation model for leakage rate based on the gradient boosting decision tree algorithm. After collecting the pressure and temperature data of the hydrogen storage tank, it is input into the dynamic compensation model for leakage rate to obtain optimized operating parameters of the hydrogen storage tank.

[0053] The hydrogen conversion optimization module is used to collect current, voltage and hydrogen flow data of the hydrogen conversion equipment, and plot the efficiency curve of the hydrogen conversion equipment based on the current, voltage and hydrogen flow data. Combined with hydrogen demand and system load data, the hydrogen demand, system load and efficiency curve data are input into the dynamic programming model to obtain the control signal of the hydrogen conversion equipment.

[0054] The system economic benefit optimization module is used to generate an equipment capacity configuration scheme that maximizes the system economic benefits based on the wind power and solar power prediction curves for the future period, the operating data of the hydrogen production equipment, the optimized operating parameters of the hydrogen storage tank, and the control signals of the hydrogen conversion equipment as inputs to a linear programming algorithm.

[0055] The simulation verification module is used to optimize the multi-energy system based on the equipment capacity configuration scheme that maximizes the system's economic benefits.

[0056] On the other hand, the present invention also provides an electronic device including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computer program.

[0057] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0058] Beneficial effects:

[0059] 1. This invention can accurately predict fluctuations in wind and solar resources using a predictive model, and optimize the storage and release process of hydrogen through scheduling strategies.

[0060] 2. This invention predicts wind and solar power output through a time series analysis model and controls the start and stop of the electrolyzer accordingly, taking the minimum value between wind and solar power and the rated power of the electrolyzer as the actual operating power.

[0061] 3. This invention utilizes a dynamic compensation algorithm for leakage rate to monitor the hydrogen loss in the hydrogen storage tank in real time, and adjusts the operating pressure or temperature when the loss is too large.

[0062] 4. This invention combines the efficiency curve of the hydrogen conversion equipment, the hydrogen demand, and the system load data, and uses a dynamic programming model to determine the optimal timing and conversion rate of hydrogen conversion.

[0063] 5. The present invention inputs the above results together with the constraint model and the revenue-cost model into a linear programming algorithm to obtain an equipment capacity configuration scheme that maximizes the economic benefits of the system, and verifies the effectiveness and stability of the scheme through a simulation platform.

[0064] 6. The comprehensive application of a series of measures in this invention enables the safe, efficient, and economical operation of the wind and solar hydrogen production, storage, and utilization system. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0066] Figure 2This is a schematic diagram of the system structure according to an embodiment of the present invention. Detailed Implementation

[0067] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0068] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0069] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0070] Example 1

[0071] like Figure 1-2 As shown, this embodiment provides a multi-energy system optimization method based on hydrogen energy storage, including:

[0072] Step S101: Obtain historical wind speed, wind direction, and solar irradiance data from the wind measurement tower and photovoltaic power station. Train a time series analysis model using the historical wind speed, wind direction, and solar irradiance data to obtain prediction curves for wind power and solar power in future periods.

[0073] Historical wind speed and direction data recorded by meteorological towers and historical solar irradiance data recorded by photovoltaic power plants were acquired. Data quality checks were performed on these data to determine if missing values ​​were present. If missing values ​​were found, linear interpolation was used to impute them. Outlier detection was then performed on the imputed historical wind speed, wind direction, and solar irradiance data using box plots. If outliers were found, the Laida criterion was used to remove them, resulting in valid data. Spatial matching was performed on the valid data based on the geographical locations of the meteorological towers and photovoltaic power plants, matching the data from the nearest meteorological towers with the data from the photovoltaic power plants to obtain a matched dataset. The matched dataset was then time-aligned to unify the temporal resolution, resulting in an aligned dataset. A Long Short-Term Memory (LSTM) network was used to extract features from the aligned dataset, yielding a multidimensional feature vector. Based on multidimensional feature vectors, wind power prediction models and solar power prediction models are constructed. The inputs to both models are multidimensional feature vectors, and the outputs are predicted power values. The wind power prediction model is trained using a support vector regression algorithm, resulting in a trained wind power prediction model. The solar power prediction model is trained using a gated recurrent unit algorithm, resulting in a trained solar power prediction model. Using both trained models, wind power and solar power predictions are used to predict future wind and solar power, yielding predicted curves for those future periods.

[0074] Specifically, historical data is acquired. For example, in a certain region from January to December 2022, wind measurement tower A recorded wind speed and direction data from 8:00 AM to 6:00 PM daily, while photovoltaic power station B recorded solar irradiance data for the same period. Data quality checks first assess data completeness. For instance, wind speed data for wind measurement tower A at 9:00 AM on March 5, 2022, and solar irradiance data for photovoltaic power station B at 2:00 PM on June 10, 2022, are missing. For the missing data, linear interpolation is used to fill in the gaps. For example, if the wind speed at 8:00 AM on March 5 was 5 m / s and at 10:00 AM it was 7 m / s, then linear interpolation can be used to estimate the wind speed at 9:00 AM as 6 m / s. Similarly, the missing value at 2:00 PM can be estimated using solar irradiance data from 1:00 PM and 3:00 PM on June 10. Data quality checks ensure data continuity and completeness, providing a reliable data foundation for subsequent analysis. Outlier detection is performed using box plots. For example, for wind speed data from January 2022, its quartiles are calculated, and a box plot is drawn based on the interquartile range. If a wind speed value exceeds the upper or lower limit of the box plot, it is identified as an outlier. For outliers, the Laida criterion is used for further evaluation. For example, the mean and standard deviation of the January wind speed data are calculated; if the difference between a wind speed value and the mean exceeds three times the standard deviation, it is identified as an outlier and removed. Outlier detection and handling can eliminate the impact of extreme data on model training, improving the model's accuracy and robustness. Spatial matching is performed based on the geographical location information of the meteorological towers and photovoltaic power stations, such as latitude and longitude coordinates, to calculate the spatial distance between them. For example, if the coordinates of meteorological tower A are (116.4 degrees east longitude, 39.9 degrees north latitude) and the coordinates of photovoltaic power station B are (116.5 degrees east longitude, 40.0 degrees north latitude), the distance between them can be calculated. The data from the nearest meteorological tower and photovoltaic power station are then matched. Spatial matching considers the spatial correlation of wind and solar energy resources, selecting the spatially closest meteorological towers and photovoltaic power plants to improve the accuracy of prediction models. Time alignment unifies the data's time resolution. For example, if the time resolution of meteorological tower data is 1 hour and the time resolution of photovoltaic power plant data is 15 minutes, the photovoltaic power plant data can be averaged or interpolated to unify their time resolution to 1 hour. Time alignment ensures the consistency of wind speed, wind direction, and solar irradiance data over time, facilitating subsequent feature extraction and model training. Feature extraction employs a Long Short-Term Memory (LSTM) network. For example, using aligned wind speed, wind direction, and solar irradiance data as input, the LTM network can automatically learn the temporal dependencies in the data and extract multi-dimensional feature vectors. For instance, the extracted feature vectors might contain information such as the average wind speed over the past 24 hours, the trend of wind direction changes, and the peak value of solar irradiance.Long Short-Term Memory (LSTM) networks can capture long-term and short-term dependencies in time-series data, and the extracted feature vectors can better characterize the dynamic changes in wind and solar energy resources. Predictive models are constructed using Support Vector Regression (SVR) and Gated Recurrent Unit (GRU) algorithms, respectively. SVR is suitable for wind power prediction because it can handle nonlinear relationships and is robust to noisy data. GRU is suitable for solar power prediction because it can better capture the periodic and trend changes in solar irradiance data. For example, using the extracted multidimensional feature vectors as input, the SVR model can output the predicted wind power for the next hour, and the GRU model can output the predicted solar power for the next hour. The trained models can predict future wind and solar power based on the input feature vectors. Prediction and error analysis are performed, using the trained models to predict future time periods and calculating the root mean square error (RMSE) and mean absolute error (MAE) of the prediction results.

[0075] Step S102: Based on the wind power and solar power prediction curves, determine the start-up and shutdown of the hydrogen production equipment. If the wind power or solar power is greater than the preset electrolyzer start-up threshold, start the hydrogen production equipment; otherwise, stop it. The actual operating power of the electrolyzer is the minimum of the predicted wind power or solar power and the rated power of the electrolyzer.

[0076] Based on the wind and solar power prediction curves, wind and solar power data for each moment within the prediction time series are obtained. According to a preset electrolyzer start-up threshold, the relationship between the wind and solar power data at each moment and the electrolyzer start-up threshold is compared to determine whether the start-up conditions are met at each moment. If the wind or solar power is greater than the electrolyzer start-up threshold, the hydrogen production equipment is started at that moment; otherwise, it is stopped. Based on the start-up and shutdown status of the hydrogen production equipment at each moment, the rated power of the electrolyzer is obtained, determining the operating status of the electrolyzer at each moment. Using the operating status of the electrolyzer at each moment, a random forest algorithm is employed, combined with the wind and solar power data, to obtain the predicted power of the electrolyzer at each moment. Based on the predicted power of the electrolyzer at each moment, combined with the rated power of the electrolyzer, the minimum value between the two is taken to obtain the actual operating power of the electrolyzer at each moment.

[0077] This also includes: using a Long Short-Term Memory (LSTM) network algorithm to predict the health status of hydrogen production equipment, thus obtaining the equipment's health status. Based on the equipment's health status, a Support Vector Machine (SVM) algorithm is used to predict equipment failures, resulting in the predicted failure outcomes.

[0078] Specifically, based on wind and solar power power forecasts, the power output of wind and solar power at each moment within a future period can be obtained. For example, predicting wind and solar power output over the next 24 hours yields a time series containing 24 data points, each representing an hourly wind or solar power value. For instance, the first hour might have a wind power output of 500 kW and a solar power output of 300 kW; the second hour might have a wind power output of 600 kW and a solar power output of 200 kW, and so on. An electrolyzer is a device that uses electricity to decompose water into hydrogen and oxygen, requiring a certain power input to start. A pre-set electrolyzer start-up threshold, such as 400 kW, is used. The wind and solar power output data at each moment is compared to this threshold. If the wind or solar power output at a given moment exceeds 400 kW, the start-up conditions are considered met, and the hydrogen production equipment can be started to electrolyze water to produce hydrogen. For example, if the wind power in the first hour is 500 kW (greater than 400 kW), meeting the startup conditions, the hydrogen production equipment is started. In the second hour, if the solar power is 200 kW (less than 400 kW), and although the wind power is 600 kW, it may not start due to control strategies or economic considerations, then the hydrogen production equipment is stopped at that time. Based on the start / stop status of the hydrogen production equipment at each moment, combined with the rated power of the electrolyzer (e.g., 1000 kW), the operating status of the electrolyzer at that moment can be determined. For example, if the hydrogen production equipment starts at a certain moment, the electrolyzer is in an operating state; if the hydrogen production equipment stops, the electrolyzer is in a stopped state. Here, the random forest algorithm is used, an ensemble learning method that makes predictions by constructing multiple decision trees and combining their predictions. Wind power data, solar power data, and the operating status of the electrolyzer are used as input features to train the random forest model, predicting the power of the electrolyzer at each moment. For example, the model might predict that the power of the electrolyzer at a certain moment is 700 kW. After obtaining the predicted power of the electrolyzer, it is compared with the rated power of the electrolyzer, and the minimum of the two is taken as the actual operating power of the electrolyzer at that moment. For example, if the predicted power is 700 kW and the rated power is 1000 kW, then the actual operating power is 700 kW; if the predicted power is 1200 kW, then the actual operating power is 1000 kW, meaning the electrolyzer will not be overloaded. This ensures that the electrolyzer operates within a safe range, avoiding equipment damage. Long Short-Term Memory (LSTM) networks are a special type of recurrent neural network that excels at processing time-series data. Using the actual operating power of the electrolyzer at each moment as input, the LSTM model is used to predict the health status of the hydrogen production equipment.

[0079] Step S103: Collect pressure and temperature data of the hydrogen storage tank, input the pressure and temperature data into a pre-established dynamic compensation algorithm for leakage rate, and obtain the real-time hydrogen loss. If the hydrogen loss is greater than a preset threshold, reduce the operating pressure or temperature of the hydrogen storage tank.

[0080] The data acquisition unit obtains pressure values ​​from the pressure sensor and temperature values ​​from the temperature sensor of the hydrogen storage tank, obtaining real-time pressure and temperature values. These real-time pressure and temperature values ​​are input into a pre-established dynamic leakage rate compensation model based on a gradient boosting decision tree algorithm to obtain a hydrogen loss compensation value. An initial hydrogen loss value is calculated based on the real-time pressure and temperature values. The hydrogen loss amount is obtained by subtracting the hydrogen loss compensation value from the initial hydrogen loss value. A pre-set hydrogen loss threshold is used to determine the relationship between the hydrogen loss amount and the threshold. If the hydrogen loss amount exceeds the threshold, an adjustment signal is output. This adjustment signal is input to the hydrogen storage tank control system to determine the adjustment range of the operating pressure. Based on the adjustment range, the reduction value of the operating pressure is calculated. This reduction value is input to the pressure control unit of the hydrogen storage tank to determine the new operating pressure, completing the reduction operation of the hydrogen storage tank's operating pressure or temperature.

[0081] Specifically, the data acquisition unit continuously monitors and acquires real-time pressure and temperature values ​​inside the hydrogen storage tank using pressure and temperature sensors installed on the tank. For example, the pressure sensor might detect a current pressure of 35 MPa and the temperature sensor a current temperature of 25 degrees Celsius. These two values ​​form the basis for subsequent calculations. The acquired real-time pressure and temperature values ​​are input into a pre-built dynamic leakage rate compensation model based on a gradient boosting decision tree algorithm. Gradient boosting decision trees are an ensemble learning method that improves prediction accuracy by combining multiple decision trees. In this model, pressure and temperature are key input features. The model is trained based on historical data and can predict the hydrogen loss compensation value under current pressure and temperature conditions due to factors such as material aging and decreased sealing performance. For example, the model might predict a hydrogen loss compensation value of 0.002 kg per hour based on the current pressure of 35 MPa and temperature of 25 degrees Celsius, combined with historical data training results. Based on the real-time pressure and temperature values, the initial hydrogen loss value can be calculated. For example, using a simplified application of the ideal gas law, and combining parameters such as current pressure and temperature, as well as the volume of the hydrogen storage tank, the change in hydrogen mass due to temperature and pressure variations can be estimated, for example, a decrease of 0.01 kg per hour. Then, the initial hydrogen loss is subtracted from the hydrogen loss compensation value to obtain the actual hydrogen loss. For example, 0.01 kg minus 0.002 kg yields a hydrogen loss of 0.008 kg per hour. A pre-set hydrogen loss threshold is used to determine if the current hydrogen loss is within an acceptable range. For example, the hydrogen loss threshold is set to 0.005 kg per hour. The calculated hydrogen loss is compared with this threshold. If the hydrogen loss exceeds the threshold, an anomaly is considered, requiring adjustment. For example, if the current hydrogen loss is 0.008 kg per hour, exceeding the set threshold of 0.005 kg, the system determines that adjustment is needed and outputs an adjustment signal. The adjustment signal is sent to the hydrogen storage tank's control system. The control system determines the adjustment range of the operating pressure according to preset rules and algorithms. For example, the control system can set a safe pressure reduction range, such as 0.1 MPa to 0.5 MPa, based on the current hydrogen consumption rate and the performance parameters of the hydrogen storage tank. Based on the determined operating pressure adjustment range, the specific operating pressure reduction value is further calculated. For instance, the control system decides to reduce the operating pressure by 0.2 MPa based on the magnitude and rate of hydrogen consumption. This reduction value is sent to the pressure control unit of the hydrogen storage tank. Upon receiving the operating pressure reduction value, the pressure control unit precisely controls the pressure relief valve or booster pump of the hydrogen storage tank to adjust the operating pressure of the tank. For example, the pressure control unit reduces the pressure of the hydrogen storage tank from 35 MPa to 34.8 MPa by controlling the opening degree of the pressure relief valve.

[0082] Step S104: Collect current, voltage, and hydrogen flow data of the hydrogen conversion equipment. Based on the current, voltage, and hydrogen flow data, plot the efficiency curve of the hydrogen conversion equipment. Combine the hydrogen demand and system load data, input the hydrogen demand, system load, and efficiency curve data into the dynamic programming model to obtain the optimal timing and conversion amount of hydrogen conversion.

[0083] Data on the operation of the hydrogen conversion equipment is collected, including current intensity, voltage values, and hydrogen flow rate. Based on the collected data, the conversion efficiency of the equipment under different operating conditions is calculated, and efficiency curves are plotted. Hydrogen demand data and system load data are obtained and used as input to a dynamic programming model. The dynamic programming model calculates the optimal hydrogen conversion timing based on the input hydrogen demand data, system load data, and efficiency curve data. Based on the optimal hydrogen conversion timing and hydrogen demand data, the optimal number of hydrogen conversions is calculated. If multiple hydrogen conversion devices exist, the optimal operating range for each device is obtained using a linear regression algorithm based on its efficiency curve. Based on the optimal operating range and the calculated optimal number of hydrogen conversions, a support vector machine algorithm is used to obtain the hydrogen allocation quantity for each device.

[0084] Data collection for hydrogen conversion equipment operation is a crucial aspect of energy management systems. By collecting data such as current intensity, voltage values, and hydrogen flow rate, a comprehensive understanding of the equipment's operating status can be obtained. For example, during operation, a water electrolysis hydrogen production unit maintains a stable current intensity of 50 amperes, a voltage of 220 volts, and a hydrogen flow rate of 10 standard cubic meters per hour. This data provides the foundation for subsequent efficiency analysis. Based on the collected data, the conversion efficiency of the equipment under different operating conditions can be calculated. For instance, when the current intensity increases to 60 amperes, the voltage and hydrogen flow rate also change accordingly. By comparing the input power and output hydrogen energy, the conversion efficiency under this condition can be calculated. Plotting the efficiency data under different operating conditions into a graph yields an efficiency curve. This curve visually illustrates the equipment's performance under different loads, providing a basis for optimizing operating strategies. For example, it was found that the equipment achieves its highest efficiency (85%) at a current intensity of 55 amperes, while efficiency decreases at lower or higher current intensities. Obtaining hydrogen demand data and system load data is a prerequisite for dynamic programming. For example, an industrial park's hydrogen demand for the next 24 hours is 1000 standard cubic meters, and the power system load data has also been obtained. These data, along with efficiency curves, are input into a dynamic programming model. The model will comprehensively consider various factors to find the optimal time for hydrogen conversion. The core of the dynamic programming model is finding the optimal solution. For instance, the model calculates that although hydrogen conversion occurs slightly during periods of low power load, the lower electricity price results in a better overall cost. Therefore, the model determines that hydrogen conversion is most economical between 2 AM and 4 AM. The model further calculates that, to meet demand and considering losses, the optimal amount of hydrogen to be converted is 1050 standard cubic meters. When multiple hydrogen conversion devices exist, meticulous management of each device is required. For example, the system has two water electrolysis hydrogen production devices with different efficiency curves. Using a linear regression algorithm to analyze the efficiency curve of each device, it can be found that device A is most efficient between 40 and 50 amperes, while device B performs best between 50 and 60 amperes. This yields the optimal operating range for each device. Based on the optimal operating range and the total number of hydrogen conversions, the hydrogen allocation for each device can be further optimized using the support vector machine algorithm.

[0085] Step S105: Based on the wind and solar power prediction curves, hydrogen production equipment operation data, hydrogen storage tank operation parameters, hydrogen conversion equipment control signals, constraint models, and revenue-cost models as inputs to the linear programming algorithm, a capacity configuration scheme that maximizes the system's economic benefits is obtained.

[0086] Based on the wind and solar power prediction curve collector, predictive data on the time-varying power of wind and solar power are collected to construct a time series prediction model, resulting in a set of wind and solar power prediction curves for future time periods. Based on this set, the input power range of the electrolyzers is pre-defined, and the corresponding hydrogen production rate and hydrogen output are derived by combining the input power of different electrolyzers, thus obtaining a set of hydrogen production equipment operation information. Based on the hydrogen output of each electrolyzer in the hydrogen production equipment operation information set, combined with the minimum and maximum allowable hydrogen storage capacity of the hydrogen storage tanks, a hydrogen storage tank capacity range is pre-established, resulting in a set of hydrogen storage tank parameters. Based on the hydrogen storage capacity of each hydrogen storage tank in the parameter set, start-up and shutdown thresholds for the hydrogen fuel cells are set, generating a set of control commands for the hydrogen conversion equipment. Based on the set of control commands for the hydrogen conversion equipment, and combined with the wind and solar power prediction curves, a constraint model including upper and lower limits of equipment power and hydrogen storage tank capacity limitations is pre-established, resulting in equipment operation constraints. Based on the equipment operation constraints, a revenue-cost model is pre-established by integrating the power generation revenue per unit time and the hydrogen production cost per unit time, resulting in the system's net revenue per unit time. Based on the system's net revenue per unit time, a linear programming algorithm is applied to the revenue-cost model and constraint model to obtain a device capacity configuration scheme that maximizes the system's total net revenue within a set time period.

[0087] The wind and solar power prediction curve acquisition device monitors the operating status of wind turbines and photovoltaic panels in real time, collecting historical data, including environmental factors such as wind speed, wind direction, solar irradiance, and temperature, as well as the corresponding power generation. For example, a wind farm might record wind speed and corresponding turbine power generation every 15 minutes over the past year, forming a large dataset. Based on this historical data, time series analysis methods, such as Autoregressive Moving Average (ARMA) or Long Short-Term Memory (LSTM) networks, are used to construct wind and solar power prediction models. For instance, using wind speed and power generation data from the past 24 hours, an ARMA model can predict wind power for the next 6 hours. After model training and validation, a set of wind and solar power prediction curves for a future period (e.g., the next 24 hours) can be obtained. This set may contain multiple curves, each representing a prediction result at different confidence levels. The input power range of the electrolyzer is pre-defined; for example, the input power range for a certain type of electrolyzer is set to 100 kW to 500 kW. By combining different input power values, the corresponding hydrogen production rate and hydrogen yield can be calculated. For example, when the input power of the electrolyzer is 200 kW, its hydrogen production rate is 4 kg / h; when the input power is 400 kW, the hydrogen production rate is 8 kg / h. This hydrogen production information at different power levels is aggregated to form a hydrogen production equipment operation information set. Based on the hydrogen production of each electrolyzer in this set, the minimum and maximum allowable hydrogen storage capacity of the hydrogen storage tank are set. For example, the minimum storage capacity of the hydrogen storage tank is set to 100 kg, and the maximum to 1000 kg, to ensure the safe and stable operation of the system. Based on the minimum and maximum storage capacities, the capacity range of the hydrogen storage tank can be determined; for example, tanks with capacities of 500 kg, 800 kg, or 1200 kg can be selected. The parameters of these hydrogen storage tanks with different capacities are aggregated to form a hydrogen storage tank parameter set. Based on the hydrogen storage capacity of each hydrogen storage tank in this parameter set, the start-up and shutdown thresholds for the hydrogen fuel cell are set. For example, when the hydrogen storage tank reaches 800 kg, the hydrogen fuel cell starts generating electricity; when the storage drops to 200 kg, the hydrogen fuel cell shuts down. These threshold settings aim to balance the hydrogen storage tank's capacity and power generation demand, avoiding overcharging and over-discharging of the storage tank. Based on these thresholds, a set of control commands for the hydrogen conversion equipment is generated to guide the start-up and shutdown of the hydrogen fuel cell. Combining the set of control commands for the hydrogen conversion equipment with wind and solar power prediction curves, a constraint model is established that includes upper and lower limits for equipment power and capacity limitations for the hydrogen storage tank. For example, the input power of the electrolyzer cannot exceed its rated power, the output power of the hydrogen fuel cell cannot exceed its rated power, and the storage capacity of the hydrogen storage tank must be within its capacity range. These constraints ensure the safety and stability of the system operation. A revenue-cost model is established by integrating the power generation revenue per unit time and the hydrogen production cost per unit time. For example, the power generation revenue can be calculated based on the electricity price and power generation, while the hydrogen production cost can be calculated based on electricity consumption and electricity price. Through the revenue-cost model, the system's net revenue per unit time can be calculated.For example, if the revenue from power generation is 100 yuan and the cost of hydrogen production is 60 yuan in a certain hour, then the net revenue is 40 yuan. A linear programming algorithm is applied to the revenue-cost model and the constraint model to solve for the optimal equipment capacity configuration scheme with the objective of maximizing the total net revenue of the system within a given time period.

[0088] Step S106: Based on the equipment capacity configuration scheme and energy conversion and distribution control signals, input to the simulation platform to obtain system operation simulation results. If the simulation results meet the preset performance indicators of economic benefits and system stability, then the equipment capacity configuration scheme and energy conversion and distribution control signals are the final results; otherwise, adjust the model parameters of the time series analysis model, leakage rate dynamic compensation algorithm, and dynamic programming model based on the simulation results.

[0089] The process begins by acquiring initial data for the equipment capacity configuration scheme and the energy conversion and distribution control signals. Based on this initial data, the system model of the simulation platform is initialized with parameters, resulting in an initialized system simulation model. The time-series data of this initialized model is then input into the simulation platform for simulation, yielding system operation simulation results. This simulation results are then input into a pre-defined economic benefit and system stability evaluation model. Based on the evaluation results output by this model, it is determined whether the simulation results meet the pre-defined performance indicators for economic benefit and system stability. If the simulation results meet these indicators, the equipment capacity configuration scheme and the energy conversion and distribution control signal data are considered the final results. If the simulation results do not meet these indicators, the data is input into a time-series analysis model. Based on the time-series analysis results, the parameter adjustment values ​​for this model are determined. Finally, based on the time-series analysis results, the simulation results are input into a dynamic leakage rate compensation algorithm. Based on the compensation results output by this algorithm, the parameter adjustment values ​​for the dynamic leakage rate compensation algorithm are determined. Based on the time series analysis results and the compensation results output by the leakage rate dynamic compensation algorithm, the simulation results data are input into the dynamic programming model. Based on the planning results output by the dynamic programming model, the parameter adjustment values ​​of the dynamic programming model are determined.

[0090] Specifically, initial data for equipment capacity configuration schemes is acquired, such as an initial capacity of 50 MW for the wind turbine generator set, 30 MW for the photovoltaic power generation system, 20 MW for the electrolyzer, 10 tons for the hydrogen storage tank, and 15 MW for the fuel cell. Simultaneously, initial data for energy conversion and distribution control signals is acquired, such as the electrolyzer start-up signal, and the fuel cell start-up and shutdown thresholds. This initial data will be used to initialize the system model of the simulation platform, which includes a wind power generation model, a photovoltaic power generation model, an electrolyzer model, a hydrogen storage tank model, and a fuel cell model. The aforementioned initial data is then input into the simulation platform to initialize the system model parameters. For example, in a wind power generation model, parameters such as the number of wind turbine generators, individual unit capacity, cut-in wind speed, rated wind speed, and cut-out wind speed are set; in a photovoltaic power generation model, parameters such as the number of photovoltaic panels, peak power of a single panel, irradiance, and temperature are set; in an electrolyzer model, parameters such as the number of electrolyzers, rated power of a single electrolyzer, and electrolysis efficiency are set; in a hydrogen storage tank model, parameters such as the number of hydrogen storage tanks, capacity of a single hydrogen storage tank, and maximum and minimum allowable hydrogen storage capacity are set; and in a fuel cell model, parameters such as the number of fuel cells, rated power of a single fuel cell, and power generation efficiency are set. Through these parameter settings, a complete system simulation model can be established. The time-series data of the initialized system simulation model, such as hourly wind power generation, photovoltaic power generation, hydrogen production rate of the electrolyzer, hydrogen storage capacity of the hydrogen storage tank, and power generation of the fuel cell, are input into the simulation platform for simulation. The simulation platform will simulate the system's operation over a period of time, such as one day, one week, or one month, based on this time-series data. During the simulation, the system model calculates the system state at the next moment based on the current input data and model parameters, and updates the time series data. This method yields simulation results of the system operation. The simulation results are then input into a pre-defined economic benefit and system stability assessment model. The economic benefit assessment model calculates indicators such as total revenue, total cost, and net revenue based on the simulation results. For example, total revenue can be calculated based on the fuel cell's power generation and electricity price; total cost can be calculated based on the operation and maintenance costs of wind and solar power generation, the operating costs of the electrolyzer, and the operating costs of the fuel cell; and net revenue is the total revenue minus the total cost. The system stability assessment model analyzes system stability indicators based on the simulation results, such as whether the hydrogen storage capacity of the hydrogen tank remains within a safe range and whether the fuel cell's output power is stable. Based on the output of the economic benefit and system stability assessment models, it is determined whether the simulation results meet the pre-defined performance indicators.For example, preset economic benefit indicators could be: the total net profit of the system during the simulation period is greater than 100,000 yuan; preset system stability indicators could be: the hydrogen storage capacity of the hydrogen storage tank is always between 2 and 8 tons, and the output power fluctuation range of the fuel cell does not exceed 5%. If the simulation results meet the preset performance indicators, the current equipment capacity configuration scheme, energy conversion, and distribution control signal data are determined as the final results. If the simulation results do not meet the preset performance indicators, the model parameters need to be adjusted. The simulation results data are input into a time series analysis model to analyze the characteristics of the time series data, such as periodicity, trend, and randomness. Based on the time series analysis results, the parameter adjustment values ​​of the time series analysis model are determined, such as the order of the autoregressive model and the order of the moving average model. These parameter adjustments can improve the prediction accuracy of the time series analysis model, thereby better guiding subsequent parameter adjustments. The simulation results data are input into a dynamic leakage rate compensation algorithm to calculate the hydrogen leakage rate based on the changes in the hydrogen storage capacity of the hydrogen storage tank and to dynamically compensate for the leakage rate. Based on the compensation results output by the dynamic leakage rate compensation algorithm, the parameter adjustment values ​​of the dynamic leakage rate compensation algorithm are determined, such as the compensation coefficient and compensation period.

[0091] On the other hand, this embodiment also provides a multi-energy system optimization method based on hydrogen energy storage, including:

[0092] The data acquisition module is used to acquire historical wind speed, wind direction, and solar irradiance data of the wind measurement tower and photovoltaic power station. The time series analysis model is trained using the historical wind speed, wind direction, and solar irradiance data to obtain the wind power and solar power prediction curves for future periods.

[0093] The time series analysis module is used to determine the start-up and shutdown status of the hydrogen production equipment at each moment based on the wind power and solar power prediction curves for the future period, and to obtain the operating data of the hydrogen production equipment.

[0094] The hydrogen loss monitoring module is used to build a dynamic compensation model for leakage rate based on the gradient boosting decision tree algorithm. After collecting the pressure and temperature data of the hydrogen storage tank, it is input into the dynamic compensation model for leakage rate to obtain optimized operating parameters of the hydrogen storage tank.

[0095] The hydrogen conversion optimization module is used to collect current, voltage and hydrogen flow data of the hydrogen conversion equipment, and plot the efficiency curve of the hydrogen conversion equipment based on the current, voltage and hydrogen flow data. Combined with hydrogen demand and system load data, the hydrogen demand, system load and efficiency curve data are input into the dynamic programming model to obtain the control signal of the hydrogen conversion equipment.

[0096] The system economic benefit optimization module is used to generate an equipment capacity configuration scheme that maximizes the system economic benefits based on the wind power and solar power prediction curves for the future period, the operating data of the hydrogen production equipment, the optimized operating parameters of the hydrogen storage tank, and the control signals of the hydrogen conversion equipment as inputs to a linear programming algorithm.

[0097] The simulation verification module is used to optimize the multi-energy system based on the equipment capacity configuration scheme that maximizes the system's economic benefits.

[0098] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor implements the method when executing the computer program.

[0099] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method.

[0100] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An optimization method for multi-energy systems based on hydrogen energy storage, characterized in that, include: Historical wind speed, wind direction, and solar irradiance data from wind measurement towers and photovoltaic power plants are obtained. Time series analysis models are trained using historical wind speed, wind direction, and solar irradiance data to obtain prediction curves for wind power and solar power in future periods. The start-up and shutdown status of hydrogen production equipment at each moment is determined based on wind power and solar power prediction curves, and the operating data of hydrogen production equipment is obtained. A dynamic compensation model for leakage rate is constructed based on the gradient boosting decision tree algorithm. After collecting the pressure and temperature data of the hydrogen storage tank, the data is input into the dynamic compensation model for leakage rate to obtain optimized operating parameters of the hydrogen storage tank. Collect current, voltage, and hydrogen flow data of the hydrogen conversion equipment, plot the efficiency curve of the hydrogen conversion equipment based on the current, voltage, and hydrogen flow data, and input the hydrogen demand, system load data, and efficiency curve data into the dynamic programming model to obtain the control signal of the hydrogen conversion equipment. Based on the wind power and solar power prediction curves, hydrogen production equipment operation data, hydrogen storage tank operation parameters, and hydrogen conversion equipment control signals as inputs to the linear programming algorithm, an equipment capacity configuration scheme that maximizes the system's economic benefits is generated. Optimize multi-energy systems based on equipment capacity configuration schemes that maximize system economic benefits.

2. The method according to claim 1, characterized in that, The process of obtaining the wind power and solar power prediction curves for future time periods includes: Acquire historical wind speed data, historical wind direction data, and historical solar irradiance data recorded by the wind measurement tower and the photovoltaic power station. Perform data quality checks on the historical wind speed data, historical wind direction data, and historical solar irradiance data to determine if there are missing values. If there are missing values, use linear interpolation to fill in the missing values. Outlier detection was performed on the historical wind speed data, historical wind direction data, and historical solar irradiance data after the filling process was completed. The box plot method was used to determine whether there were outliers in the data. If outliers were found, the Laida criterion was used to remove them and obtain valid data. Based on the geographical location information of the wind measurement towers and the photovoltaic power stations, spatial matching is performed on the valid data. The data of the wind measurement towers with the closest spatial distance are matched with the data of the photovoltaic power stations to obtain the matched dataset. The matched dataset is time-aligned to unify the data time resolution and obtain an aligned dataset. A long short-term memory network is then used to extract features from the aligned dataset to obtain a multi-dimensional feature vector. Based on the multidimensional feature vector, wind power prediction model and solar power prediction model are constructed, and wind power and solar power prediction curves for future periods are obtained based on the wind power prediction model and solar power prediction model.

3. The method according to claim 1, characterized in that, The process of determining the start-up and shutdown status of the hydrogen production equipment at each moment based on the predicted wind and solar power curves for the future time period, and obtaining the operating data of the hydrogen production equipment, includes: Based on the wind power and solar power prediction curves for the future period, wind power data and solar power data for each moment in the prediction time series are obtained. Based on the preset electrolyzer start-up threshold, the wind power data and solar power data at each time point are compared with the electrolyzer start-up threshold to obtain the start-up and shutdown status of the hydrogen production equipment at each time point. Based on the start-up and shutdown status of the hydrogen production equipment at each time point, the rated power of the electrolyzer is obtained, and the operating status of the electrolyzer at each time point is determined. By using the operating status of the electrolyzer at each time point, and employing the random forest algorithm in conjunction with wind power data and solar power data, the predicted power of the electrolyzer at each time point is obtained. Based on the predicted power of the electrolyzer at each time point and combined with the rated power of the electrolyzer, the minimum value between the two is taken to obtain the operating data of the hydrogen production equipment at each time point.

4. The method according to claim 1, characterized in that, The process of obtaining optimized operating parameters for the hydrogen storage tank includes: The data acquisition unit obtains the pressure value from the pressure sensor and the temperature value from the temperature sensor of the hydrogen storage tank to obtain the real-time pressure value and real-time temperature value. The real-time pressure and temperature values ​​are input into the leakage rate dynamic compensation model based on the gradient boosting decision tree algorithm to obtain the hydrogen loss compensation value. The initial value of hydrogen loss is calculated based on the real-time pressure value and the real-time temperature value, and the amount of hydrogen loss is obtained based on the initial value of hydrogen loss and the hydrogen loss compensation value. A hydrogen loss threshold is set, and the amount of hydrogen loss is compared with the hydrogen loss threshold to obtain a comparison result. Based on the comparison result, the operating parameters of the hydrogen storage tank are optimized to obtain optimized operating parameters of the hydrogen storage tank.

5. The method according to claim 1, characterized in that, The process of obtaining the control signal for the hydrogen conversion device includes: Collect operating data of the hydrogen conversion equipment, including current intensity, voltage values, and hydrogen flow rate data; Based on the collected current intensity, voltage values ​​and hydrogen flow data, the conversion efficiency of the equipment under different operating conditions was calculated, and the efficiency curve was plotted. Obtain hydrogen demand data and system load data, and use the hydrogen demand data, system load data and efficiency curve data as input to the dynamic programming model to obtain the optimal hydrogen conversion timing. The control signal for the hydrogen conversion device is obtained based on the optimal hydrogen conversion timing.

6. The method according to claim 5, characterized in that, The process of obtaining the optimal hydrogen conversion timing includes: Historical hydrogen demand data, system load data, and equipment efficiency curve data are acquired and preprocessed using data cleaning methods to obtain preprocessed hydrogen demand data, preprocessed system load data, and preprocessed equipment efficiency curve data. Based on the preprocessed hydrogen demand data and preprocessed system load data, a time series forecasting algorithm is used to obtain the predicted hydrogen demand data and the predicted system load data for future time periods. Based on the predicted hydrogen demand data and the predicted system load data at future times, a linear regression algorithm is used to obtain the efficiency values ​​of the equipment under different loads. The efficiency values ​​of the equipment under different loads and the predicted hydrogen demand data at future times are calculated to obtain the optimal operating power of the hydrogen production equipment, the capacity of the hydrogen storage equipment, and the power of the hydrogen release equipment. The operating power, hydrogen storage capacity, and hydrogen release power of the optimal hydrogen production equipment are compared with the real-time operating power, hydrogen storage capacity, and hydrogen release power of the hydrogen production equipment to obtain the comparison results. Based on the comparison results, the hydrogen production equipment is adjusted, and the updated hydrogen capacity in the hydrogen storage equipment is obtained based on the adjusted equipment parameter data. Based on the updated hydrogen capacity in the hydrogen storage equipment, the optimal hydrogen conversion timing is obtained.

7. The method according to any one of claims 1-6, characterized in that, The process of generating a device capacity configuration scheme that maximizes the system's economic benefits includes: A set of hydrogen production equipment operation information is obtained based on the wind power and solar power prediction curves for the future time period. Based on the hydrogen production of each electrolyzer in the hydrogen production equipment operation information set, and combined with the minimum and maximum allowable hydrogen storage capacity of the hydrogen storage tank, a set of parameters for the hydrogen storage tank is obtained. Based on the hydrogen storage capacity of each hydrogen storage tank in the parameter set, the start-up and shutdown thresholds of the hydrogen fuel cell are set, and a set of control commands for the hydrogen conversion device is generated. Based on the set of parameters of the hydrogen storage tank and the set of operating information of the hydrogen production equipment, a constraint model is constructed that includes the upper and lower limits of equipment power and the capacity limit of the hydrogen storage tank. Based on the constraint model, the operating constraints of the equipment are obtained. Based on the equipment operation constraints, a revenue-cost model is established by integrating the power generation revenue per unit time and the hydrogen production cost per unit time, and the net system revenue per unit time is obtained based on the revenue-cost model. Based on the linear programming algorithm, the operating constraints of the equipment and the net system revenue per unit time are calculated to obtain the equipment capacity configuration scheme that maximizes the total net system revenue within a set time.

8. A multi-energy system optimization system based on hydrogen energy storage, characterized in that, include: The data acquisition module is used to acquire historical wind speed, wind direction, and solar irradiance data of the wind measurement tower and photovoltaic power station. The time series analysis model is trained using the historical wind speed, wind direction, and solar irradiance data to obtain the wind power and solar power prediction curves for future periods. The time series analysis module is used to determine the start-up and shutdown status of the hydrogen production equipment at each moment based on the wind power and solar power prediction curves for the future period, and to obtain the operating data of the hydrogen production equipment. The hydrogen loss monitoring module is used to build a dynamic compensation model for leakage rate based on the gradient boosting decision tree algorithm. After collecting the pressure and temperature data of the hydrogen storage tank, it is input into the dynamic compensation model for leakage rate to obtain optimized operating parameters of the hydrogen storage tank. The hydrogen conversion optimization module is used to collect current, voltage and hydrogen flow data of the hydrogen conversion equipment, and plot the efficiency curve of the hydrogen conversion equipment based on the current, voltage and hydrogen flow data. Combined with hydrogen demand and system load data, the hydrogen demand, system load and efficiency curve data are input into the dynamic programming model to obtain the control signal of the hydrogen conversion equipment. The system economic benefit optimization module is used to generate an equipment capacity configuration scheme that maximizes the system economic benefits based on the wind power and solar power prediction curves for the future period, the operating data of the hydrogen production equipment, the optimized operating parameters of the hydrogen storage tank, and the control signals of the hydrogen conversion equipment as inputs to a linear programming algorithm. The simulation verification module is used to optimize the multi-energy system based on the equipment capacity configuration scheme that maximizes the system's economic benefits.

9. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-7.

Images